Kolloquiumsvortrag 12. November 2024, Christian Zöllner (Betreuer: Hielscher)
Real time trajectory planning and vehicle control for Formula Student Driverless
To enable a vehicle to drive autonomously one has to develop a system that can replace a human driver and make correct decisions in real time for safe operation and good driving performance. The system must solve three major classes of problems, first it must perceive its surroundings to detect obstacles and viable paths, secondly it must plan ahead and decide on one specific path to follow, and finally it must issue the correct control inputs to keep the vehicle on this path. With my team at High-Voltage Motorsports e.V. I have developed a system to drive our autonomous race car FAUmax Rho on Formula Student competitions around Europe. In this paper I present the planning and control algorithms we use to transform the latest frames of sensor data into actuator input that steers our car without hitting any cones or leaving the track boundaries to avoid scoring penalties or disqualification. Formula Student racetracks are marked out by traffic cones in a way that is easy for humandrivers to follow but requires significant effort when designing an automated control system, as abstract human intuition must be translated into concrete rules. We tackle this problem by using the layout of these cones to find the approximate track center line that is then used to compute a new target state from which we eventually derive the according actuator input to control the vehicle. During development we make heavy use of our virtual simulation environment together with an extensive set of parameters to validate and tune our algorithms for optimal performance. We show that one of the core problems of Formula Student Driverless can be solved reliably with smart application of geometric algorithms and basic control theory.
Zeit: 10:15 Uhr
Ort: Raum 04.137, Martensstr. 3, Erlangen
oder
Zoom-Meeting beitreten:
https://fau.zoom-x.de/j/68350702053?pwd=UkF3aXY0QUdjeSsyR0tyRWtLQ0hYUT09
Meeting-ID: 683 5070 2053
Kenncode: 647333