TY  - JOUR
AU  - Jin, Xuebo
AU  - Liu, Songzheng
AU  - Kong, Jianlei
AU  - Bai, Yuting
AU  - Su, Tingli
AU  - Ma, Huijun
PY  - 2024
DA  - 2024/06/08
TI  - GPS Tracking Based on Stacked-Serial LSTM Network
JO  - Chinese Journal of Information Fusion
T2  - Chinese Journal of Information Fusion
JF  - Chinese Journal of Information Fusion
VL  - 1
IS  - 1
SP  - 50
EP  - 62
DO  - 10.62762/CJIF.2024.361889
UR  - https://www.icck.org/article/abs/CJIF.2024.361889
KW  - trajectory estimation
KW  - recurrent neural network
KW  - GPS
KW  - filtering algorithm
KW  - LSTM fusion networks
KW  - stacked serial structure
AB  - Maneuvering target tracking, as a core task in multi-sensor information fusion, is widely used in unmanned vehicles, missile navigation, and underwater ship localization, where real-time and robust state estimation is critical. Due to the uncertainty of the moving characteristics of maneuvering targets and the low sensor measurement accuracy, trajectory tracking has always been an open research problem and challenging work. This paper proposes a Bayesian-inspired stacked LSTM fusion network (SLSTM) for uncertain motion characteristics. The network consists of two LSTM fusion networks with stacked serial relationships, one of which is used to predict the movement dynamics, and the other is used to update the track's state. Compared with the classical Kalman filter based on the maneuver model, the method proposed here does not need to model the motion characteristics and sensor characteristics. It can achieve high-performance tracking by learning historical data dynamics and sensor characteristics. Experimental results show that this method can effectively improve the trajectory estimation performance when the target motion is unknown and uncertain.
SN  - 2998-3371
PB  - Institute of Central Computation and Knowledge
LA  - English
ER  - 

@article{Jin2024GPS,
  author = {Xuebo Jin and Songzheng Liu and Jianlei Kong and Yuting Bai and Tingli Su and Huijun Ma},
  title = {GPS Tracking Based on Stacked-Serial LSTM Network},
  journal = {Chinese Journal of Information Fusion},
  year = {2024},
  volume = {1},
  number = {1},
  pages = {50-62},
  doi = {10.62762/CJIF.2024.361889},
  url = {https://www.icck.org/article/abs/CJIF.2024.361889},
  abstract = {Maneuvering target tracking, as a core task in multi-sensor information fusion, is widely used in unmanned vehicles, missile navigation, and underwater ship localization, where real-time and robust state estimation is critical. Due to the uncertainty of the moving characteristics of maneuvering targets and the low sensor measurement accuracy, trajectory tracking has always been an open research problem and challenging work. This paper proposes a Bayesian-inspired stacked LSTM fusion network (SLSTM) for uncertain motion characteristics. The network consists of two LSTM fusion networks with stacked serial relationships, one of which is used to predict the movement dynamics, and the other is used to update the track's state. Compared with the classical Kalman filter based on the maneuver model, the method proposed here does not need to model the motion characteristics and sensor characteristics. It can achieve high-performance tracking by learning historical data dynamics and sensor characteristics. Experimental results show that this method can effectively improve the trajectory estimation performance when the target motion is unknown and uncertain.},
  keywords = {trajectory estimation, recurrent neural network, GPS, filtering algorithm, LSTM fusion networks, stacked serial structure},
  issn = {2998-3371},
  publisher = {Institute of Central Computation and Knowledge}
}