Dynamic on-road driving scenarios require robust methods for planning a safe and feasible vehicle motion coping with both static and dynamic obstacles. Many of the different approaches which have been proposed to tackle this challenge are based on optimal control and employ local continuous or discrete optimization schemes. While discrete methods possess the ability to find reasonable solutions in a search space incorporating local minima, they tend to sacrifice optimality for real-time performance. On the other hand, local continuous methods require suitable initialization to handle the combinatorial nature of on-road driving scenarios but are capable of quickly returning an optimized solution. The presented work proposes a hybrid solution which embraces both strategies to unite their distinct advantages. A discrete optimization scheme is augmented by parametric optimization to achieve both low planning times as well as robust (re-)initialization.
Hybrid discrete-parametric optimization for trajectory planning in on-road driving scenarios
01.11.2016
548868 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
TOWARDS RISK MINIMIZING TRAJECTORY PLANNING IN ON-ROAD SCENARIOS
British Library Conference Proceedings | 2018
|TRAJECTORY PLANNING IN AUTONOMOUS DRIVING VEHICLES FOR UNFORESEEN SCENARIOS
Europäisches Patentamt | 2024
|