Vehicle behavior prediction based on attention mechanism
1. Automotive Engineering Research Institute, Jiangsu University, Zhenjiang, Jiangsu 212013, China; 2. School of Automotive and Traffic Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, China
Abstract：To predict the future behavior of surrounding vehicles in intelligent driving scenarios, the vehicle behavior prediction method was investigated based on the attention mechanism of long and short-term memory(LSTM) network model. The time series data division method of non-uniform step-size was proposed to classify the vehicle time series information of specific behaviors. With LSTM as the basic neural network framework, the attention mechanism was used to assign different weight values for each time step according to the importance in the time series information. The historical trajectory information of the target vehicle and its surrounding vehicles was used as algorithm input to predict the future motion behavior of the target vehicle. The results show that the algorithm can solve the problems of information omission and increasing of computing resource burden caused by the fixed step-size classification method. The algorithm can effectively improve the accuracy of behavior prediction and reduce the time using of vehicle behavior prediction.
YANG Q, KOUTSOPOULOS H N. A microscopic traffic simulator for evaluation of dynamic traffic management systems［J］. Transportation Research Part C, 1996, 4(3): 113-129.
［2］
MORRIS B T, TRIVEDI M M. Trajectory learning for activity understanding: unsupervised, multilevel, and long-term adaptive approach［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(11): 2287-2301.
［3］
DUEHOLM J V, KRISTOFFERSEN M S, SATZODA R K, et al. Trajectories and maneuvers of surrounding vehicles with panoramic camera arrays［J］. IEEE Transactions on Intelligent Vehicles, 2016, 1(2): 203-214.
［4］
KHOSROSHAHI A, OHN-BAR E, TRIVEDI M M. Surround vehicles trajectory analysis with recurrent neural networks［C］∥2016 IEEE 19th International Confe-rence on Intelligent Transportation Systems. Piscataway, USA: IEEE, 2016: 2267-2272.
［5］
SCHREIER M, WILLERT V, ADAMY J. Bayesian, maneuver-based, long-term trajectory prediction and criticality assessment for driver assistance systems［C］∥17th International IEEE Conference on Intelligent Transportation Systems. Piscataway, USA:IEEE, 2014: 334-341.
［6］
HOUENOU A, BONNIFAIT P, CHERFAOUI V, et al. Vehicle trajectory prediction based on motion model and maneuver recognition［C］∥2013 IEEE／RSJ Internatio-nal Conference on Intelligent Robots and Systems. Pisca-taway, USA:IEEE, 2013: 4363-4369.
［7］
HOCHREITER S, SCHMIDHUBER J. Long short-term memory［J］. Neural Computation, 1997, 9(8): 1735-1780.
［8］
ONDRUSKA P, POSNER I. Deep tracking: seeing beyond seeing using recurrent neural networks［C］∥Thirtieth AAAI Conference on Artificial Intelligence. Pisca-taway, USA:IEEE, 2016: 3361-3367.
［9］
KRAJEWSKI R, BOCK J, KLOEKER L,et al. The highd dataset:a drone dataset of naturalistic vehicle tra-jectories on german highways for validation systems［C］∥2018 21st International Conference on Intelligent Transportion Systems. Piscataway, USA:IEEE, 2018:2118-2125.
［10］
HOCHREITER S, SCHMIDHUBER J.LSTM can solve hard long time lag problems［C］∥Advances in Neural Information Processing Systems, 1997: 473-479.
2020, Vol. 41 
