Learning faster to perform autonomous lane changes by constructing maneuvers from shielded semantic actions
| Autor: | Lennart Svensson, Andrew Backhouse, Dapeng Liu, Mattias Brännström |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
050210 logistics & transportation
Focus (computing) Process (engineering) Computer science business.industry 05 social sciences 010501 environmental sciences 01 natural sciences law.invention law 0502 economics and business Shielded cable Systems design Markov decision process Artificial intelligence business 0105 earth and related environmental sciences |
| Zdroj: | ITSC |
| DOI: | 10.1109/itsc.2019.8917221 |
| Popis: | This paper introduces a new method to solve tactical decision making problems for highway lane changes. In the system design, reference sets for low level controllers are employed to formulate semantic meaningful actions used by reinforcement learning algorithm. Safety is ensured by preemptively shielding the Markov decision process (MDP) from unsafe actions. This frees the agent to focus on learning how to interact efficiently with the surrounding traffic. By introducing human demonstration with supervised loss as better exploration strategy, the learning process and initial performance are boosted further. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|