Nissan has posted a video showcasing its latest achievements in the self-driving part of the industry. The car filmed was able to avoid all obstacles that suddenly came into view.
The system preforms intensive real-time calculations to determine a wide range of possible scenarios arising from sudden obstacles on the road, and then decided on the optimum method to use to prevent collision. It is demonstrated on a 3D-rendered SUV prototype that Nissan plans to launch as a replacement for the Leaf in the near future.
Right now, the company is looking for ways to train the autopilot to make decisions in more challenging situations. The engineering team considers expanding the equipment array on board with a LiDAR, as well as a number of real-time cameras and sensors. The on-board computer will use these to predict multiple scenarios and the ways they can evolve, and settle on the safest action possible.
The obstacles considered involve sudden road accidents right ahead of the car or inanimate objects rolling onto the road. Some scenarios are demonstrated in virtual reality, while others are simulated on a proving ground with a real Nissan Skyline. The car was able to brake successfully in front of another car that pulled back from its parking spot all of a sudden. It also hit the brakes to avoid colliding with a pedestrian and changed lanes quickly to dodge a wheel running loose.
Nissan hopes to wrap up the development by mid-2020s and have the system running on every production car by 2030.