Vehicle Localization and Control Simulation based on non-holonomic kinematics based on bicycle model
This project simulates a 2D robotic vehicle using a Controller for autonomous steering. The vehicle operates in both a global and local frame, navigating through a virtual world using feedback from its position and orientation.
The simulation includes:
- Real-world dynamics: Continuous updates of the robot’s position, velocity, and steering.
- Localization: A low-frequency sensor-like localization update.
- Stanley Controller: A control algorithm used in autonomous driving to correct the robot's path based on cross-track and heading errors.
- OpenCV Visualization: Displays the robot's movement using OpenCV to track its local and global positions.
- Control noise simulation: Adds steering noise to simulate real-world control imperfections.
The project is written in Python and uses libraries such as numpy, opencv.
- Edited Stanley Controller: Adjusts the steering angle based on cross-track and heading errors to follow a desired path.
- Real-time visualization: Displays the robot’s path and position using OpenCV.
- Noise simulation: Adds randomness to steering inputs to simulate real-world vehicle dynamics.
- Feedback mechanism: Uses both position and orientation feedback for control.
- Configurable parameters: Modify velocity, steering noise, and control gains in the code to observe different behaviors.