ITS2807智能交通世界大会ITS智慧城市社区人工智能AI物联网IT报告课件教案.pdfVIP

Monitoring-based driving behavior modeling for a self-learning car Introduction Rapid advancement in technology Available Information Advanced driving assistance Personalized driving experience Self-learning car Main ideas Modeling the driver behavior based on the monitoring of the Main title driver-vehicle-environment (DVE) system Formal modeling of the DVE system Content The monitoring ensures personal driving behavior models The modeling of the whole DVE system ensures an efficient Situation-adaptive assistance The formal models facilitate the verification of the safety of the driving behavior Smart Driving Behavior Recording System System architecture Content Driving behavior recorder  Recording driving behavior as a probabilistic automaton. Main title  Driving behavior: A change in the dynamic state of the vehicle made in response to a specific driving situation Content  Recording driving behavior is based on vehicle dynamics, driver’s control action and driving environment information  DV system (Driver + Vehicle): A hybrid automaton that receives environment information as inputs and performs control actions as outputs.  Driving trip: Transitions enabled by the driver action between the state of the automaton Formal model of driving behavior - Probabilistic Rectangular Hybrid Input/Output Automata - Main title  V: Variables ( ≜∪∪)  H: internal variables  U: input variables Content  Y: output variables  ⊆ : state space  Θ: initial states  : → : rectangular predicates over H  A: actions ( ≜∪∪)  I: input actions