Samočinné řízení modelu vozidla

The aim of this thesis is to demonstrate options for self-driving model cars, focused on local path planning methods and obstacle avoidance. As a part of the project, the model was supplemented by a computing platform Raspberry Pi and appropriate sensors. Specifically, a 2D LiDAR sensor was used for detection and measuring the distance of surrounding objects, an incremental rotary encoder for measuring the distance travelled and current speed, and a gyroscope to keep track of the vehicle's relative orientation. Subsequently, a control system was implemented. This system is able to receive and process sensor data, use it to estimate vehicle's current location, compute an optimal trajectory in an uncharted environment, and control the vehicle's actuators accordingly. The result is a functional model car able to navigate in an unknown environment and reach specified goals by following a trajectory, dynamically generated depending on the surrounding obstacles.

self-driving, model car, local path planning, obstacle avoidance, NXP Cup Alamak, Raspberry Pi, LiDAR, PID Controller, kinematic vehicle models, Dynamic Window Approach