A maneuvering tracking method based on LSTM and CS model
Siwei Li, Cheng Hu, Rui Wang, Chao Zhou, Jing Yang · 2019 IEEE International Conference on Signal, Information and Data Processing (ICSIDP) · 2019
Maneuvering target tracking is an important research field in radar tracking. In recent years, the development of machine learning provides a new idea for maneuvering target tracking. This paper presents a trajectory recognition method based on LSTM (Long Short-Term Memory) and a maneuvering tracking method using matched CS (current statistics) model parameter. This method makes use of the characteristics of LSTM that can effectively combine the above information to realize the recognition of target motion states. Then, clustering analysis is used to obtain the optimal filtering parameters of each motion mode in the statistical sense, and filtering is carried out according to the recognition results of LSTM. Compared with the traditional maneuvering tracking method, this method can maintain stable filtering gain in the duration of maneuvering, and the filtering accuracy is improved.