Speaker
Mr
Zhengyang Mao
(National Key Laboratory of Aerospace Flight Dynamics, School of Astronautics, Northwestern Polytechnical University)
Description
As a branch of the field of target tracking, maneuvering target tracking plays an increasing role in military and civilian fields. A novel maneuvering target tracking algorithm is investigated. Drawing on the experience of combination idea of the modular structure and the fuzzy interacting multiple model algorithm (FIMM), a modular fuzzy interacting multiple model algorithm (MFIMM) is presented. The MFIMM algorithm consists of three independent modules working in parallel, called non-maneuver, weak maneuver and strong maneuver respectively. And the change of a target motion is also divided into three levels: no, small, and big. The motion of a target is detected by a fuzzy control motion detector, which imitates the thoughts of human beings to detect a target’s motion. Once the maneuver is detected, the MFIMM algorithm selects one of the three modules matching the actual movement of the target every moment according to maneuver condition, and the state vector and covariance matrix is compensated, so that the modified state can suit the actual motion well. Afterwards, the MFIMM algorithm estimates the state of the target through interactive multiple model algorithm (IMM) based on square root unscented Kalman Filter (SRUKF) of the selected module. Therefore, under the architecture of the proposed algorithm, the fuzzy motion detector deals with the level of motion and the modules switching, whereas the IMM-SRUKF accounts for the estimation of the dynamic system. At the end of the paper, simulation is performed on the problem of maneuvering target tracking in two-dimensional space. In order to evaluate the effectiveness of the MFIMM method, Root Mean Square Error (RMSE) of the estimated state is used. The simulation test tracks a same target with determined trajectory by MATLAB emulation comparing the two algorithms, MFIMM and IMM. Results demonstrate that the proposed MFIMM algorithm improves the tracking precision and reduces the computational burden compared with traditional IMM.
| Applicant type | First author |
|---|
Primary author
Mr
Zhengyang Mao
(National Key Laboratory of Aerospace Flight Dynamics, School of Astronautics, Northwestern Polytechnical University)
Co-author
Prof.
Zhanxia Zhu
(National Key Laboratory of Aerospace Flight Dynamics, Northwestern Polytechnical University)
