Current State of the Transportation Sector
In recent years, there has been an urgent need to improve the convenience of public transportation and to secure new means of transportation, such as on-demand transportation, for the purpose of solving the issues of an aging society and regional development. The issue of unprofitable routes has not been resolved nationwide, with buses running even though there are few passengers due to fixed operation schedules.
We are researching and developing algorithms and software for future urban development in which transportation systems are optimized as a whole city, with individual mobility systems determining the optimal routing based on data such as traffic congestion information and weather forecasts, and on-demand transportation that responds to ever-changing passenger demand as IOT and DX become increasingly prevalent in the future.
Automated Lane Change Using a Multi-Agent System
In a world where automated driving has become commonplace, individual cars changing lanes on their own can cause accidents and traffic jams. By having multiple cars give way to each other, all vehicles in the vicinity can reach their destinations safely and quickly. This is made possible by a multi-agent system that enables “team play.
Deep Reinforcement Learning for Combinatorial Optimization of Traffic Networks
In recent years, there has been a lot of research on using reinforcement learning to solve combinatorial optimization problems. It is difficult for humans to construct optimal routes from multiple locations (mainly large stations) to multiple stops. State-of-the-art machine learning techniques and mathematical optimization methods must be used to realize on-demand transportation with predicted demand.