Safety issues related to autonomous vehicles are of great concern both in the academy and industry, identifying the braking system performance as a crucial research field. In this work, an autonomous braking system based on deep reinforcement learning is proposed, employing an intelligent agent trained in city scenarios to manage both pedestrians' safety and passengers' comfort. The agent is modelled via the deep deterministic policy gradient algorithm in a software environment and its performance is tested showing good results in maximizing both pedestrians' safety and passengers' comfort.
Safety and Comfort in Autonomous Braking System with Deep Reinforcement Learning / Fanti, M. P.; Mangini, A. M.; Martino, D.; Olivieri, I.; Parisi, F.; Popolizio, F.. - 2022-:(2022), pp. 1786-1791. (Intervento presentato al convegno 2022 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2022 tenutosi a cze nel 2022) [10.1109/SMC53654.2022.9945383].
Safety and Comfort in Autonomous Braking System with Deep Reinforcement Learning
Fanti M. P.;Mangini A. M.;Martino D.;Olivieri I.;Parisi F.;Popolizio F.
2022-01-01
Abstract
Safety issues related to autonomous vehicles are of great concern both in the academy and industry, identifying the braking system performance as a crucial research field. In this work, an autonomous braking system based on deep reinforcement learning is proposed, employing an intelligent agent trained in city scenarios to manage both pedestrians' safety and passengers' comfort. The agent is modelled via the deep deterministic policy gradient algorithm in a software environment and its performance is tested showing good results in maximizing both pedestrians' safety and passengers' comfort.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.