Report No: 1
Team : Fusionfleet
Date:16.12.2024
1 Introduction
This report highlights our project's initial activities, current progress, and key details. Our
primary focus has been on advanced software development, including the creation of
algorithms for accurate lane detection, tracking, and traffic sign identification. Driven by
enthusiasm, our team has maintained seamless collaboration, both remotely and in the
lab, while effectively managing time constraints. This document reviews our
achievements thus far and lays the groundwork for the tasks ahead.
2 Planned activities
The activities planned are as follows:
Architecture
Project plan
Lane detection and tracking
Sign detection and classification
Traffic light Classification
Architecture
Status: Completed
Implementation: Our team has successfully designed an architecture leveraging
our collaborative efforts and shared expertise
Difficulties: A significant challenge in real-time development was the uncertainty
surrounding potential add-ons before project initiation. Determining additional
features and functionalities required careful evaluation and consensus-building
among team members, making the process both intricate and collaborative.
Project plan
Status: Completed
Implementation: Aligned project development with the established
architectureand planned activities based on the allocated time.
Difficulties: The real-time challenge was adapting the project plan to include
unforeseen complexities. Decisions on resource allocation and timelines
required continual adjustment.
1
Lane detection and tracking
Status: ongoing
Implementation: The assignment was to create a program that tracks the lane
usingvideo input of a track present in BFMC and detects lane.
Difficulties: The developed program can identify lanes in a straight line, but
implementing it in a curved path appeared to be challenging. Higher ambient light
intensity causes some lane tracking to be inefficient, which is difficult to fix.
Traffic Sign detection and classification
Status: ongoing
Implementation: Collected dataset and added data to the existing one under
various lighting conditions, trained the model and got accuracy of 96%while
testing the model, if the mode is changed from manual to auto the vehicle
starts running when the stop sign is detected the vehicle is being stopped
Difficulties:
Intergrating the real time video footage for object detection was complex and
working on it
Traffic Light Classification
Status: ongoing
Implementation: Collected data for new environment and currently in the
processof training the model.
Difficulties: Classification on the traffic light on which it differentiate is
really complex and working on it
3 General status of the project
The project has made significant strides since the completion of the project plan,
architecture definition and controlling the vehicle in manual mode. Real-time challenges
persist in making crucial decisions on addons and improvements that can elevate the project's
performance, especially in a high-end competition setting. We are currently working on lane
dectection on straight pathways and traffic sign detection on with real-time challenges
revolving around decisions on strategic add on to enhance adaptability to curved paths and
dynamic lighting conditions.
4 Upcoming activities
1. Perform real-time lane detection and tracking, focusing on improving efficiency on
curved pathways.
2. Continue real-time implementation of traffic sign detection and classification with
enhanced stability and accuracy.
3. Initiate the development of a real-time intersection detection module.
4. Begin implementing a real-time traffic light detection system.
2
