Driver intention recognition method based on comprehensive lane-change environment assessment

A driver intention recognition method designed for lateral driving assistance systems is proposed based on comprehensive lane-change environment assessment. A new symbol, Comprehensive Decision Index is designed using fuzzy method to assess the influence of surrounding traffic environment on drivers´ lane-change decisions. Meanwhile, Hidden Markov Model is applied to recognize driver intention. In the model structure, multiple observation variables are used and the elements´ density functions in observation matrix are given the form of Gaussian 3-component mixture model. Finally, data of lane changes performed on driving simulator are used to testify the performance of the proposed method. The results show that the Comprehensive Decision Index is able to make an effective assessment on the environment´s influence on drivers´ lane-change decisions and the algorithm using it as one of the observation signals can both guarantee the accuracy of recognition results and improve the real-time performance.