融合轨迹预测的AEB VRU_TW系统控制策略研究
Research on Control Strategy for AEB VRU_TW System Integrated with Trajectory Prediction
摘要
针对车辆与二轮车混行情况下,自动紧急制动(AEB)系统应对二轮车类弱势交通参与者时响应滞后、难以兼顾安全性和舒适性的问题,提出了一种融合轨迹预测和自适应碰撞时间(TTC)阈值的控制策略.利用自注意力机制和编码器-解码器架构进行二轮车轨迹预测;根据预测结果设计双层制动阈值控制逻辑,其中,一级制动采用模糊控制法和阈值限制保障驾乘人员舒适性,二级制动基于碰撞风险动态触发.模型在环仿真及硬件在环验证结果表明:在30~85 km/h车速范围内,相较于传统策略,所提策略的速度减少量平均提升24.52%;当路面附着系数分别为0.5和0.35时,所提策略的碰撞速度分别为26.75 km/h和31.14 km/h,低于传统策略的40.60 km/h和52.30 km/h.
Aiming at the problem that the Automatic Emergency Braking(AEB)system in mixed traffic scenarios between vehicles and two-wheelers exhibits delayed response and difficulty in simultaneously ensuring safety and comfort when responding to vulnerable road users such as two-wheelers,this paper proposes a control strategy integrating trajectory prediction and an adaptive Time-to-Collision(TTC)threshold.By employing a self-attention mechanism and an encoder-decoder architecture,the trajectory of two-wheelers is predicted.Based on the prediction results,a control logic with two-level braking thresholds is designed:the first-level braking strategy adopts fuzzy control and threshold constraints to ensure occupant comfort,while the second-level braking is dynamically triggered according to collision risk.The model-in-the-loop simulation and hardware-in-the-loop validation results over the vehicle speed range of 30~85 km/h demonstrate that,compared with the conventional strategy,the proposed strategy achieves an average improvement of 24.52%in speed reduction;when the road adhesion coefficients are 0.5 and 0.35,the collision speeds of the proposed strategy are 26.75 km/h and 31.14 km/h,respectively,lower than the corresponding 40.60 km/h and 52.30 km/h obtained with the conventional strategy.
年,卷(期):
2026, (10)
分类号:
U461.91
机标分类号:
U463.33
TM351
TP273.4
在线出版日期:
2026年11月2日
基金项目:
52202414:国家自然科学基金,