Prediction Capabilities for Cyber Physical Vehicles

Abstract

Cyber physical systems open new ground in the automotive domain. Autonomous vehicles will try to adapt to the changing environment, and decentralized adaptation is a new type of issue that needs to be studied. This article investigates the effects of adaptive route planning when real-time online traffic information is exploited. Simulation results show that if the agents selfishly optimize their actions, then in some situations, the cyber physical system may fluctuate and sometimes the agents may be worse off with real-time data than without real-time data. The proposed solution to this problem is to use anticipatory techniques, where the future state of the environment is predicted from the intentions of the agents. This article concludes with this conjecture: if simultaneous decision-making is prevented, then intention-aware prediction can limit the fluctuation and help the cyber physical system converge to the Nash equilibrium, assuming that the incoming traffic can be predicted.