State Estimation of a Dubin's Car Model using Extended Kalman Filter

Data Required

Each file in the \dataset directory is paired, with controls_observationsN.txt and ground_truth_statesN.txt, where $N$ ranges from 1 to 4.

• Each line of controls_observationsN.txt provides a control/observation pair.
• Each line of ground_truth_statesN.txt provides an instance of the robot's state.

Start with the initial state distributions given by:

• $p_x\sim N(0,(10000m{)}^2)$
• $p_y\sim N(0,(10000m{)}^2)$
• $\theta \sim N(0,(2rad{)}^2)$
• $v=0$
• $\varphi =0$

Assume the vehicle's steering radius is $L=2m$.

Running the Code

To run the Extended Kalman Filter, use:

python3 ekf.py

For some intense documentation, click here.