Reliable Data Compression and Fusion for Cooperative Perception and Vehicular Communications

2022 Fiscal Year Final Research Report

• Project/Area Number: 21K21300
• Research Category: Grant-in-Aid for Research Activity Start-up
• Allocation Type: Multi-year Fund
• Review Section: 1001:Information science, computer engineering, and related fields
• Research Institution: National Institute of Informatics
• Principal Investigator: Aoki Shunsuke 国立情報学研究所, アーキテクチャ科学研究系, 助教 (20910475)
• Project Period (FY): 2021-08-30 – 2023-03-31
• Keywords: 自動運転 / センサ融合 / 無線通信 / 協調センシング / 深層強化学習 / コネクテッドカー

Outline of Final Research Achievements

In this research project, we worked on research and development of cooperative sensing and cooperative perception, in which sensor information is shared and utilized in a distributed manner between automated and connected vehicles and urban IoT infrastructure. In this research, we first defined the granularity and type of information from blind spots, then designed and developed a data sharing mechanism to prevent "information flooding" and "dissemination of false information," and conducted implementation and evaluation experiments on simulators and actual vehicles. In this evaluation experiment, we first conducted experiments using the vehicle simulator CARLA and the communication network simulator SUMO, and then conducted experiments using actual intersections, radio equipment, and autonomous mobile robots. The results of this research and experiments were presented externally in international magazines and at international conferences.

Free Research Field

ロボティックス

Academic Significance and Societal Importance of the Research Achievements

本研究課題の遂行によって、自律移動ロボットのセンサ情報共有基盤を構築することができた。本基盤ソフトウェアを用いることによって、T字路・十字路などの死角のある場面でも衝突・デッドロックを避けながら安全に屋内移動ロボットを利活用することが可能となる。