CHEN Jun, LI Da-wei, CAO Xiao-hua, PENG Yong, JIA Hong-fei, WANG Yan, HE Chong-qi
With the continuous development of next-generation information and communication technologies, artificial intelligence, autonomous driving, and vehicle-road-cloud collaboration, transportation systems are rapidly evolving from informatization and intelligence toward autonomy. In this process, perception is no longer confined to information acquisition within the local environment of a single vehicle, but is gradually becoming a fundamental capability for supporting the safe operation, collaborative organization, and closed-loop regulation of autonomous transportation systems. Although substantial progress has been made in in-cabin monitoring, external environment perception, vehicle-road collaboration, cloud-edge collaboration, and cognitive interaction, existing studies still mainly focus on single objects, single spatial scopes, or local technical chains, and there remains a lack of unified and systematic reviews from the perspective of the overall operation process of autonomous transportation systems. Therefore, a systematic review of holographic perception for autonomous transportation systems was presented from the perspectives of demand characteristics, conceptual connotation, hierarchical architecture, key technologies, and development trends. First, the expansion characteristics of perception objects, perception space, information modalities, and supporting conditions under autonomous operation were analyzed, and the basic connotation of holographic perception, together with its relationship with traditional single-vehicle perception and cooperative perception, was clarified. Then, a holographic perception architecture consisting of the in-cabin perception layer, external perception layer, roadside cooperative perception layer, cloud-edge collaboration layer, and cognitive interaction and feedback layer was constructed. Furthermore, the research progress of key technologies, including in-cabin state perception, visible-range external perception, roadside cooperative perception, cloud-edge collaboration, and multi-source cognitive interaction and feedback, was systematically reviewed. The results show that perception technologies for autonomous transportation systems are shifting from local single-vehicle environmental perception to multi-level cooperative perception for the overall operational state of “human-vehicle-road-cloud” systems. In-cabin perception, visible-range external perception, roadside cooperative perception, cloud-edge collaboration, and cognitive interaction and feedback play key roles in human-state understanding, local environment modeling, beyond-line-of-sight information supplementation, cross-node data and computing support, and structured cognitive output, respectively, and jointly constitute a holographic perception system for operational state acquisition and environmental understanding in autonomous transportation systems. Finally, the major challenges of holographic perception in cross-domain fusion, spatiotemporal alignment, cooperative robustness, real-time communication, trustworthiness and security, and unified high-level semantic representation were summarized, and its future development toward multi-level collaboration, global state understanding, and integrated perception-cognition-feedback closed-loop systems was discussed. This review can provide a reference for the framework construction, key technology development, and system integration of holographic perception for operational and environmental state awareness in autonomous transportation systems.
