A Robust Diffusion Minimum Kernel Risk-Sensitive Loss Algorithm over Multitask Sensor Networks

Sensors - Tập 19 Số 10 - Trang 2339
Xinyu Li1,2, Qing Shi2, Shuangyi Xiao2, Shukai Duan1,2, Feng Chen1,2
1College of Artificial Intelligence, Southwest University, Chongqing, 400715, China
2Key Laboratory of Nonlinear Circuits and Intelligent Information Processing, and College of Electronic and Information Engineering, Southwest University, and Chongqing Collaborative Innovation Center for Brain Science, Chongqing 400715, China

Tóm tắt

Distributed estimation over sensor networks has attracted much attention due to its various applications. The mean-square error (MSE) criterion is one of the most popular cost functions used in distributed estimation, which achieves its optimality only under Gaussian noise. However, impulsive noise also widely exists in real-world sensor networks. Thus, the distributed estimation algorithm based on the minimum kernel risk-sensitive loss (MKRSL) criterion is proposed in this paper to deal with non-Gaussian noise, particularly for impulsive noise. Furthermore, multiple tasks estimation problems in sensor networks are considered. Differing from a conventional single-task, the unknown parameters (tasks) can be different for different nodes in the multitask problem. Another important issue we focus on is the impact of the task similarity among nodes on multitask estimation performance. Besides, the performance of mean and mean square are analyzed theoretically. Simulation results verify a superior performance of the proposed algorithm compared with other related algorithms.

Từ khóa


Tài liệu tham khảo

Sayed, 2014, Adaptation, learning, and optimization over networks, Found. Trends Mach. Learn., 7, 311, 10.1561/2200000051

Lorenzo, P.D., Barbarossa, S., and Sayed, A.H. (September, January 29). Bio-inspired swarming for dynamic radio access based on diffusion adaptation. Proceedings of the 2011 19th European Signal Processing Conference (EUSIPCO), Barcelona, Spain.

Chen, 2010, Distributed collaborative control for industrial automation with wireless sensor and actuator networks, IEEE Trans. Ind. Electron., 57, 4219, 10.1109/TIE.2010.2043038

Sayed, 2013, Diffusion strategies for adaptation and learning over networks, IEEE Signal Process. Mag., 30, 155, 10.1109/MSP.2012.2231991

Fax, 2007, Consensus and cooperation in networked multi-agent systems, Proc. IEEE, 95, 215, 10.1109/JPROC.2006.887293

Kar, 2009, Distributed consensus algorithms in sensor networks: Link failures and channel noise, IEEE Trans. Signal Process., 57, 355, 10.1109/TSP.2008.2007111

Wang, J., Peng, D., Jing, Z., and Chen, J. (2018). Consensus-Based Filter for Distributed Sensor Networks with Colored Measurement Noise. Sensors, 18.

Nedic, 2009, Distributed subgradient methods for multiagent optimization, IEEE Trans. Autom. Control, 54, 48, 10.1109/TAC.2008.2009515

Nedic, 2001, Incremental subgradient methods for nondifferentiable optimization, SIAM J. Optim., 12, 109, 10.1137/S1052623499362111

Rabbat, 2005, Quantized incremental algorithms for distributed optimization, IEEE J. Sel. Areas Commun., 23, 798, 10.1109/JSAC.2005.843546

Lopes, 2007, Incremental adaptive strategies over distributed networks, IEEE Trans. Signal Process., 48, 223

Chen, 2012, Diffusion adaptation strategies for distributed optimization and learning over networks, IEEE Trans. Signal Process., 60, 4289, 10.1109/TSP.2012.2198470

Cattivelli, 2010, Diffusion LMS strategies for distributed estimation, IEEE Trans. Signal Process., 58, 1035, 10.1109/TSP.2009.2033729

Zhao, 2012, Performance limits for distributed estimation over LMS adaptive networks, IEEE Trans. Signal Process., 60, 5107, 10.1109/TSP.2012.2204985

Tu, 2012, Diffusion strategies outperform consensus strategies for distributed estimation over adaptive networks, IEEE Trans. Signal Process., 60, 6217, 10.1109/TSP.2012.2217338

Chen, 2018, Diffusion generalized maximum correntropy criterion algorithm for distributed estimation over multitask network, Digit. Signal Process., 81, 16, 10.1016/j.dsp.2018.02.008

Liu, 2012, Distributed estimation over complex networks, Inf. Sci., 197, 91, 10.1016/j.ins.2012.02.008

Chen, 2017, Broken-motifs diffusion LMS algorithm for reducing communication load, Signal Process., 197, 91

Chen, 2019, Complementary performance analysis of general complex-valued diffusion LMS for noncircular signals, Signal Process., 160, 237, 10.1016/j.sigpro.2019.02.026

Cattivelli, F.S., Lopes, C.G., and Sayed, A.H. (2007, January 17–20). A diffusion RLS scheme for distributed estimation over adaptive networks. Proceedings of the 2007 IEEE 8th Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Helsinki, Finland.

Cattivelli, 2008, Diffusion recursive leasts-quares for distributed estimation over adaptive networks, IEEE Trans. Signal Process., 56, 1865, 10.1109/TSP.2007.913164

Gao, 2017, Kernel Least Mean p-Power algorithm, IEEE Signal Process. Lett., 24, 996, 10.1109/LSP.2017.2702714

Shao, X., Chen, F., Ye, Q., and Duan, S. (2017). A Robust Diffusion Estimation Algorithm with Self-Adjusting Step-Size in WSNs. Sensors, 17.

Wen, 2013, Diffusion least-mean P-power algorithms for distributed estimation in alpha-stable noise environments, Electron. Lett., 49, 1355, 10.1049/el.2013.2331

Ni, 2016, Diffusion sign-error LMS algorithm: formulation and stochastic behavior analysis, Signal Process., 128, 142, 10.1016/j.sigpro.2016.03.022

Liu, 2007, Correntropy: Properties and applications in non-Gaussian signal processing, IEEE Trans. Signal Process., 55, 5286, 10.1109/TSP.2007.896065

Chen, 2017, Maximum correntropy Kalman filter, Automatica, 76, 70, 10.1016/j.automatica.2016.10.004

Chen, 2016, Generalized correntropy for robust adaptive filtering, IEEE Trans. Signal Process., 64, 3376, 10.1109/TSP.2016.2539127

Chen, 2017, Kernel Risk-Sensitive Loss: Definition, Properties and Application to Robust Adaptive Filtering, IEEE Trans. Signal Process., 65, 2888, 10.1109/TSP.2017.2669903

Chen, 2015, Diffusion LMS over multitask networks, IEEE Trans. Signal Process., 63, 2733, 10.1109/TSP.2015.2412918

Chen, 2013, Distributed Pareto optimization via diffusion strategies, IEEE J. Sel. Top. Signal Process., 7, 205, 10.1109/JSTSP.2013.2246763

Zhao, X., and Sayed, A.H. (2012, January 28–30). Clustering via diffusion adaptation over networks. Proceedings of the 2012 3rd International Workshop on Cognitive Information Processing (CIP), Parador de Baiona, Spain.

Zhao, 2015, Distributed clustering and learning over networks, IEEE Trans. Signal Process., 63, 3285, 10.1109/TSP.2015.2415755

Chen, J., Richard, C., Hero, A.O., and Sayed, A.H. (2014, January 21–24). Diffusion LMS for mu1titask problems with overlapping hypothesis subspaces. Proceedings of the 2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP), Reims, France.

Bogdanovic, N., Plata-Chaves, J., and Berberidis, K. (2014, January 4–9). Distributed diffusion-based LMS for node-specific parameter estimation over adaptive networks. Proceedings of the 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Florence, Italy.

Chen, 2014, Multitask diffusion adaptation over networks, IEEE Trans. Signal Process., 62, 4129, 10.1109/TSP.2014.2333560

Nassif, 2016, Proximal multitask learning over networks with sparsity-inducing coregularization, IEEE Trans. Signal Process., 64, 6329, 10.1109/TSP.2016.2601282

Ma, 2016, Diffusion maximum correntropy criterion algorithms for robust distributed estimation, Digit. Signal Process., 58, 10, 10.1016/j.dsp.2016.07.009

Sayed, A.H. (2008). Adaptive Filters, Wiley.

Haykin, S. (2002). Adaptive Filter Theory, Prentice-Hall.

Kelley, C.T. (1999). Iterative Methods for Optimization, SIAM.

Abadir, K.M., and Magnus, J.R. (2005). Matrix Algebra, Cambridge University Press.

Chan, 2004, A recursive least M-estimate algorithm for robust adaptive filtering in impulsive noise: fast algorithm and convergence performance analysis, IEEE Trans. Signal Process., 52, 975, 10.1109/TSP.2004.823496

Sayed, 2016, Robust adaptation in impulsive noise, IEEE Trans. Signal Process., 64, 2851, 10.1109/TSP.2016.2535239