We have developed a path planning technology that achieves smooth driving by taking into account the surrounding environment, and a safe and secure vehicle control technology that maximizes vehicle performance.

These technologies contribute to the realization of safe, secure and comfortable autonomous driving.

We have developed a unique path planning method called "PF-RRT." Using a particle filter^*1 (PF), the system estimates the change in state needed for the vehicle to become closer to the target state, such as being positioned in the center of the target lane, and generates a path.

The tree (path candidates) can be expanded using a rapidly exploring random tree^*2 (RRT), and the best path with the lowest achievement cost (e.g. distance) is selected.

The method enables the generation of a smooth path even in complicated environments.

We have developed a controller that delivers integrated longitudinal and lateral vehicle control using nonlinear model predictive control (NMPC).

By optimizing vehicle steering angle and velocity using a nonlinear vehicle and tire model, the vehicle can be controlled to accurately follow the target path.

The controller realizes smooth driving that circumvents obstacles with minimal path errors, and a comfortable ride.