.png){kind=icon}
.png){kind=icon}
Return
RFID-Based Vehicle Detection and Positioning for Autonomous Driving
.png){kind=icon}
PRE
AI
R
刘
J
H
S
J
A
DOI：10.1109/TIV.2026.3668860.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Vehicle detection is crucial for environmental perception tasks in autonomous driving, particularly at higher SAE automation levels (e.g., Levels 4–5), where reliable sensing enables the removal of the human driver from the control loop. Autonomous vehicles typically rely on cameras, LiDAR, and radar to perceive surrounding vehicles, yet detection remains challenging due to heterogeneous surroundings. Existing statistical or neural-network-based methods heavily depend on prior vehicle feature knowledge that may not always be available, while road vehicles are passive, relying on others to detect them. This paper proposes an active RFID-based approach to enhance detectability: multiple active RFID tags attached to a vehicle’s surfaces periodically broadcast vehicle-specific information. Nearby equipped vehicles use RFID readers to identify and localize the target directly via tag IDs and data—offering more direct identification than probabilistic sensor methods, although performance remains environment-dependent. Four key challenges are addressed: (1) simplification of 3D models for tag storage, (2) design of efficient data structures, (3) infrastructure-independent two-way ranging for positioning, and (4) real-time angle-of-arrival orientation estimation. The approach enables detection in common adverse conditions (e.g., rain, low light), position/orientation estimation, 3D boundary recovery for occluded vehicles via tag redundancy, and detectability using rechargeable tags. Experiments and simulations validate proof-of-concept feasibility in controlled scenarios, indicating the need for larger-scale testing in dynamic environments. Initial adoption should target niche pilots (e.g., ports, mining, industrial compounds) to complement sensors and reduce risks, with gradual future integration via cooperating OEMs.
Keywords:
Active RFID
autonomous driving
occlusion and invisibility
model-based segmentation
two-way ranging
angle of arrival
orientation estimation
vehicle detection
vehicle safety
Journal
I
IF:
14.3
Papers:
1.2K
Citations:
1.2W
Organization
Cited Papers