Multisensor Multitarget Tracking Arithmetic Average Fusion Method Based on Probabilistic Time Window
Kuiwu Wang, Qian-Han Zhang, Xiaolong Hu · IEEE Sensors Journal · 2023
The deployment of sensors in multisensor network systems needs to measure the actual state of the target. There may be significant errors in transmitting information to each sensor node through the sensor network system, which cannot accurately reflect the actual target state, resulting in the sensor network generating wrong measurements and instructions. The accuracy of target state estimation is affected by two factors: random noise in the process of sensor measurement and random delay of sensor data information in the network transmission process. Because of the above problems, this article designs a probabilistic time-window multisensor arithmetic average (AA) fusion multitarget tracking algorithm that can resist these interference factors, compensate for the random delay, and calculate the correct probability cumulative value of different sensor nodes in the time window, and then adjusts the probability hypothesis density filter update value as the multisensor fusion weight to reduce the negative impact of random interference on the sensor measurement. The simulation results show that under the condition of accurate probability information of random noise and random delay, compared with the traditional AA fusion and generalized covariance intersection (GCI) fusion algorithms, the proposed probabilistic time-window multisensor network AA fusion algorithm has the most minor tracking error and is more suitable for complex tracking scenarios with clutter and detection probability changes.