RESEARCH ARTICLE

Yeast-Driven and Bioimpedance-Sensitive

Biohybrid Soft Robots Citation: Soliman M, Forbes F,
MennaAllah Soliman1*, Frederick Forbes1,2, and Dana D. Damian1,2,3* Damian DD. Yeast-Driven and
Bioimpedance-Sensitive Biohybrid
1
School of Electrical and Electronic Engineering, University of Sheffield, Sheffield S1 3JD, UK. 2Insigneo Soft Robots. Cyborg Bionic Syst.
2025;6:Article 0233. https://doi.
Institute for in silico Medicine, University of Sheffield, Sheffield, UK. 3Sheffield Robotics, University of org/10.34133/cbsystems.0233
Sheffield, Sheffield, UK.
Submitted 31 August 2024
Revised 8 December 2024
*Address correspondence to: [email protected] (D.D.D.); [email protected] (M.S.) Accepted 23 January 2025
Published 25 April 2025
Biohybrid robots integrate biological components with synthetic materials to harness the unique capabilities
of living systems for robotic functions. This study focuses on leveraging yeast fermentation dynamics to Copyright © 2025 MennaAllah
enable actuation and sensing in soft robotic systems. By leveraging yeast’s natural ability to produce Soliman et al. Exclusive licensee
Beijing Institute of Technology Press.
carbon dioxide and generate pressure during fermentation, we demonstrate the feasibility of creating No claim to original U.S. Government
biohybrid robots with lifelike behavior and adaptability. Our research integrates bioimpedance sensing Works. Distributed under a Creative

Downloaded from https://spj.science.org on May 07, 2025

into track yeast behavior and metabolic dynamics in real time. We developed an adjustable single-resistor Commons Attribution License
oscillator circuit by using a digital potentiometer to measure impedance frequency and model the yeast (CC BY 4.0).
growth rate. Experimental results reveal the sensitivity of the single-resistor oscillator circuit to variations
in yeast concentration and demonstrate the correlation between yeast behavior and actuation power.
Furthermore, we highlight the potential of yeast-driven robots for various applications by demonstrating
a yeast-driven soft limb capable of rotating 140° tested at different temperatures, an inflatable membrane
actuator functioning as a tactile sensor detecting forces up to 4.5 N, a palpation probe for differentiating
tissue stiffness, and a gripper capable of manipulating objects. This work lays the foundation for advancing
biohybrid robotics by integrating yeast fermentation dynamics with bioimpedance sensing, enhancing the
functionality of robotic systems.

Introduction Tissue-based biohybrid robots are constructed by incorpo-

rating tissue such as cardiomyocytes obtained from neonatal
Soft robots, inspired by the adaptability of biological organisms, rats or derived from human induced pluripotent stem cells onto
combine sensing, actuation, and control to create highly flexible robotic or soft structures. One notable example is xenobots, a
and versatile robotic systems [1]. However, replicating the intri- method for creating in vitro biological robots introduced by
cate communication and intelligence living cells exhibit remains Blackiston et al. [14]. Recently, Kim et al. [15] presented a new
a challenge [2]. Rossiter [3] asserts the necessity of conceptual- type of robot that combines living cells with electronic compo-
izing future robots as robotic organisms rather than machines, nents to create motion. These biohybrid machines can be wire-
advocating for a paradigm shift in robotics systems. Recent lessly controlled and perform walking, turning, and object
advancements, such as those highlighted by Zhou et al. [4], transportation functions. The study by Filippi et al. [16] inte-
emphasize the integration of actuation and sensing to achieve grated artificial structures with living biosystems and was
intelligent soft robots capable of adapting to their environ- advancing toward autonomous control and state awareness
ments. Biohybrid robots integrate biological cells with artificial through innovations like piezoresistive sensors that detect and
components to harness cells’ inherent actuation and sensing respond to muscle contractions, thus enhancing their function-
capabilities [5,6]. These microsystems are designed for micro- ality and intelligence. Advancements in control mechanisms
environment operation, including ex vivo and in vivo settings for muscle-powered robots, highlighting the potential of bio-
[7], and have the potential for intelligent drug delivery, with logical actuators in soft robotics, are discussed by Bawa and
applications in medicine and healthcare and challenges for Raman [17]. Additionally, the progression from biomimicry to
clinical translation [8]. Furthermore, stimuli-responsive living biohybrid system design emphasizing dynamic adaptation in
materials and biosensing technologies open new boundaries engineered systems is explored by Raman and Bashir [18].
in environmental adaptation and controllability over biohybrid Incorporating such control and design principles has broad-
robots [9]. Biohybrid robots can be categorized into 4 groups ened the scope of biohybrid robotics, particularly in creating
based on their composition and design: tissue-based biohybrids adaptable and intelligent systems.
[10], microbial biohybrids [11], DNA biohybrid robots [12], Microbial biohybrid robots consist of unicellular microorgan-
and cyborg robots [13]. Among these categories, tissue-based isms, predominantly bacteria, integrated with micro/nanoscale
biohybrid robots are the most well-known. functional components. They have demonstrated controlled

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 1

actuation for cargo transport and assembly, and recent advance- The first step for biohybrid control is feedback from biological
ments have led to the development of specialized microrobotic component status and how to sense it. Bioimpedance is a non-
systems for therapeutic cargo delivery [11], biofilm treatment invasive technique used to measure the electrical properties of
[19], and high-throughput micro-object manipulation [20]. biological tissues. The living organism’s components act as elec-
Another group used yeast and demonstrated thin film ZnO/Si trical impedance elements by applying electrical current to the
surface acoustic waves to manipulate microparticles and yeast tissue [26]. Bioimpedance technology offers a promising solution
cells in a microchamber, enabling 3-dimensional patterning in by enabling the development of biohybrid robots that can sense
the study by Tao et al. [21]. Using yeast as a biosensing platform, their internal state and detect external environmental effects,
the study by Crnković et al. [22] demonstrated that a living yeast enhancing their observability and controllability. Different elec-
biosensor can effectively detect and differentiate single-amino- trical models, including integer-based [27] and fractional order
acid changes in peptide ligands, suggesting potential applications models [28], represent the electrical behavior of biological tis-
in diagnostics and scalable communication systems. sues. Bioimpedance was used in soft robotics as a sensor for
Critical challenges biohybrid robots face include biohybrid external parameters, and a portable soft gripper was designed in
robot design and modeling, scalable fabrication, autonomous a study [29] to measure the quality of fruits, such as apples, with-
control, and translating them beyond the lab [23]. These com- out damaging their tissues, as bioimpedance and soft materials
plexities appear due to the multifaceted nature of biological and are nonharmful to fruits. Additionally, a single-resistor oscil-
robotic components and their dynamic interactions. Designing lator (SCRO) circuit combined with an analog multiplexer was
such robots necessitates a profound understanding of biological employed in a different study [30] to design soft E-skin using
behaviors and robotic principles, demanding the harmonious inte- aloe vera gel.
gration of these domains while considering nonlinear biological In this study, our contributions include employing bioimped-
responses and uncertainties in modeling [24]. Achieving auton- ance technology to characterize the electrical properties of bio-
omous control in unpredictable biological dynamics demands logical tissues and cells using an adjustable SCRO as a sensor,
advanced control algorithms that can adapt to variations. which uses a digital potentiometer to measure the impedance

Downloaded from https://spj.science.org on May 07, 2025

Translating biohybrid robots beyond controlled lab environ- frequency for a biohybrid system described in Fig. 1. We inves-
ments involves navigating challenges related to stability to live tigated the relationship between actuator motion and the state of
in uncontrolled environments and ethical considerations [25]. yeast cells using impedance frequency readings from the SCRO
While cell-based biohybrid robots rely on muscle tissue for actua- circuit, aiming to leverage yeast as a potential proprioceptive and
tion, our study demonstrates a microbial approach using yeast exteroceptive sensor for biohybrid robotic systems. Changes in
fermentation for actuation and sensing. This approach explores impedance frequency reflect alterations in the yeast’s metabolic
the potential of yeast as a scalable, renewable, and biocompatible activity and cell structure, providing a direct means to track gas
actuation source, differing from cell-driven systems. By leverag- production and growth dynamics. The SCRO offers several
ing yeast’s unique metabolic properties, we aim to expand the advantages over traditional bioimpedance measurement meth-
possibilities for biohybrid systems in applications that demand ods, including simplicity, cost-effectiveness, and compactness,
environmental adaptability and simplicity in design. making it particularly suitable for portable and scalable systems.

Frequency cell activity

F (Hz)
F (Hz)
Frequency
Frequency

Time Time

Robot Move

Stimulus signal
(increase temperature, sugar concentration)
Fig. 1. Bioimpedance sensing for a biohybrid robot, where yeast fermentation plays a crucial role in actuation and sensing. During fermentation, yeast cells undergo division,
which affects the yeast culture’s impedance. Concurrently, carbon dioxide is produced as a byproduct of fermentation, which the robot utilizes as a power source for motion.
The bioimpedance frequency serves as a feedback sensor for monitoring the biological components in biohybrid.

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 2

Furthermore, its design allows for miniaturization, enabling its T
 is the doubling time of the culture. After several doubling times,
use in biomedical applications, such as implantable biosensors yeast enters a stationary phase in which ethanol is produced,
for tracking cellular or tissue activity. Unlike conventional imped- characterized by slowed growth due to nutrient depletion. Cells
ance analyzers, the SCRO provides continuous, noninvasive in this phase undergo physiological changes, including carbohy-
monitoring of impedance frequency changes, directly reflecting drate accumulation and enhanced cell wall resilience. This phase
yeast metabolic activity. Additionally, our implementation fea- can persist until conditions improve, but if not, cells enter the
tures a digital potentiometer, allowing precise sensitivity adjust- death phase [31]. The exponential model describes the yeast
ments to accommodate varying fermentation conditions. growth phase, as shown in Eqs. (1) and (2). A detailed visualiza-
In addition, we studied the effect of yeast growth on mechani- tion of this process, including the impedance frequency changes
cal properties such as pressure, volume, and force and established over time, is provided in Fig. S3. There are 2 primary types of
correlations with yeast frequency. We also modeled the relation- yeast fermentation: aerobic fermentation, which occurs in open
ship between pressure and yeast frequency to correlate the actua- air, and anaerobic fermentation. Anaerobic fermentation occurs
tion power with the dynamic behavior of yeast. Furthermore, we when yeast ferments in a closed chamber, producing carbon
studied the regeneration of the actuation power using yeast and dioxide pressure as a byproduct of yeast growth. In this study, we
its impact on frequency sensing. Yeast offers several advantages utilize anaerobic fermentation as an actuation power while simul-
compared to traditional actuation and sensing systems. Unlike taneously monitoring the dynamics of yeast growth, which are
conventional systems requiring pumps, valves, or batteries, yeast then transferred to actuator dynamics for further analysis.
fermentation provides a natural and biodegradable source of
actuation power through carbon dioxide generation. This unte- Oscillator circuit and bioimpedance
thered approach simplifies the design by eliminating the need for model for yeast dynamics
additional mechanical components and external energy sources. Bioimpedance presents a noninvasive approach for tracking
Moreover, yeast systems are highly adaptive and cost-effective, various parameters of yeast fermentation, such as cell density,
making them suitable for scalable and sustainable robotic applica- growth rate, and environmental factors (e.g., temperature or

Downloaded from https://spj.science.org on May 07, 2025

tions both inside the body as biomedical applications and in envi- pH). It enables continuous monitoring of yeast behavior as it
ronmental settings. Bioimpedance allows yeast to act as both an metabolizes nutrients and produces metabolites. The method-
actuation element and a sensor, providing a dual-purpose capabil- ology involves measuring changes in impedance frequency
ity without the need for additional external sensors. Our study over time between 2 electrodes immersed in the fermentation
analyzed yeast’s actuation and sensing capabilities across 2 soft medium using the oscillator circuit. As yeast cells undergo
robotic systems: an inflatable membrane actuator and a soft limb growth and metabolic processes, they alter the electrical proper-
actuator. We evaluated the influence of temperature on controlling ties of the medium, leading to changes in impedance that the
a yeast-actuated soft limb and its effect on yeast impedance. electrodes can detect. The electrical properties of the biological
Additionally, we explore using an inflated membrane actuator component are modeled using the Cole model of bioimped-
filled with yeast to study how external forces affect internal pres- ance shown in Eq. (3). This model effectively characterizes
sure and how impedance can detect these changes, enabling its biological tissue by combining resistance and capacitance.
application as tactile sensing. This method allows the use of yeast Under low-current conditions, the cell wall behaves as a
impedance to monitor not only the internal states of robotic sys- capacitance, while the intracellular liquid exhibits low resis-
tems but also environmental changes. This study highlights the tance, referred to as Rcw. In contrast, the extracellular liquid
potential of yeast-driven systems in practical applications by dem- displays high-resistance characteristics referred to as REx with
onstrating a yeast-driven probe for tissue palpation, capable 𝜏 as the time constant and 𝛼 as the Warburg coefficient. The
of differentiating tissue stiffness, and a yeast-driven gripper Warburg coefficient represents the impedance contribution
for object grasping, showcasing their sensitivity and mechani- from diffusion-controlled processes, such as the movement
cal responsiveness through bioimpedance sensing. of ions or molecules within the fermentation medium. In our
bioimpedance model, this coefficient captures the diffusion
dynamics associated with yeast metabolic activity and contrib-
Materials and Methods utes to understanding the system’s electrochemical behavior
Yeast’s metabolism characteristics over time. This model helps elucidate yeast cells’ complex elec-
In microbiology, yeast strains, such as Saccharomyces cerevisiae, trical behavior during fermentation processes.
are easily cultivated in nutrient-rich media containing carbon 
and nitrogen sources, vitamins, salts, and essential minerals at R −R
Z(𝜔) = REx + cw( Ex (3)
a temperature of around 30 to 40 °C. The growth follows a
)𝛼
1 + j𝜔𝜏
recognizable pattern—beginning with a lag phase after injec-
tion, followed by an exponential growth phase where cells mul- The yeast impedance measurement methodology followed is
tiply rapidly and produce more carbon dioxide. The exponential an indirect approach, as introduced by Mohsen et al. [32]. This
growth can be mathematically described, with the growth rate method employs an SCRO circuit in Fig. 2A. Its variable com-
varying with temperature in Eq. (1). ponents are adjusted based on the Cole model of yeast growth
 rate shown in Table 1 and Eq. (3), depicted in Eqs. (4) and (5).

N = N0 ekt (1)
( )
2 (𝛼 + 1)𝜋
Rcw 900.49 × 103 REx − 12.14 × 107 𝜔𝛼+1 cos
( )
Real = 21Ccw
2
k = ln(2) ∕ T (2)

( ) 𝛼 ( 𝛼𝜋 ) −6
+ Ccw Rcw 44.7REx − 2.4 𝜔 cos + 4.47 × 10 REx Rcw
N is the number of yeast cells over time, and N0 is the initial 2
−9
number. k is the growth constant represented in Eq. (2), where + 44.7REx + 100 × 10 Rcw − 2.4 = 0 (4)

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 3

 A

Intercellular
Rcw

Extracellular
REx

Downloaded from https://spj.science.org on May 07, 2025

B

Actuation chamber

Arduino Pressure sensor

Yeast chamber

NI DAQ device SCRO circuit

Fig. 2. Bioimpedance and pressure measurement system for yeast-driven actuation during anaerobic fermentation. (A) Indirect bioimpedance measurement system using a
single-resistor oscillator (SCRO) and a yeast solution equivalent circuit, where Rcw is the capacitance and resistance of the cell wall and inside the cell liquid, while REx is the
resistance of the extracellular liquid. (B) Experimental setup for measuring the frequency and pressure of yeast during anaerobic fermentation. The yeast chamber is connected
to the designed SCRO circuit to read impedance frequency, and the produced gas is trapped inside a 5-ml actuation chamber. This chamber is connected to a pressure sensor
to detect changes in pressure due to yeast growth and impedance changes. GND, ground; DAQ, data acquisition; NI, National Instruments.

Imag. = in the SCRO circuit has been replaced with a digital potenti-
 (𝛼 + 1)𝜋
( ) ometer controlled by an Arduino microcontroller. The Arduino
2
Rcw 900.49 × 103 REx − 12.14 × 107 𝜔𝛼+1 sin
( )
21Ccw
2
+ CcwRcw adjusts the resistance values in real time to achieve oscillation
( ) (
𝛼𝜋
) ( ) in the circuit, enabling precise tuning for optimal bioimpedance
44.7REx − 2.4 𝜔𝛼 sin
2
+ 21Ccw 2.03REx + 2.03Rcw + 0.2 𝜔 = 0 measurements. This approach ensures sensitivity and adapt-
(5) ability in dynamic conditions, making it particularly effective
for monitoring yeast fermentation processes. In this study, the
Notably, the digital potentiometer in the study is adjusted SCRO circuit was simulated using the yeast Cole model param-
to achieve oscillation around the bioimpedance frequency. eters reported by Wang et al. [33] (see Table 1) to account for
Central to this concept is the inverse relationship between the bioimpedance characteristics of yeast at different growth
capacitance and frequency. By applying an appropriate voltage stages. The digital potentiometer (Rv) was iteratively adjusted
to the oscillator, frequency amplification is achieved, enabling during the simulations to achieve oscillation within the desired
effective measurement. In this work, the digital potentiometer frequency range. The resulting SCRO component values, spanning

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 4

10 to 15 kΩ, correspond to the impedance range of yeast during (pressure) and yeast growth kinetics. The pressure dynamics
fermentation and are summarized in Table 2. paralleled yeast growth behavior, showing an initial lag phase
By utilizing yeast dynamics (Eqs. (1) and (2)), represented followed by a rapid increase during exponential growth. By
by the biological factors of fermentation, along with the SCRO correlating the frequency impedance, which represents yeast
circuit equations (Eqs. (4) and (5)), we model the yeast growth growth, with the pressure changes observed in the experiment,
rate to generate actuation power as an electric model. This we demonstrated a direct relationship between yeast activity
model correlates impedance frequency with actuator outputs, and gas production within the chamber.
such as pressure or deflection angle, enabling the system to
predict and control desired behaviors. For instance, the desired Yeast impedance frequency
behavior in a soft robotic membrane actuator might involve The initial phase of the study aims to assess the SCRO circuit’s
achieving a target pressure for inflation. Similarly, the desired capability to detect yeast dynamics. For testing, a yeast mix-
behavior in a soft limb actuator could be a controllable deflec- ture consisting of 10 ml of water at 38 to 40 °C mixed with 2.25 g
tion angle (e.g., 140°). This capability makes bioimpedance a of yeast (22.5% concentration) is utilized. We exploit the sugar
dual-purpose sensor: it monitors the yeast’s state and provides concentration to examine its impact on yeast frequency and
feedback to predict actuator performance. output actuation power, employing concentrations of 3% (0.3 g),
6% (0.6 g), 12% (1.2 g), and 21% (2.1 g) sugar in yeast mixture
Yeast-based biohybrid actuation power preparation as shown in Fig. S1.
This investigation focuses on the potential of yeast-produced The yeast mixture is added into the yeast chamber, as illus-
gas to actuate a soft pneumatic actuator. Central to this mecha- trated in Fig. 2B, where it undergoes anaerobic fermentation,
nism is the gas volume generated by yeast and the resulting generating carbon dioxide and inducing pressure in the actua-
pressure exerted. The experimental setup, shown in Fig. 2B, is tor chamber, monitored by the pressure sensor. The experiment
designed to measure pressure within a fixed volume, utilizing includes 5 trials for each sugar concentration, each lasting
10 min. These parameters are determined based on findings

Downloaded from https://spj.science.org on May 07, 2025

a 5-ml syringe as the actuation chamber and applying the ideal
gas law to analyze the process. The yeast chamber is linked to from the yeast literature [34], as excessive sugar can be lethal
the actuation chamber, which is connected to a pressure sensor to yeast, while insufficient sugar may impede yeast reactivity.
(XGZP6847A Pressure Sensor Module) with a range of 0 to Fig. 3A shows the frequency changes over time for 5 trials,
7 bar. The pressure sensor reads pressure with an Arduino Uno with the mean and standard deviation. Initially, the frequency
every second. The yeast chamber was constructed from a 10-ml increases with higher sugar concentrations, indicating increased
fully closed syringe, with impedance electrodes positioned 3 cm yeast activity and growth rate. However, after 1.5 min, the fre-
apart at the bottom. These electrodes are connected to the quency for the 21% concentration drops below those of the
SCRO circuit described above to measure the impedance fre- 12% and 6% concentrations. Notably, the 6% concentration
quency. The circuit is interfaced with Arduino Uno to regulate consistently exhibits a higher frequency than the 12% and 21%
the digital potentiometer and provide a 5-V power supply. concentrations.
A National Instruments data acquisition system acquired the The yeast frequency for 3%, 12%, and 21% sugar concentra-
impedance frequency readings using the DAQ Express soft- tions are evaluated in aerobic fermentation, as depicted in Fig.
ware. The experiment measured pressure and yeast impedance S2, by measuring yeast frequency in a Petri dish equipped with
frequency to investigate the correlation between physical power an impedance electrode. Results show increased frequency with
escalating sugar concentration. The disparity observed in the
frequency results between aerobic and anaerobic fermentation
can be attributed to the influence of pressure from the anaerobic
fermentation on yeast-produced carbon dioxide. In anaerobic
Table 1. Yeast solution Cole model parameters [33] fermentation, per Henry’s law, a portion of the dissolved carbon
dioxide gas in the yeast mixture transforms into foam as it
Cole model parameters is in a closed container, affecting the output pressure and the
Yeast Ccw (nF) Rcw (Ω) REx (Ω) measured frequency [35]. This phenomenon may occur due to
the altered physical and chemical conditions within the fer-
5% living 17.68 2.82 25.74 mentation environment, leading to variations in gas dissolution
10% living 15.40 3.14 20.68 and subsequent foam formation. This explanation aligns with
20% living 7.42 5.62 40.52 scientific principles governing gas–liquid interactions and pro-
vides insight into the observed differences in frequency results
between the 2 fermentation conditions.
In Fig. 3B, the imaginary and real parts of the frequency
were calculated using Eqs. (4) and (5), respectively, with a 0.9
alpha coefficient based on the yeast Cole model. Due to the
experimental setup, which involves capturing yeast frequency
Table 2. SCRO oscillator parameters data for 10 min from the yeast’s significantly longer lifespan
(which can last for months), the lifespan frequency of yeast
R1 R2 R3 R4 R5 C Rv shown in Fig. S3, the Nyquist plot displays only a partial curve
(kΩ) (kΩ) (kΩ) (kΩ) (kΩ) (nF) (kΩ) due to a restricted frequency range. Despite this limitation, the
15 300 20 20 10 22 10–15 Nyquist plot reveals that the sugar concentrations exhibit simi-
lar frequency ranges. In contrast, the rate of frequency change
over time differs, as illustrated in Fig. 3A. This distinction

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 5

A B Pressure over time at anaerobic fermentation
Frequency of yeast impedance at anaerobic fermentation
8
0.8

Pressure (bar)
Frequency (kHz)

0.6

4
0.4

2 Sugar concentration 3% Sugar concentration 3%

0.2
Sugar concentration 6% Sugar concentration 6%
Sugar concentration 12% Sugar concentration 12%
Sugar concentration 21% Sugar concentration 21%
0 0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
Time (s) Time (s)

C D
10 9 Nyquist plot for Anaerobic fermentation frequency Pressure vs. Frequency of yeast at anaerobic fermentation
3.5 8

3
6

Frequency (kHz)
Imaginary part

2.5

2 4

1.5
Sugar concentration 3% 2 Sugar concentration 3%
1 Sugar concentration 6% Sugar concentration 6%
Sugar concentration 12% Sugar concentration 12%
Sugar concentration 21% Sugar concentration 21%

Downloaded from https://spj.science.org on May 07, 2025

0.5 0
0 2 4 6 8 10 12 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Real part 20
10 Pressure (bar)

Fig. 3. Yeast impedance frequency and yeast pressure at different sugar concentrations during anaerobic fermentation inside the actuation chamber. (A) Change in frequency over
time at different sugar concentrations. The shaded area represents the 95% confidence level over 5 trials for each concentration, each with a 10-min duration, except for the 21%
concentration due to foam filling the actuation chamber around 8 min. (B) Change in pressure over time for 5 trials, each with a 10-min duration, except for the 21% concentration
due to foam filling the actuation chamber around 8 min affecting pressure reading. The solid line represents the mean, and the shaded areas represent the standard deviation over
5 trials for each concentration. (C) Nyquist plot of impedance frequency’s real and imaginary parts at different sugar concentrations. (D) Change in impedance frequency inside
the yeast chamber versus pressure inside the actuation chamber over 5 trials, each with a 10-min duration. The shaded area indicates the 95% confidence level for the 5 trials.

underscores the distinct dynamics associated with varying sugar pressure changes, facilitating the establishment of a model
concentrations during fermentation. Eventually, the SCRO correlating yeast state with output actuation power at constant
circuit effectively detects and represents the changes occurring volume.
within the yeast, highlighting its ability to capture the nuanced The relationship between pressure and the frequency rate
dynamics of yeast fermentation. of change over time for each sugar concentration is illustrated
in Fig. 4. The experimental data average was fitted to a second-
Pressure profile with the yeast growth rate order exponential equation, resembling the behavior of yeast
The pressure was measured concurrently with frequency over dynamics described in Eq. (1). The coefficient of determina-
5 trials for sugar concentrations of 3%, 6%, 12%, and 21% during tion (R2) values were found to be 0.97, 0.91, 0.93, and 0.94 for
anaerobic fermentation. The accumulation of pressure in the sugar concentrations of 3%, 6%, 12%, and 21%, respectively. These
actuation chamber was monitored by a pressure sensor over R2 values signify a correlation between pressure and the rate
10 min, as displayed in Fig. 3B. Notably, the 3% concentration of yeast growth, as represented by the rate of change in imped-
reached 0.6 bar, the 12% one reached 0.65 bar, the 6% one ance frequency (df/dt). The 12% model was validated by evalu-
peaked at 0.75 bar, and the 21% concentration attained a maxi- ating its ability to predict pressure based on the rate of change
mum pressure of 0.4 bar. These pressure values obtained from in frequency, as shown in Fig. S5. This approach mirrors real-
yeast fermentation proved sufficient for serving as actuation world application scenarios where the rate of change in fre-
power for the soft actuator. In Fig. 3B, it is evident that the 21% quency can be used to estimate the pressure of the actuator
concentration exhibited the lowest pressure over time, while the without needing a dedicated pressure sensor, thereby enabling
3%, 6%, and 12% concentrations showed no significant differ- control of the actuator’s motion. By combining these models
ence in output pressure. However, a significant development with the yeast impedance equations (Eqs. 4 and (5)), we can
occurred as the pressure approached 0.1 bar, as illustrated in elucidate the dynamics of pressure as actuation power in rela-
Fig. S4. At this juncture, yeast foam began infiltrating the actua- tion to yeast growth dynamics. This integration highlights
tion chamber, influenced by the gas bubbles within the yeast the potential of the SCRO circuit as a sensor for the biohybrid
chamber. In particular, the experiment with the 21% concentra- system, capable of accurately monitoring and predicting yeast
tion had to be halted prematurely at around 8.5 min, as the yeast behavior and its influence on actuation power.
foam inside the actuation chamber exceeded its capacity, affect-
ing the experimental setup. The relationship between pressure Yeast-powered biohybrid soft robotics characteristics
change and frequency is depicted in Fig. 3D, exhibiting similar Yeast-driven inflatable membrane actuator: Volume and force
behavior to the frequency change over time shown in Fig. 3A. This experiment investigated actuation metrics in a simple soft
This observation indicates that frequency can effectively detect actuator, specifically an inflatable membrane actuator from

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 6

 Curve fitting (3% averaged data) 10-3 Curve fitting (6% averaged data)
0.02
10
Experimental data
Fitted curve
8
0.015

df/dt (kHz/s)
6
df/dt (kHz/s)

0.01
4
df/dt = 0.008 · exp(0.89 · P) – 0.006 · exp(–13.88 · P) df/dt = 316.9 · exp(–4.268 · P) – 316.9 · exp(–4.268 · P)
R 2: 0.9792 2 R 2: 0.9177
0.005 RMSE: 0.0005086 RMSE: 0.0005652
0

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Pressure (bar) Pressure (bar)

Curve fitting (12% averaged data) Curve fitting (21% averaged data)
0.01 0.01

0.008 0.008

Downloaded from https://spj.science.org on May 07, 2025

df/dt (kHz/s)

df/dt (kHz/s)
0.006 0.006

0.004 0.004
df/dt = 12.26 · exp(–5.6 · P ) – 12.26 · exp(–5.7 · P )

df/dt = 0.04 · exp(–4.003 · P) – 0.03811 · exp(–6.997 · P) R 2: 0.9401

0.002 R 2: 0.9373 0.002 RMSE: 0.0004269
RMSE: 0.0006014

0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45
Pressure (bar) Pressure (bar)

Fig. 4. Mathematical model of pressure vs. rate of change in impedance frequency at different sugar concentrations. The equations describe the relationship between the
averaged data over 5 trials of the rate of change in frequency versus the change in pressure for each concentration. These equations relate the growth rate of yeast as a
biological actuation element to the pressure output for actuation power using the rate of change in impedance frequency as a sensing part. RMSE, root mean square error.

Jones and Damian [36]. This actuator, made from silicone rub- solution. While the actuator was primarily inflated by carbon
ber (Ecoflex 00-50), features an inner diameter of 8 mm and a dioxide, the process led to a foamlike mixture with the yeast
membrane thickness of 1 mm. The inflatable membrane actua- solution. This phenomenon stemmed from the mechanics of
tor replaced the actuator chamber from the previous pressure the inflation process, where gas bubbles within the yeast solu-
experiment and is connected to the yeast chamber, as shown tion attempted to traverse into the actuator from the yeast
in Fig. 5A. Two key experiments were conducted. The first chamber.
experiment focused on measuring the peak position of the The subsequent experiment measured the force exerted by
actuator across varying sugar concentrations (3%, 12%, and the inflatable membrane actuator, specifically at its tip. This
21%) within the yeast solution under no-load conditions. The experiment was realized by firmly affixing the inflatable mem-
peak position was measured through recorded videos by a brane actuator at 5 mm from the strain gauge (strain gauge load
vision tracker system, as illustrated in Fig. 5B. The peak position cell) to allow the actuator to inflate. The strain gauge served a
was monitored using the Tracker video analysis and modeling dual purpose by measuring the force and functioning as a
tool software [37]. This peak position measurement reflects the motion constraint for the actuator, as shown in Fig. 5C. This
actuator’s volume alteration in response to yeast gas effects, configuration ensured that the actuator’s displacement did not
albeit considering the complexities arising from the actuator’s exceed the designated limit, thus maintaining the integrity of
hyperelastic properties that impede straightforward volume the experimental setup. The observed force reached a range of
measurements. 10 N, and its rate of change demonstrated escalation with the
As illustrated in Fig. 5D, the experimental results of 3 trials increase in sugar concentration from 3% to 12%. Nonetheless,
demonstrate a direct correlation between increased sugar con- assessing the force under a 21% sugar concentration posed a
centration and enhanced gas production, amplifying the actua- challenge. This was attributed to the instability of the power
tor’s peak position from 3% to 12%. However, the peak position generated at this concentration level. Notably, the force read-
decreased from 12% to 21%, confirming the pressure experi- ings at 21% were lower than those at 12%, as shown in Fig. 5E.
ment’s findings. The inflation dynamics of the actuator revealed Consequently, this interaction resulted in a portion of the yeast
an intriguing interplay between carbon dioxide and the yeast solution finding its way inside the actuator. The fully inflated

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 7

 A B

Soft membrane Peak position

F motion
Yeast Origin point
Yeast chamber
YGrowth CCell
Yeast chamber
Oscillator circuit
CCell FYeast
C
Strain gauge

Inflatable membrane

Downloaded from https://spj.science.org on May 07, 2025

D E
Peak position of inflatable with ANOVA P values Force exerted from inflatable membrane with ANOVA P values
30 30

* *
25 * 25
*
* 20
*
20
Y (mm)
Y (mm)

15
15

10
10
5
5
0
3% 12% 21% 3% 12% 21%
Sugar concentrations Sugar concentration

Fig. 5. Soft robotic system: The inflatable membrane with yeast fermentation changes the capacitance of cell walls, which produces gas, altering the impedance frequency
and resulting in the dynamic motion of the soft membrane. (A) The actuator design with an experimental setup was used to test yeast actuation power at different sugar
concentrations. (B) The inflatable membrane actuator’s peak position changes due to pressure variations in the experimental setup. (C) The output force of the inflatable
membrane actuator is exerted on the strain gauge fixed 5 mm above the actuator experimental setup. (D) Peak position of the inflatable membrane actuator under different
sugar concentrations of yeast solution. (E) Force exerted by the inflatable membrane actuator under different sugar concentrations of yeast solution. * indicates a significant
change between the 3 sugar concentrations using the analysis of variance (ANOVA) test (P value <0.05).

actuator showcased a distinct configuration, with gas entrapped inflated the membrane actuator for approximately 15 min. A
over a foamlike liquid yeast layer. This unique observation 10-ml yeast mixture with a consistent 12% sugar concentration
accentuates the complexity of the yeast-based actuation system, was employed throughout the experiment due to its demon-
where multiple factors, including gas diffusion and fluid dynam- strated stability in previous experiments, showing reasonable
ics, contribute to the overall behavior of the actuator. pressure, volume, and force characteristics.
Observations revealed that yeast could initiate inflation for
Regeneration of actuation power in yeast-driven approximately 3 cycles across 3 trials, as depicted in Fig. 6A
inflated membrane actuators with observing the frequency change over time shown in Fig.
This experiment investigated the possibility of regenerating the 6B. In the first cycle, inflation reached around 9 mm, while in
actuation power without additional yeast or sugar. A peristaltic the second cycle, inflation ranged between 12 and 14 mm,
pump was integrated as a reverse system for the inflated mem- attributed to the lag phase of yeast growth. Notably, the actua-
brane actuator from Fig. 5 to pump foam outside the actuator. tor’s peak positions were more consistent across cycles in trial
This setup facilitated the eviction of gas and foam from the actua- 1, allowing for more cycles than in trial 3, where higher peak
tor after yeast growth. The pump was programmed to deflate the positions in cycle 2 made it more difficult to maintain consis-
actuator for 1 s whenever the frequency rate of change deviated tent actuation power. However, after these initial cycles, the
by 1 kHz, a threshold determined from the observations in pump effectively removed all yeast from the actuator, halting
the pressure experiment shown in Fig. 3. Meanwhile, the yeast further inflation. Furthermore, each successive cycle gradually

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 8

 A Peak position vs. time
15
Trial 1 Trial 2 Trial 3 Average

10
Y (mm)

0
0 100 200 300 400 500 600 700 800 900 1,000
Time (s)

B Frequency vs. time

Downloaded from https://spj.science.org on May 07, 2025

6
Frequency (Hz)

Trial 1 Trial 2 Trial 3 Average

2
0 100 200 300 400 500 600 700 800 900 1,000
Time (s)
Fig. 6. Assessing the regeneration of a 10-ml yeast solution with a 12% sugar concentration for actuating and peristaltic pump deactivating the inflated membrane actuator.
(A) Cycles of the peak position of the inflatable membrane over time for 3 trials, averaging 3 cycles over approximately 16 min. (B) Impedance frequency readings of the yeast
inside the chamber during the experiment.

diminished the yeast’s ability to reinflate the actuator. This with 3 air pillow geometries featuring curved oval shapes with
decline can be attributed to the peristaltic pump’s removal of a width of 5 mm each. These air pillows were arranged in a
yeast and the resulting decrease in yeast density, evidenced curved configuration to achieve approximately a 180° rotation,
by a reduction in solution mass. Residual foam left from the with the geometry selected based on the familiar shape of air
pump’s operation contributed to this change in yeast density. pillows. The soft limbs are fabricated using Ecoflex 00-50 mate-
Maintaining yeast-driven actuation over multiple cycles holds rial, while the foot and holder parts are made of Mold Star
significant promise for advancing this biohybrid actuation 15 material to attach the limb to the yeast chamber shown
mechanism. in Fig. 7C. The second system consists of an inflatable mem-
brane actuator with a yeast chamber made of Mold Star 30,
Results which is a stiffer material to prevent chamber inflation, featur-
ing an 8-mm diameter to accommodate a 2-mm membrane for
The practicality of utilizing yeast for actuation and sensing in actuation as shown in Fig. 5A. The yeast chamber in the inflated
real soft robotics systems was examined using a soft limb and membrane actuator is equipped with 2 electrodes connected
an inflatable membrane actuator. The underlying concept to the SCRO impedance sensor. The manufacture and molds
aimed to correlate impedance frequency with the motion of of the 2 actuators are shown in Fig. S6.
these soft robotics systems and establish a relationship between
yeast growth, as indicated by frequency, and robot motion. In Yeast-powered biohybrid flexion of a soft limb
both setups, experiments utilized a 12% sugar concentration The yeast-based actuation was investigated in a soft flexion
yeast mixture. The first system comprises a soft limb designed actuator under different temperatures. The work envelope,

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 9

 A B C

Hot
water bath

Yeast chamber

Downloaded from https://spj.science.org on May 07, 2025

with SCRO

D E
Soft limb actuator motion with impedance
Soft limb actuator impedance frequency
140 4.5

120 4
Deflection angle (°)

Frequency (kHz)

100 3.5

80 3

60 2.5
20 °C room temp.
40 °C water bath
40 2
60 °C water bath
20 °C room temp. 60 °C water bath
80 °C water bath
20 40 °C water bath 80 °C water bath
1.5
1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 0 50 100 150 200 250 300
Frequency (kHz) Time (s)

F G Actual vs. predicted deflection angles (GPR)

Soft limb actuator dynamics
140
150
120
Deflection angle (°)
Deflection angle (°)

100
100

80 50
Actual
60 0 Predicted
20 °C room temp.
40 °C water bath 4
40 3.5 80
60 °C water bath 70
3 60
80 °C water bath 2.5 50
40
20 2 30
0 50 100 150 200 250 300 20
Frequency (Hz) Temperature (°C)
Time (s)

Fig. 7. Soft robotic systems: soft limb actuator. (A) Soft limb actuator prior to inflation, showing the deflection angle (𝜃) and impedance frequency inside the yeast chamber.
(B) Path of the soft limb actuator after inflation, highlighting the increase in impedance frequency. (C) Design of the soft limb actuator and experimental setup for measuring
the actuator path at different temperatures using a water bath. (D) Deflection angle versus corresponding yeast frequency at various temperatures, with bars representing
the standard deviation over 3 trials for each temperature. (E) Changes in frequency over time during limb motion. (F) Change in deflection over time. (G) Relationship between
yeast temperature and the soft limb dynamics with yeast impedance using Gaussian process regression (GPR).

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 10

depicting the utmost extent of motion achievable by the actua- variations are intrinsic to bio-driven systems and highlight the
tor through yeast-powered actuation, is graphically illustrated importance of conducting multiple trials to capture the system’s
in Fig. 7. This envelope, determined by utilizing the imped- dynamic behavior. Key variables measured included the applied
ance frequency of yeast as an observation parameter for yeast force, the resulting pressure within the actuator, and the fre-
growth rate as a proprioceptive sensor, offers insights into the quency response of the actuator at equilibrium. The actuator
actuator’s maximum coordinate points. Notably, the actuator reached an equilibrium phase when the pressure reached
demonstrates an impressive deformability angle range span- 0.1 bar, and the actuator inflated to approximately 0.5 cm. The
ning from 20° to 140° in 5 min. The limb’s work envelope was experiment used a controlled linear actuator attached to a strain
monitored using the Tracker video analysis and modeling tool gauge to apply a controlled force signal over the inflated mem-
software [37] as shown in Fig. 7A and B. brane actuator filled with yeast and connected to the SCRO
The impact of varying medium temperatures on the actua- circuit and pressure sensor, as shown in Fig. 8A. We used fast
tor’s behavior was thoroughly investigated, as shown in Fig. 7D Fourier transform (FFT) analysis on each trial’s recorded force,
and F. The yeast actuator limb was mounted on a yeast chamber pressure, and frequency signals. The FFT results revealed dis-
in hot water bath at temperatures of 40, 60, and 80 °C, com- tinct peaks corresponding to the frequency of the applied sine-
pared to room temperature at 20 °C, illustrated in Fig. 7C. The wave force and its harmonics, as shown in Fig. S8.
alteration in temperature shows a substantial influence on the To investigate the frequency characteristics of the actuator
actuator’s performance. A noticeable trend emerged as the tem- response, we performed a cross-correlation analysis to examine
perature increased from 20 to 40 to 60 °C, causing a significant the relationship between the applied force and the resulting
acceleration in the actuator’s speed of motion, as shown in Fig. pressure and frequency responses of the actuator. Cross-
7F, where it took 300 s to complete the path at 20 °C and only correlation measures the similarity between 2 signals as a func-
150 s at 60 °C. However, the dynamics shifted when the actuator tion of the time lag between them, providing insights into the
was subjected to an 80 °C environment, where it took 200 s to temporal relationship and phase shift between the signals. By
complete the path. In this case, the exceptionally high tempera- calculating the cross-correlation between the applied force and

Downloaded from https://spj.science.org on May 07, 2025

ture had the opposite effect, slowing down the motion of the the pressure, as well as between the applied force and the fre-
actuator. This deceleration can be attributed to the tempera- quency, we aimed to understand the time delays and phase
ture’s adverse impact on yeast metabolism, impeding natural differences in the actuator response. The cross-correlation
processes. The initial temperature-induced acceleration is likely results revealed significant positive correlations between the
due to the heightened metabolic activity of yeast in response applied force and the pressure response and between the
to elevated thermal conditions, which can be sensed with applied force and the frequency response, with distinct peaks
impedance, as shown in Fig. 7D and E. These plots demonstrate at specific time lags in Fig. 8. These peaks indicate the time
how temperature variations impact yeast impedance, affecting delays at which the pressure and frequency responses are
the actuator’s performance, which is sensed by the bioimped- most similar to the applied force. These time lags suggest that
ance SCRO circuit. the actuator’s response is not instantaneous but follows the
Complementing this analysis, the relationship between applied force with a certain delay, probably due to the yeast
actuator dynamics and yeast temperature can be utilized to cells’ dynamics and the SCRO delay.
control robot motion in real-time applications by using yeast In particular, pressure and force signals are directly propor-
impedance as a sensor. To achieve this, we modeled the rela- tional. At the same time, yeast frequency has an inverse relation
tionship between temperature (as a control signal) and actuator with both pressure and force, decreasing as pressure increases
deflection angle, with impedance frequency serving as the due to changes in the arrangement of yeast cells. These relations
sensing element, using Gaussian process regression, as shown for 4 trials are represented in Fig. 9A, and the comparison
in Fig. 7G. The model explains 83% of the deflection angle between the amplitude of the time series is shown in Fig. 9B
variance based on the temperature and frequency data, indicat- for 1 trial. The plot presented in Fig. 9C reveals the sensitivity
ing a good fit and suggesting that the Gaussian process regres- of force and pressure to changes in frequency as indicated by
sion model effectively captures most of the relationship between the slope of the curves. The initial changes in frequency do not
these variables. The model’s validity was confirmed by predict- correspond to changes in force and show only a slight change
ing the actuator’s angles on untrained data, as shown in Fig. S7. in pressure, representing the equilibrium phase where the
membrane is allowed to deflect. The first trial displays the short-
Yeast as a tactile sensor est curve due to higher noise levels than the other 3 trials. The
In this experiment, we investigated the dynamic response of plot demonstrates that force and pressure exhibit variable sen-
a yeast-inflated membrane actuator subjected to a sine-wave sitivity across different frequencies. The minimum detectable
force. The inflated membrane actuator was filled with 3 ml of changes in force and pressure are illustrated in Fig. 9D. The
12% sugar concentration yeast and consequently identified its resolution of the sensor system demonstrates the SCRO circuit’s
sensing capability. The primary objective was to observe how capability to detect variations in the actuator’s environment.
the pressure and frequency of the actuator changed in response We calculate the resolution in force and pressure by determin-
to the applied force. The experiment was conducted over 4 trials, ing the smallest nonzero difference in these 2 parameters,
each capturing time-varying force, pressure, and frequency data resulting in a measurable change in frequency. Notably, the
of the actuator. Each trial consisted of 15 cycles of sine-wave maximum resolution for force was 7 × 10−5 N, while for pres-
force application with a maximum force of 4.5 N, and data were sure, it was 9 × 10−6 bar. The data for each trial are filtered using
recorded at uniform time intervals. The applied stimulus (sine- a Butterworth filter and then smoothed using a moving average.
wave force) was identical across the 4 trials. However, biological The comparative analysis of sensitivity and resolution provides
variability in yeast activity and environmental conditions led valuable insights into how effectively frequency variations can
to differences in the pressure and frequency responses. These be used to sense changes in force and pressure.

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 11

 A B

Cross-correlation between force and pressure

1
Trial 1
Trial 2

Correlation coefficient
0.8 Trial 3
Trial 4
0.6
Probe
0.4
Yeast
sensor 0.2

0
-1,500 -1,000 -500 0 500 1,000 1,500
Lag (s)

C D
Cross-correlation between force and frequency Cross-correlation between pressure and frequency
1
1
Trial 1
Trial 1
Trial 2
Trial 2
Correlation coefficient

0.8 0.8 Trial 3

Downloaded from https://spj.science.org on May 07, 2025

Trial 3 Correlation coefficient
Trial 4
Trial 4
0.6 0.6

0.4 0.4

0.2 0.2

0
0
-1,500 -1,000 -500 0 500 1,000 1,500 -1,500 -1,000 -500 0 500 1,000 1,500
Lag (s) Lag (s)

Fig. 8. Tactile sensing using yeast inside the inflatable membrane. (A) Experiment setup for data acquisition, featuring yeast injected into the inflatable membrane actuator,
equipped with SCRO electrodes, where a linear actuator with a strain gauge applies a sinusoidal force wave after the actuator’s inflation. (B) Cross-correlation between the input
force sine signal and the output pressure across 4 trials shows a positive correlation at zero lag, while peaks indicate the time lag between the 2 signals. (C) Cross-correlation
between the input force sine signal and yeast impedance frequency as a sensing variable across 4 trials, with a positive correlation at zero lag and peaks representing time
lags. (D) Cross-correlation between the pressure sine signal and output frequency across 4 trials, demonstrating a positive correlation at zero lag.

Despite the inverse relationship between pressure and fre- of a yeast-driven inflatable membrane actuator to differentiate
quency, the cross-correlation may yield positive values as shown between tissue types by performing a palpation task. We used
in Fig. 8D. This can occur due to shared dynamics as both pres- the previously tested inflated membrane actuator. We increased
sure and frequency respond to the same input, which is force, the yeast chamber to 15 mm, allowing it to be filled with
leading to the same temporal patterns, even if their magnitude approximately 2.5 ml of 12% yeast concentration, as shown in
moves in opposite directions. In addition, the positive cross- Fig. 10E, to create a yeast-driven probe. To simulate differences
correlation indicates that while pressure and frequency may in tissue stiffness, a skin phantom was created using 5% concen-
exhibit an inverse relationship with respect to their values, their tration gelatin, representing healthy tissue, and silicone rubber
temporal patterns of change are aligned in specific contexts. (Ecoflex 00-50), representing cancerous tissue [38,39]. The actua-
These findings highlight the potential of using yeast bioimped- tor was filled with yeast and allowed to ferment for 13 min until
ance as an exteroceptive sensor to detect external changes in the internal pressure stabilized at 0.12 bar. This pressure was
yeast-driven robots. The yeast-driven inflatable membrane’s abil- maintained consistently throughout the experiment, optimizing
ity to detect changes in volume and pressure through bioimped- the actuator’s sensitivity to variations in tissue stiffness.
ance frequency allows the robot to sense when it has encountered The experimental protocol began with inflating the yeast-
an obstacle, facilitating effective control and avoidance. driven actuator until it reached a stable internal pressure of 0.12
bar. A linear actuator was then used to move the inflated mem-
Yeast-driven systems for tissue palpation brane toward the skin phantom surface until a change in load
and gripping applications was detected, indicating contact with the phantom. Subsequently,
This section presents yeast-driven systems’ capabilities in bio- the actuator was displaced 4 mm into the phantom to simulate
hybrid robotics, focusing on tissue palpation and object grasp- tissue palpation, ensuring consistent penetration across materials
ing. Using yeast fermentation for actuation and sensing, the for comparison. For 3 cycles, the probe was pressed 4 mm into
experiments show how these systems can detect differences in the phantom, which was less than the inflated membrane height.
tissue stiffness and manage objects of varying properties, illus- At the beginning of the experiment, the internal pressure of the
trating their practical potential for soft robotics and biomedical actuator was consistently maintained at 0.12 bar. This process
applications. In this experiment, we investigated the capability was repeated until the actuator’s internal pressure increased from

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 12

 A Sine-wave force applied to actuator
B
5
Force (N)

3
5

0 Pressure (bar)
0 100 200 300 400 500 600 700 800 Frequency (kHz)
4 Force (N)
Time (s)
Pressure response

Scaled amplitude
Pressure (bar)

0.2 3
3

Amplitude
0.1
2 2
0
0 100 200 300 400 500 600 700 800
Time (s) 1
1
Frequency response
Frequency (kHz)

5 0
400 410 420 430 440 450 460 470 480 490 500
Time (s)
4 0
0 100 200 300 400 500
Trial 1 Trial 2 Trial 3 Trial 4
3
0 100 200 300 400 500 600 700 800
Time (s)

C D
Sensitivity of force and pressure to frequency changes 10 -5 Resolution of force and pressure
0.1 7
Change in force

Min. detectable force change (N)

Minimum detectable change

0.05 6 Min. detectable pressure change (bar)
(N)

0 Trial 1 Trial 2 5
Trial 3 Trial 4
-0.05 4
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
Change in frequency (kHz) 3
Change in pressure

10-3
10
2

Downloaded from https://spj.science.org on May 07, 2025

(bar)

5 1
Trial 1 Trial 2
Trial 3 Trial 4 0
0 1 2 3 4
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Trial
Change in frequency (kHz)

Fig. 9. Effect of pressure change on yeast frequency due to external force. (A) Time-series data of the applied sine-wave force, pressure, and frequency response for 4 trials
after the pressure inside the actuator reaches a specified value. (B) A comparison of the scale magnitude of the 3 signals from 1 trial illustrates the inverse relationship between
force and pressure and the corresponding impedance frequency. (C) The sensitivity of force and pressure to changes in frequency. (D) Resolution: minimum detectable changes
in force and pressure for frequency change for each trial.

0.12 to 0.13 bar due to active yeast fermentation. During the during object manipulation. The gripper fingers filled with
repeated measurements across different materials within the approximately 1.5 ml of 12% yeast mixture, as shown in Fig.
phantom, the gradual pressure increase from 0.12 to 0.13 bar did 10J, were tested by gripping objects made of gelatin and silicone
not significantly affect the sensitivity of the actuator. However, with identical weight and thickness, as illustrated in Fig. S9.
this pressure change is crucial for maintaining the yeast fermen- The gripper closed fully and grabbed a piece of paper after
tation-driven probe and ensuring consistent system performance around 10 min of yeast-driven actuation Fig. 10J. The analysis
throughout the experiment. revealed distinct gripping phases, marked by peaks in the
The key parameters measured during the experiment included deflection angle, as shown in the path profile of the gripper
the force exerted on the tissue phantom and the corresponding fingers in Fig. 10G. These peaks correspond to mechanical
percentage change in impedance, which served as a sensing met- transitions during object engagement, highlighting the grip-
ric, as shown in Fig. 10. The results demonstrated a clear distinc- per’s sensitivity to grasping. Simultaneously, changes in imped-
tion between the tissue types: gelatin exhibited a minor percentage ance during gripping, as shown in Fig. 10I, provide insights
change in impedance compared to silicone, reflecting the differ- into the gripper’s response to deflection changes and its adapt-
ences in stiffness between the 2 materials. For example, the mean ability to the physical properties of the objects being handled.
applied force during palpation was higher for silicone tissues Time-stamped photos of the gripper holding objects in Fig.
compared to gelatin, as shown in Fig. 10A, with gelatin requiring 10H link these behavioral changes to specific operation phases,
lower force due to its softer composition. Similarly, the percentage offering a synchronized view of impedance dynamics, deflec-
change in impedance was more pronounced in silicone tissues, tion angle changes, and mechanical performance.
demonstrating the probe’s sensitivity to stiffness differences, as To evaluate the accuracy and response time of the driving
illustrated in Fig. 10B. The comparison of force and impedance system, the mean rate of change for the deflection angle was
measurements between gelatin and silicone tissues during calculated as the derivative of the deflection angle with respect
pressing further emphasized the actuator’s ability to distinguish to time. The results indicate that the right finger exhibited a
between the 2, as shown in Fig. 10D. These findings highlight faster rate of change (0.51136/s). In comparison, the left finger
the effectiveness of the yeast-driven actuator in detecting and demonstrated a slower rate of change (−0.15096/s), suggesting
differentiating tissue stiffness variations. The experiment setup, different mechanical responses and yeast fermentation between
depicted in Fig. 10C, shows the inflated membrane actuator press- the 2 fingers. Similarly, the mean rate of impedance change,
ing on the skin phantom. determined as the derivative of impedance over time, was
Second, we designed the yeast-driven gripper system using higher for the right finger (0.019 Hz/s) than for the left finger
the soft limb actuator as gripper fingers. This system demon- (0.01212 Hz/s), reflecting faster impedance dynamics in the
strated variations in deflection angles and changes in impedance right finger. These metrics were derived from the time-series

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 13

 A Force by tissue type F Impedance percentage change Deflection angle percentage change
by object type by object type

Impedance frequency change (%)

1.5 200 10

Deflection angle change (%)

5
Force (N)

1 150
0

0.5 100 -5

-10
0 50
Big round Gelatin Small round Square
silicone silicone silicone -15
Right finger
Tissue type
Left finger
0 -20
B Percentage change in impedance by tissue type
Gelatin Paper Silicone Gelatin Paper Silicone
Object type Object type
Percentage change
in impedance (%)

0
G 130
Deflection angle over time

-10 120

Deflection angle (°)

110
-20 100

90
Big round Small round Square
Gelatin
silicone silicone silicone 80
Tissue type 70

Downloaded from https://spj.science.org on May 07, 2025

60
C Right finger
Left finger
50
0 100 200 300 400 500 600
Time (s)

t = 323.75 s t = 363.66 s t = 387.73 s t = 417.11 s t = 442.73 s t = 447.99 s

I Impedance frequency over time

D Force at position 4 at Difference of impedance at
6
pressing 4 mm inside the skin pressing 4 mm inside the skin
1.4
1.4 5
Frequency (kHz)

1.2
1.2
Change of impedance (kHz)

1 4
1

0.8
Force (N)

0.8 3
0.6
0.6
2
0.4
0.4 Right finger
Left finger
0.2 0.2 1
0 100 200 300 400 500 600
0 Time (s)
Gelatin Silicone Gelatin Silicone
Tissue type Tissue type

J
E

15 mm h

Fig. 10. Using the yeast-driven probe and gripper. (A) Mean applied force across different tissue types during palpation. Error bars represent the standard deviation of
measurements. (B) Percentage change in impedance across different tissue types during palpation. Error bars represent the standard deviation of measurements. (C) The
experiment setup shows an inflated membrane pressing on the skin. (D) The difference between gelatin and silicone force and impedance at pressing. Error bars represent
the standard deviation of measurements. (E) Yeast-driven probe size. (F) Percentage change in deflection angle and its effect on impedance change in a yeast-driven gripper
system gripping different objects. (G) Path profile of the gripper fingers highlighting gripping phases as peaks in the deflection angle. (H) Time-stamped photos of the yeast-
driven gripper holding objects, reflected in changes in impedance shown in (H) and deflection angles during gripping phases in (G). (I) Impedance changes correspond to
the gripping phases for various objects (gelatin, silicone, and paper). (J) Comparison of the yeast-driven gripper to a 1-pound coin and fully closed gripper filled with yeast.

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 14

data of the gripper’s deflection angles and impedance, with advantages like biodegradability and untethered operation,
baseline values established during the “none” state, where the eliminating the need for external valves and pumps. At the same
gripper was not holding any object. time, its dual functionality as both an actuator and a sensor
This comprehensive analysis highlights the dynamic behav- further enhances its versatility in robotic systems.
ior of the yeast-driven gripper, showcasing its responsiveness This paper explores using 2 types of soft actuators—a soft
and mechanical sensitivity during object manipulation. While limb and an inflatable membrane—to test the capability of yeast
the response speed is relatively slow, the system’s response to actuate robots while sensing both internal and external states
enables it to sense and grasp various objects. These findings using bioimpedance. The experiments with soft limb designs
underscore the potential of yeast-driven grippers for advanced revealed a complex interaction between yeast fermentation and
biohybrid robotic applications requiring careful handling and robot motion, particularly regarding deflection angle under
high sensitivity. varying temperatures. By correlating impedance frequency
with robot movement, we established a method for monitoring
yeast dynamics and predicting actuator behavior as a proprio-
Discussion ceptive sensor. The results indicate that temperature variations
This study explored the feasibility of using yeast as an actuation significantly affect the acceleration and deceleration of limb
and sensing mechanism in soft robotics. We investigated the motion, suggesting that temperature can effectively control
potential of yeast fermentation to generate mechanical power yeast actuation. Our findings also demonstrate that yeast
through pressure, which can be sensed using yeast bioimped- impedance frequency is a reliable indicator of yeast state, pro-
ance. Through experiments with inflated membrane actuators viding valuable insights into actuator performance and dynam-
and soft limb designs, we demonstrated the ability of yeast- ics. In our experiments involving an inflatable membrane
driven systems to function as actuators, proprioceptive sensors, actuator without a separate yeast chamber, we gained insights
and exteroceptive sensors. Our results indicate that yeast fer- into the practical application of yeast-driven actuation by incor-
mentation and bioimpedance technology can effectively power porating yeast directly into the actuator. The inflatable mem-

Downloaded from https://spj.science.org on May 07, 2025

and monitor microbial biohybrid robotic systems. Additionally, brane actuator effectively responded to external forces, with
our research highlights the development of a circuit capable of noticeable changes in pressure and frequency. Cross-correlation
measuring impedance in biological matter and adjusting by analysis showed significant positive correlations between applied
using a digital potentiometer, emphasizing its sensitivity to force, pressure, and frequency responses. While pressure and
variations in yeast concentration. force were directly proportional, frequency showed an inverse
First, the study investigated the potential of yeast as a source relationship with both. FFT analysis confirmed distinct peaks
of mechanical power and the ability of a bioimpedance circuit corresponding to the applied force frequency and its harmon-
to detect pressure changes. The experiments demonstrated that ics. These findings underscore the potential of using yeast bio-
yeast can generate sufficient pressure for actuation, with pres- impedance as an exteroceptive sensor to detect external changes
sure levels directly correlating with yeast impedance at a fixed in biohybrid robotic systems.
volume. These findings underscore the potential for dynamic, We demonstrated the potential of yeast-driven biohybrid
real-time tracking of yeast-driven actuation, enabling timely robots through 2 applications: a tissue stiffness probe and a
adjustments for optimized fermentation performance. By inte- gripper. These demonstrations highlight the sensitivity and
grating yeast dynamics equations with SCRO circuit equations, capability of yeast fermentation for actuation and sensing,
we successfully modeled yeast growth rate and actuation power, particularly in handling delicate tasks. Despite a relatively
paving the way for enhanced sensing and actuation capabilities slower response speed than conventional methods, the inher-
in biohybrid systems. Additionally, we observed a direct rela- ent sensitivity and dual functionality of yeast-driven systems
tionship between sugar concentration, actuator peak position, underscore their potential for soft robotics and biomedical
and the exerted force of the soft inflatable membrane actuator, applications demanding delicate handling and precise sensing
indicating the influence of fermentation dynamics on actuator capabilities. If miniaturized, such systems could be effec-
behavior. Furthermore, integrating a peristaltic pump allowed tively applied in biomedical contexts, including tissue evalu-
for the regeneration of actuation power and showed the poten- ation and precise object manipulation, with the advantage
tial for sustained yeast-driven actuation over multiple cycles. that yeast-driven systems could potentially be ingestible and
Although we tested the effects of temperature and sugar con- biodegradable.
centration, humidity was not examined due to the closed nature Understanding the factors that influence yeast-driven actua-
of the yeast chamber, which maintains anaerobic fermentation tion is crucial for optimizing performance. One key factor is the
conditions where humidity is not a significant factor. The yeast- design and configuration of the electrodes used for bioimped-
driven system exhibits a slower response speed than traditional ance sensing. We observed that the length and material of the
actuation methods due to the gradual nature of fermentation. electrodes and the amount of yeast mixture affected the reading
However, this characteristic makes it ideal for applications frequency. Adding a digital potentiometer to the SCRO circuit
requiring adaptive and energy-efficient responses, such as envi- to address these variations proved effective. More significant
ronmental monitoring and biomedical systems. For example, amounts of yeast mixture need a higher value of the digital
it can provide long-term sensing of parameters like pressure, potentiometer, requiring adjustments in circuit parameters to
temperature, or gas composition in dynamic environments maintain oscillation frequency. Conversely, smaller amounts
or enable controlled, gentle movements for handling delicate necessitate changes to achieve optimal circuit performance.
objects like biological tissue in soft robotics. In medical applica- Additionally, the distance between electrodes is crucial for
tions, it shows potential for adaptive prosthetics, tactile sensing accurately capturing yeast dynamics, influencing the sensitivity
for soft tissue evaluation, implants, and biosensing platforms and resolution of the sensing system; therefore, we fixed the
to monitor body conditions over time. Additionally, yeast offers electrodes in the same position during experiments. Our findings

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 15

 Table 3. Comparison of biohybrid robot systems

Feature Yeast-driven Cell-driven [17] Bacteria-driven [11]

Driving mechanism CO2 production via fermentation Contraction of muscle cells Flagellar motion or metabolic
activity
Control difficulty Moderate (temperature, sugar High (requires precise stimulus Moderate (chemical or
regulation) control) light stimuli)
Scalability High (low-cost and easy to culture) Low (complex cell culture Moderate (scalable but
requirements) contamination risk)
Response speed Slow Moderate to fast Fast
Environment Actuation, sensing, environmental High-precision biohybrid robotics Microrobots, therapeutic delivery
tasks, biomedical uses inside
the body
Ability to operate Works inside the body and outside Typically limited to controlled lab Works outside the lab and requires
the lab environments extensive genetic modification to
ensure safety in biomedical
application
Limitations Slow response speed High cost, complex control Limited force generation,
contamination risk

Downloaded from https://spj.science.org on May 07, 2025

also revealed that yeast actuation is driven by carbon dioxide study investigates the feasibility of using yeast for actuation
production and the flow of yeast foam into the actuator and sensing in soft robotics, demonstrating that yeast fermen-
chamber. This foam formation, resulting from gas bubbles tation can generate mechanical power and be sensed through
within the yeast solution interacting with the actuator’s mem- bioimpedance. We used yeast-driven systems as actuators
brane, contributes to the actuator’s inflation and movement. and sensors, showing their potential in proprioceptive and
While yeast-driven actuation demonstrates significant poten- exteroceptive applications. Experiments with inflated mem-
tial, several challenges and limitations must be addressed to opti- brane actuators and soft limbs revealed that yeast-generated
mize its use. For example, the high growth rate of yeast can cause pressure correlates with impedance changes, while tempera-
rapid fluctuations in frequency readings and actuator perfor- ture variations influence actuator performance. Additionally,
mance, complicating precise control and long-term operation of incorporating yeast directly into the actuator and using bio-
yeast-driven robots. One critical area for future research is the impedance for real-time monitoring highlighted the effec-
deactivation process of yeast-driven actuation. Although our tiveness of this approach for detecting external changes
study used a peristaltic pump for deactivation, developing a more and optimizing robot control. Our findings pave the way
efficient system without electronics is possible. Introducing a self- for enhanced biohybrid systems with integrated sensing and
release valve made from soft materials could facilitate controlled actuation capabilities.
deflation of the actuator, eliminating the need for electronic com-
ponents. This innovation could improve the efficiency of yeast-
driven robotic systems. Acknowledgments
Yeast-driven biohybrid systems offer unique advantages,
including scalability, cost-effectiveness, and straightforward We acknowledge Marco Pontin, Joanna Jones, Mostafa A.
handling, making them well-suited for in-body and external Mousa, and Mahmood Saleh’s valuable contributions. We
applications. The dual functionality of these systems, as both an acknowledge Prof. Ahmed G. Radwan’s assistance in bio-
actuator and a sensor, along with their biodegradability, high- impedance technology. We also appreciate the feedback
lights their potential as versatile options in biohybrid robotics. from Sheffield Biomedical Robotics Lab members.
Unlike cell-driven systems, which provide faster and more pre- Funding: The work is supported by 2 EPSRC DTP PhD schol-
cise actuation at a higher cost and complexity and are needed arships from the School of Electrical and Electronic Engineering
to work in lab medium, or bacteria-driven systems, which at the University of Sheffield.
excel in microscale applications but lack scalability and require Author contributions: M.S. conceived the idea, designed and
extensive genetic modification to ensure safety in biomedical conducted the experiments, wrote the manuscript, and ana-
applications, yeast systems simplify design and have minimal lyzed the result. F.F. assisted with the computer-aided design,
operational demands, enabling them to work both inside the contributed to the experimental design, conducted the experi-
body and outside the laboratory. This comparison underscores ments, and reviewed the manuscript. D.D.D. designed the stud-
the complementary roles of yeast-driven systems in expanding ies and reviewed the results and the manuscript. All authors
the capabilities of biohybrid robotics (see Table 3). contributed equally to the writing of the manuscript.
In conclusion, yeast-driven actuators hold immense poten- Competing interests: The authors declare that they have no
tial for powering and controlling soft robotics systems. This competing interests.

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 16

Data Availability of muscle-driven miniature robots with battery-free wireless
optoelectronics. Sci Robot. 2023;8(74):Article eadd1053.
All data are available from the corresponding authors upon 16. Filippi M, Balciunaite A, Georgopoulou A, Paniagua P,
reasonable request. Drescher F, Nie M, Takeuchi S, Clemens F, Katzschmann R.
Integrated closed-loop control of bio-actuation for
Supplementary Materials proprioceptive bio-hybrid robots. bioRxiv. 2024. https://www.
biorxiv.org/content/10.1101/2024.01.30.577987v1
Supplementary 1 17. Bawa M, Raman R. Taking control: Steering the future of
Figs. S1 to S9 biohybrid robots. Sci Robot. 2024;9(94):Article eadr9299.
Data File S1 18. Raman R, Bashir R. Biomimicry, biofabrication, and biohybrid
Movies S1 and S2 systems: The emergence and evolution of biological design.
Adv Healthc Mater. 2017;6(20):Article 1700496.
References 19. Hwang G, Paula AJ, Hunter EE, Liu Y, Babeer A, Karabucak B,
Stebe K, Kumar V, Steager E, Koo H. Catalytic antimicrobial
1. Chi Y, Zhao Y, Hong Y, Li Y, Yin J. A perspective on miniature robots for biofilm eradication. Sci Robot. 2019;4(29):
soft robotics: Actuation, fabrication, control, and applications. Article eaaw2388.
Adv Intell Syst. 2024;6(2):Article 2300063. 20. Kim D, Liu A, Diller E, Sitti M. Chemotactic steering
2. Yuan Z, Guo Q, Jin D, Zhang P, Yang W. Biohybrid soft of bacteria propelled microbeads. Biomed Microdevices.
robots powered by myocyte: Current progress and future 2012;14:1009–1017.
perspectives. Micromachines. 2023;14(8):Article 1643. 21. Tao X, Dai Nguyen T, Jin H, Tao R, Luo J, Yang X, Torun H,
3. Rossiter J. Soft robotics: The route to true robotic organisms. Zhou J, Huang S, Shi L, et al. 3D patterning/manipulating
Artif Life Robot. 2021;26(3):269–274. microparticles and yeast cells using ZnO/Si thin film surface
4. Zhou S, Li Y, Wang Q, Lyu Z. Integrated actuation and acoustic waves. Sensors Actuators B Chem. 2019;299:

Downloaded from https://spj.science.org on May 07, 2025

sensing: Toward intelligent soft robots. Cyborg Bionic Syst. Article 126991.
2024;5:Article 0105. 22. Crnković T, Bokor BJ, Lockwood ME, Cornish VW. Peptide
5. Zhang F, Li Z, Chen C, Luan H, Fang RH, Zhang L, variant detection by a living yeast biosensor via an epitope-
Wang J. Biohybrid microalgae robots: Design, fabrication, selective protease. BioDes Res. 2023;5:Article 0003.
materials, and applications. Adv Mater. 2024;36(3): 23. Webster-Wood VA, Guix M, Xu NW, Behkam B, Sato H,
Article e2303714. Sarkar D, Sanchez S, Shimizu M, Parker KK. Biohybrid robots:
6. Nguyen PQ, Courchesne N-MD, Duraj-Thatte A, Recent progress, challenges, and perspectives. Bioinspir
Praveschotinunt P, Joshi NS. Engineered living materials: Biomim. 2022;18(1):Article 015001 10.1088/1748-3190/ac9c3b.
Prospects and challenges for using biological systems 24. Pagan-Diaz GJ, Zhang X, Grant L, Kim Y, Aydin O,
to direct the assembly of smart materials. Adv Mater. Cvetkovic C, Ko E, Solomon E, Hollis J, Kong H, et al.
2018;30(19):Article e1704847. Simulation and fabrication of stronger, larger, and
7. Carlsen RW, Sitti M. Bio-hybrid cell-based actuators for faster walking biohybrid machines. Adv Funct Mater.
microsystems. Small. 2014;10(19):3831–3851. 2018;28(23):Article 1801145.
8. Li J, Dekanovsky L, Khezri B, Wu B, Zhou H, Sofer Z. 25. Xu N, Lenczewska O, Wieten S, Federico C, Dabiri J. Ethics
Biohybrid micro- and nanorobots for intelligent drug delivery. of biohybrid robotic jellyfish modification and invertebrate
Cyborg Bionic Syst. 2022;2022:Article 9824057. research. Bioinspirat Biomimet. 2020; https://philpapers.org/
9. Rivera-Tarazona LK, Campbell ZT, Ware TH. Stimuli- rec/XUEOBV.
responsive engineered living materials. Soft Matter. 26. Martinsen OG, Grimnes S. Bioimpedance and bioelectricity
2021;17(4):785–809. basics. Amsterdam (The Netherlands): Academic Press;
10. Filippi M, Yasa O, Kamm RD, Raman R, Katzschmann RK. 2011.
Will microfluidics enable functionally integrated biohybrid 27. Juansah J, Budiastra I, Dahlan K, Seminar K. Electrical
robots? Proc Natl Acad Sci USA. 2022;119(35):Article e2200741119. behavior of Garut citrus fruits during ripening changes in
11. Alapan Y, Yasa O, Yigit B, Yasa IC, Erkoc P, Sitti M. resistance and capacitance models of internal fruits.
Microrobotics and microorganisms: Biohybrid autonomous Int J History Eng Technol. 2012;12(4):1–8.
cellular robots. Ann Rev Contr Robot Autonom Syst. 28. Radwan AG, Elwakil AS, Soliman AM. Fractional-order
2019;2:205–230. sinusoidal oscillators: Design procedure and practical
12. Liu J, Jing X, Liu M, Li F, Li M, Li Q, Shi J, Li J, Wang L, examples. IEEE Trans on Circuits Syst I Reg Papers.
Mao X, et al. Mechano-fluorescence actuation in single 2008;55(7):2051–2063.
synaptic vesicles with a DNA framework nanomachine. 29. Saleh MA, Soliman M, Mousa MA, Elsamanty M,
Sci Robot. 2022;7(73):Article eabq5151. Radwan AG. Design and implementation of variable inclined
13. Ma S, Chen Y, Yang S, Liu S, Tang L, Li B, Li Y. The air pillow soft pneumatic actuator suitable for bioimpedance
autonomous pipeline navigation of a cockroach bio-robot with applications. Sensors Actuators A Phys. 2020;314:
enhanced walking stimuli. Cyborg Bionic Syst. 2023;4: Article 112272.
Article 0067. 30. Mousa MA, Soliman M, Saleh MA, Radwan AG. Tactile
14. Blackiston D, Lederer E, Kriegman S, Garnier S, Bongard J, sensing biohybrid soft E-skin based on bioimpedance using
Levin M. A cellular platform for the development of synthetic aloe vera pulp tissues. Sci Rep. 2021;11(1):Article 3054.
living machines. Sci Robot. 2021;6(52):Article eabf1571. 31. Deak T. Environmental factors influencing yeasts. In: Péter G,
15. Kim Y, Yang Y, Zhang X, Li Z, Vázquez-Guardado A, Rosa C, editors. Biodiversity and ecophysiology of yeasts.
Park I, Wang J, Efimov AI, Dou Z, Wang Y, et al. Remote control Berlin (Germany): Springer; 2006. p. 155–174.

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 17

32. Mohsen M, Said LA, Elwakil AS, Madian AH, Radwan AG. products, and yeast stress indicators. J Ind Microbiol Biotechnol.
Extracting optimized bio-impedance model parameters 2023;50(1):Article kuad001.
using different topologies of oscillators. IEEE Sensors J. 36. Jones J, Damian DD. A soft fluidic sensor-actuator for active
2020;20(17):9947–9954. sensing of force and displacement in biomedical applications.
33. Wang L, Hu S, Liu K, Chen B, Wu H, Jia J, Yao J. A hybrid In: 2022 IEEE/RSJ international conference on intelligent robots
genetic algorithm and Levenberg–Marquardt (GA–LM) and systems (IROS). New York (NY): IEEE; 2022. p. 6913–6919.
method for cell suspension measurement with electrical 37. Brown D. Video modeling: Combining dynamic model
impedance spectroscopy. Rev Sci Instrum. 2020;91(12): simulations with traditional video analysis. Paper presented at:
Article 124104. American Association of Physics Teachers (AAPT) Summer
34. Sherman F. Getting started with yeast. In: Methods in Meeting; 2008 Jul 19–23; Edmonton, Canada.
enzymology. Amsterdam (The Netherlands): Elsevier; 2002. 38. Lo MD, Ackley SH, Solari P. Homemade ultrasound phantom
Vol. 350, Guide to yeast genetics and molecular and cell for teaching identification of superficial soft tissue abscess.
biology—part B. p. 3–41. Emerg Med J. 2012;29(9):738–741.
35. Guadalupe-Daqui M, Goodrich-Schneider RM, Sarnoski PJ, 39. Yan Y, Pan J. Fast localization and segmentation of tissue
Carriglio JC, Sims CA, Pearson BJ, MacIntosh AJ. The effect abnormalities by autonomous robotic palpation. IEEE Robot
of CO2 concentration on yeast fermentation: Rates, metabolic Autom Lett. 2021;6(2):1707–1714.

Downloaded from https://spj.science.org on May 07, 2025

Soliman et al. 2025 | https://doi.org/10.34133/cbsystems.0233 18