Cadence | 1:51
Cadence has been closely collaborating with the University of Nagoya on AUTOWARE an open source software for highly automated driving. A vehicle was equipped with Lidar, IMU, GNSS, cameras, high-definition digital maps and a high-performance computing platform. Since Autonomous driving requires the processing and transfer of a massive amount of sensor data a low power compute platform is key for efficient use of batteries. Hence a high-performance low-power computing platform is running the AUTOWARE software that is processing all the sensor data and controlling the car accordingly. In order to validate certain automated driving functions the car is driving on a test circuit to prove that it can handle typical traffic situations and safely navigate to a specific destination. A Lidar sensor is scanning the environment around the test vehicle. AUTOWARE’s NDT algorithm matches then the lidar data to a high-definition map and localize the car on it. In parallel camera data is processed by leveraging an AI-based algorithm called YOLO for object detection like pedestrians, cars, traffic signs, etc. A high-performance, low-power SoC with four Tensilica Vision P6 cores is leveraged to process AUTOWARE’s compute-intensive algorithms. As shown in the video the car can drive fully automated to the final destination.