Distributed Driving System for Coupled Small EV Using Neural Network and Load Cell

Japan is facing difficulties with providing general public transport systems in depopulated cities. This paper proposes a new means of transport for application to such cities. The system is based on small electric vehicles with mechanical couplers that are designed to attach to devices installed at each home. Unfortunately, this system incurs an issue of enhancing the operability for drivers of the coupled EVs. Therefore, this study set out to develop a distributed driving system that ensures that a group of vehicles maintains the same driving characteristics as those of a single vehicle, regardless of the number of coupled vehicles. This system suppresses fluctuations in the driving characteristics by means of online learning by the vehicle model using a neural network and by feedback control using a force gauge. This paper describes the details of this system and evaluation experiments using two small EVs.