Distributed consensus Kalman filtering for asynchronous multi-rate sensor networks

• Published in: Signal, image and video processing Vol. 18; no. 8-9; pp. 6419 - 6429
• Main Author: Shao, Teng
• Format: Journal Article
• Language: English
• Published: London Springer London 01.09.2024; Springer Nature B.V
• Subjects: Algorithms; Computer Imaging; Computer Science; Cost analysis; Cost function; Error analysis; Image Processing and Computer Vision; Kalman filters; Multimedia Information Systems; Networks; Original Paper; Pattern Recognition and Graphics; Sampling; Sensors; Signal,Image and Speech Processing; Vision; Multi-rate sampling; Asynchronous measurements; Sensor network; Optimal distributed filter; Kalman consensus filtering
• ISSN: 1863-1703; 1863-1711
• DOI: 10.1007/s11760-024-03326-7

Distributed filtering over sensor networks has received much attention due to its wide applications and has been extensively studied especially under the consensus strategy. For simplicity, most existing distributed filtering algorithms assume that all sensors observe the system state synchronously, i.e., all sensors have the same sampling rate and initial time. However, different sensors may have different sampling rates in practice. For such scenarios, a synchronous distributed filter under consensus strategy cannot be applied directly due to missing measurements of some sensors at some times. In this paper, to deal with the asynchronous measurements caused by different sampling rates in sensor networks, the optimal multi-rate distributed consensus Kalman filter for asynchronous sensor networks is proposed. First, the measurement model of multi-rate sampling is re-modeled by the indicator function. Then a multi-rate distributed consensus Kalman filter is designed for each sensor. Finally, the globally optimal filter gain and consensus gain are obtained by minimizing the cost function, i.e., the mean squared error of the whole network. Depending on the indicator function, two extreme cases of the proposed multi-rate distributed consensus Kalman filter are analyzed. To reduce the computational complexity of the proposed multi-rate Kalman filter, a suboptimal filter is also suggested. Numerical experimental results further verify the effectiveness of the proposed distributed filter.