- Object Detection - Object Detection is implemented using HoG feature vectors to detect pedestrians from a live video feed.
- State Estimation - Object detection is combined from moving targets and preprocessed using Bayesian (Kalman) filtering to estimate state and detect the location of vehicle in the space.
- Motion Planning - Motion planning of vehicle is optimized using A* path finding algorithm to provide the best route and assist in parking. Obstacle avoidance is also designed to steer the vehicle from any potential danger.
Data sources are provided by TU Delft