汽车防撞系统中目标跟踪与防撞决策研究

为了实现汽车主动安全系统中的目标跟踪与防撞,提出了混合式汽车防撞系统信息融合结构模型,采用分级信息融合实现目标跟踪,推导出了基于跟踪残留误差和预测残留误差共同校正的融合算法,并给出了算法的实现结构;提出了基于局部分析的映射变换方法,实现驾驶模型特征向量连续、实时的修正,在此基础之上,利用模糊积分方法融合多种相关信息,确定汽车应采用的安全运行模式,实现主动安全防撞决策。经过大量试验证明:该算法具有很好的稳定性和准确率。

Research on target tracking and decision-making for collision avoidance in vehicle collision avoidance system

To realize target tracking and collision avoidance in the system of vehicle active safety, a hybrid information fusion structure model of vehicle collision avoidance system is presented to implement the target tracking with the method of hierarchical fusion. A new algorithm for multisensor fusion based on residual track error and predicted residual error is proposed and the realization structure for the new algorithm is also presented. Then, a transform method based on local reconstruction is presented to modify the character vector continuously and real-time, and based on it, an information fusion method based on fuzzy integral is presented to make a decision for travelling model. Lots of experiments show that the new algorithm is stable and exact.