Artificial Intelligence is massively used in achieving fully autonomous vehicles. From processing of sensor data to decision making in different scenarios, AI is constantly learning and improving the perception of the vehicles. LiDAR, as a primary component of the sensors, scans the surroundings of a vehicle and produces accurate geo-referenced 3D point clouds that assist in object detection and tracking. The input from LiDAR plays an important role in the Advanced Driver-assistance Systems (ADAS) by providing much higher depth resolution than images captured by cameras. Despite the rapid evolving of computer vision technology, I believe there is also great potential for point-cloud processing algorithms. My research focuses on improving such algorithms using artificial intelligence solutions to better facilitate object detection, recognition and classification.
Meanwhile, I’m also drastically interested in the role of autonomous vehicles in the growing industry of Internet of Things (IoT). As 5G becomes available, connections and communications among “things” will be much faster with extremely low latency. It will also be a driving force for autonomous vehicles by improving V2X where the information sharing of vehicles, infrastructures, pedestrians and networks are much better connected.