INTELLIGENT COOPERATIVE PERCEPTION WITH INFORMATION-CENTRIC NETWORKING FOR SUPPORT OF MULTI-DOMAIN MISSIONS
iCOOPER is a distributed, deep reinforcement learning-based, intelligent cooperative perception software platform. It enables Autonomous Vehicles (AVs) and Autonomous Air Vehicles (AAVs) to share point cloud sensory information with other AVs/AAVs and auxiliary nodes. Thereby, extending view coverage, removing blind spots, and improving object detection accuracy.
iCOOPER software is deployed on multiple LiDAR sensor nodes in an ad hoc network. It transmits individual node point clouds to identified nodes and performs data fusion. Thereby, providing each node a complete LiDAR image for increased spatial awareness.
iCOOPER is capable of integrating with any sensor node (e.g. passenger vehicle console) or auxiliary node (e.g. smartphone), and location-based visualization platforms.
METHOD: Cooperative 3D sensor perception
KEYWORDS: 3D point cloud, Compression, Decompression
TECHNICAL SPECS: Information-centric networking, Deep reinforcement learning, Data fusion
RESEARCH: M. Bui, L. -C. Chang, H. Liu, Q. Zhao and G. Chen, “Comparative Study of 3D Point Cloud Compression Methods,” 2021 IEEE International Conference on Big Data (Big Data), 2021, pp. 5859-5861, doi: 10.1109/BigData52589.2021.9671822.