Empowering Internet of Vehicles with Embodied Intelligence: A Survey of Technologies, Evaluation, and Outlook
Year: 2026
Type: preprint
Abstract: The Internet of Vehicles (IoV) advances intelligent transportation by allowing vehicles, roadside infrastructure, pedestrian devices, and edge and cloud resources to share observations, computing resources, and traffic information. However, the traditional IoV systems often process perception, scene understanding, driving decisions, and vehicle control sequentially without continuously using feedback from vehicle motion execution to update later understanding and decisions, allowing state estimation and decision errors to persist across driving cycles and resulting in outdated vehicle states, decisions inconsistent with traffic conditions, and infeasible or unsafe maneuvers. Embodied intelligence forms a closed-loop in which vehicles perceive traffic conditions, understand scene changes, make decisions, execute control actions, and use environmental feedback to update later perception and decisions. This closed-loop reduces error accumulation, aligns traffic understanding with current road conditions, and improves vehicle action feasibility and adaptability. Existing surveys have reviewed embodied intelligence and related autonomous driving technologies. However, the closed-loop relationships among vehicle-to-everything (V2X) observations, traffic understanding, driving decision-making, vehicle execution, and environmental feedback have not been systematically examined. This survey organizes relevant studies around intelligent vehicle perception, traffic scene understanding, driving policy generation, and vehicle motion control, and clarifies the closed-loop process in which vehicles integrate multi-source observations collected by sensors to understand the current traffic scene, generate driving policies, control vehicle motion, and incorporate vehicle motion feedback into the next cycle of perception and decision making. The survey also summarizes representative datasets, metrics, challenges, and future directions for connected vehicles.
Source: Research Square
Language: English
Subfield: Automotive Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7225620370