Reliable and safe control Navigation for Autonomous Vehicles in Dynamic Urban Environments
Contrôle et navigation fiable de véhicules autonomes en environnements urbains dynamiques
Résumé
In this thesis is presented an algorithmic architecture for systematic risk evaluation, mitigationand management intended for autonomous transportation vehicles. The methodspresented span low level control, trajectory tracking and multi-vehicle coordination. Atask separation between low level steering control and trajectory tracking has been implementedto spread the design e_ort across two functional blocks. A robust low levelcontroller has been designed, and a comfortable and flexible Model Predictive Controller(MPC) has been implemented for trajectory tracking. This controller has been associatedwith a supervision mechanism that monitors its performance in real time to evaluate theprobability to underperform. When such a risk is identified, the speed of the system isadapted. The multi-vehicle coordination block fulfils the planning task. It is a decentralized,probabilistic optimization algorithm that is naturally risk-adverse. It is based on theProbability Collectives (PC) algorithm and operates a multi-stage negociation betweenvehicles. It has been made compatible with mixed-tra_c scenarios with human drivers onthe road. Results show that risks are monitored and managed across the whole architecture.Furthermore, easy to understand risk metrics are outputted to make the algorithmsdecisions understandable by the users and engineers working on the system. The work inthis thesis thus proposes systematic risk management techniques transposable to all autonomousvehicles systems. It has been tested in simulations and on the autonomous testvehicles available at the Institut Pascal.
Résumé indisponible.
Domaines
Automatique / RobotiqueOrigine | Version validée par le jury (STAR) |
---|