基于HMM的电动汽车行车轨迹纠偏优化

Optimization of electric vehicle trajectory correction based on hidden Markov model

为准确获取汽车行车轨迹跟踪数据,提出一种结合电子地图导航定位数据的行车轨迹纠偏优化方法,基于隐马尔可夫模型(hidden Markov model,HMM)构建观测概率、状态概率和状态转移概率,提出使用隐马尔可夫纠偏定位点优化算法优化原始定位点,剔除偏差较大轨迹点,提升轨迹纠偏的正确率。通过在实际的行车轨迹数据上应用Viterbi算法求解建立的模型,验证了基于HMM的行车轨迹纠偏方法的可行性和应用性。

To obtain vehicle trajectory tracking data accurately,an optimization method for trajectory correction based on navigation and positioning data of electronic map was presented.Observing probability,state probability and state transition probability were constructed based on hidden Markov model.The genetic algorithm was used to optimize the original location points and eliminate the trajectory points with large deviation,so as to improve the accuracy rate of trajectory correction model.The feasibility and applicability of the HMM-based trajectory correction method are verified by applying Viterbi algorithm to the actual trajectory data.