Speaker: 甘叠，中关村国家实验室

Title: Distributed Adaptive Estimation Algorithms for Sparse Stochastic Dynamic Systems

Language: Chinese

Time & Venue: 2024.12.11 09:00-09:30 思源楼415

Abstract: Sparse parameter identification is an essential issue that has received considerable attention in the fields of signal processing, machine learning, and systems control. When the system exhibits sparsity, the classical system identification theory often becomes inadequate. For sensor networks with randomly switching communication links, a compressed distributed Kalman filter algorithm based on the Markov switching topology is proposed. In this algorithm, each sensor first compresses the regression vector into a low-dimensional space, then generates distributed Kalman filter estimates of the compressed unknown parameters by the covariance intersection fusion rule, and finally recovers the original high-dimensional sparse unknown parameters using signal reconstruction algorithms. Under the compressed cooperative excitation condition, an upper bound of the estimation error is given. The above results are applicable to dynamic switching topologies and effectively improve the adaptability and robustness of the algorithm in the face of occasional communication failures.