This package contains an implementation of an Extended Kalman Filter (EKF) for state estimation using odometry and IMU measurements. The EKF estimates the position and velocity of a mobile robot in a 2D space.

The following dependencies are required to run the code:

• ROS1 (Robot Operating System)
• Python (version 3.x)
• numpy
• filterpy

1. Install ROS by following the official installation instructions: ROS Installation

2. Install the required Python packages using pip:

   pip install numpy filterpy
   

3. Clone this package to your ROS workspace:

   cd <path_to_your_ros_workspace>/src
   git clone https://github.com/mohdwaseem27/ekf.git
   

4. Build and source the ROS workspace:

   cd <path_to_your_ros_workspace>
   catkin_make
   source devel/setup.bash
   

1. Launch the robot:

   roslaunch ekf ekf.launch
   

2. Add some noise to the odom:

   rosrun ekf odom_noise.py
   

3. Run the EKF node:

   rosrun ekf ekf.py
   

4. Subscribe to the /odom_filtered topic to receive the filtered odometry data.

5. Publish the odometry and IMU messages to the respective topics (/odom_noisy and /sk/imu) for the EKF to process.

• Subscribers:
  • /odom_noisy (nav_msgs/Odometry): Odometry data with noise
  • /sk/imu (sensor_msgs/Imu): IMU measurements
• Publishers:
  • /odom_filtered (nav_msgs/Odometry): Filtered odometry data.

• The initial state of the EKF and the initial state covariance can be adjusted in the __init__ function of the EKFNode class.
• The process noise covariance matrix (Q) and the measurement noise covariance matrix (R) can be modified in the __init__ function as well.

• If needed, you can modify the motion model (motion_model function), the measurement model (Hx function), and their respective Jacobian matrices (F_jacobian and H_jacobian functions) to suit your specific application.
• Additional sensor data or measurements can be incorporated into the EKF by adding new subscriber callbacks and modifying the measurement vector (z) and covariance matrices accordingly.