Unmanned vehicle products of Mitsubishi Heavy Industries, Ltd. (MHI)-including unmanned forklifts, underwater vehicles, etc.-require route planning to bring the vehicle group to a target position under complicated constraints such as maximizing carrying efficiency, avoiding collisions between vehicles, maintaining a constant formation, etc. However, if all the information is aggregated on the server that plans the route for all the vehicles and a server overload condition or communication interruption occurs, the server cannot process the route plan and the entire vehicle group stops. For this reason, it is necessary to reduce the server load by decentralizing the process to each vehicle. We developed a swarm intelligence technology for upwards of 10 vehicles based on model predictive route planning that decentralizes route planning under complicated constraints to each vehicle through the joint research with Nagoya University, which is presented in this report.
