Multi-Sensor Fusion-Based Highway Vehicle Behavior Prediction and Its Application in Adaptive Cruise Control

Zihan Wang,Yang Liu,Lihui Peng,Danya Yao,Yi Zhang

DOI: https://doi.org/10.1109/tiv.2024.3486037

IF: 8.2

2024-01-01

IEEE Transactions on Intelligent Vehicles

Abstract:We propose a vehicle behavior prediction approach using multi-sensor data fusion to address the challenges of high-level autonomous driving on expressways and deploying on vehicles with limited computation hardware capabilities. The proposed approach significantly reduces the computational intensity and memory requirements by establishing a sequential fusion framework and incorporating a hierarchical target association. Furthermore, the interactions between the ego-vehicle and background vehicles are analyzed to extract critical features such as congestion levels and occupancy states. These features are fed into a Long Short-Term Memory (LSTM) model to predict the behaviors of key targets within the ego-vehicle's vicinity. Experimental results demonstrate that the proposed vehicle behavior prediction approach reduces the time required for background vehicle behavior analysis and enhances reliability and safety by providing valuable insights for defensive driving and planning strategies. Additionally, the ACC real-vehicle test incorporating the proposed behavior prediction reveals that the system can anticipate and respond appropriately to potential hazards, maintaining a safe distance between the ego-vehicle and surrounding traffic, thereby contributing to safer and more efficient autonomous driving.