Lei Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Lei Li has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Shiyuan Wang
Southwest University
6 shared publications - Yunfei Zheng
Southwest University
5 shared publications - Xiao Liu
Southwest University
4 shared publications - Mingjing Cui
Southwest University
3 shared publications - Yi-Fei Pu
College of Computer Science, Sichuan University, Chengdu 610065, China
3 shared publications - Dongyuan Lin
College of Electronics and Information Engineering, Shenzhen University, Shenzhen, China
2 shared publications
Latest Publications
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 72 (4) , 9742-9757
KalmanKAN: A Numerically Stable Neural Network-Aided State Estimation Method
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Consumer Electronics
- Vol. 72 (4) , 9742-9757
KalmanKAN: A Numerically Stable Neural Network-Aided State Estimation Method
- 26 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Signal Processing
- Vol. 74 (1053587X) , 545-560
Treating the Filter Weights as Learnable Functions: An Efficient Nonlinear Filtering Framework and Its Adaptive Algorithms
Adaptive filters, constrained by a linear filtering framework, often struggle with nonlinear modeling in complex processes. Kernel adaptive filters (KAFs) offer a promising solution by mapping input signals into diverse feature spaces. However, their computational efficiency and filtering accuracy may still not meet the demands of practical applications. To this end, based on Kolmogorov-Arnold (KA) representation theorem, this paper proposes a novel nonlinear filtering framework by treating filter weights as learnable functions. Specifically, the learnable functions are represented as a linear combination of multiple Gaussian basis functions with different centers. To determine the coefficients that define these learnable functions, the weight-learning-based least mean square (WL-LMS) and weight-learning-based recursive least squares (WL-RLS) algorithms are further proposed based on minimum mean square error (MMSE). In addition, to ensure convergence and assess the steady-state and transient performance of proposed algorithms, a thorough theoretical analysis of the convergence conditions and excess mean square error (EMSE) is provided. Finally, linear-in-parameters system identifications validate the correctness of theoretical analysis, while chaotic time-series prediction and nonlinear system identifications demonstrate the superiorities of the proposed WL-LMS and WL-RLS algorithms.
- 26 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Signal Processing
- Vol. 74 (1053587X) , 545-560
Treating the Filter Weights as Learnable Functions: An Efficient Nonlinear Filtering Framework and Its Adaptive Algorithms
Adaptive filters, constrained by a linear filtering framework, often struggle with nonlinear modeling in complex processes. Kernel adaptive filters (KAFs) offer a promising solution by mapping input signals into diverse feature spaces. However, their computational efficiency and filtering accuracy may still not meet the demands of practical applications. To this end, based on Kolmogorov-Arnold (KA) representation theorem, this paper proposes a novel nonlinear filtering framework by treating filter weights as learnable functions. Specifically, the learnable functions are represented as a linear combination of multiple Gaussian basis functions with different centers. To determine the coefficients that define these learnable functions, the weight-learning-based least mean square (WL-LMS) and weight-learning-based recursive least squares (WL-RLS) algorithms are further proposed based on minimum mean square error (MMSE). In addition, to ensure convergence and assess the steady-state and transient performance of proposed algorithms, a thorough theoretical analysis of the convergence conditions and excess mean square error (EMSE) is provided. Finally, linear-in-parameters system identifications validate the correctness of theoretical analysis, while chaotic time-series prediction and nonlinear system identifications demonstrate the superiorities of the proposed WL-LMS and WL-RLS algorithms.
- 15 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-14
Robust Nonlinear Fractional-Order Kalman Filters for False Data Injection Attacks
Maximum correntropy fractional-order central difference Kalman filter (MCFCDKF) is capable of addressing the state estimation problem for nonlinear fractional-order systems in non-Gaussian noise environments. However, its performance significantly degrades under false data injection (FDI) attacks, which intentionally distort measurement data. The key challenges include modeling attack uncertainties, distinguishing attacks from inherent non-Gaussian noises, and ensuring robust estimation accuracy. To tackle the challenges posed by FDI attacks and non-Gaussian noises, this paper first reconstructs the fractional-order measurement model using Bernoulli random variables. We then propose a robust maximum correntropy fractional-order central difference Kalman filter (RMCFCDKF) based on the maximum correntropy criterion (MCC). Additionally, by leveraging the contraction mapping theorem, we establish sufficient conditions for the convergence of the RMCFCDKF algorithm, ensuring the existence and uniqueness of the fixed-point iterative solution. Furthermore, an adaptive kernel width strategy is proposed and integrated into the RMCFCDKF algorithm, leading to the development of the adaptive RMCFCDKF (ARMCFCDKF) with no requirement for manual selection of kernel widths. Finally, evaluations on both a synthetic and a public real-world battery dataset, under FDI attacks and non-Gaussian noises, demonstrate the effectiveness of the proposed algorithms, with ARMCFCDKF reducing the average root mean square error (ARMSE) compared to MCFCDKF by 54.9% and 43.4%, respectively.
- 15 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-14
Robust Nonlinear Fractional-Order Kalman Filters for False Data Injection Attacks
Maximum correntropy fractional-order central difference Kalman filter (MCFCDKF) is capable of addressing the state estimation problem for nonlinear fractional-order systems in non-Gaussian noise environments. However, its performance significantly degrades under false data injection (FDI) attacks, which intentionally distort measurement data. The key challenges include modeling attack uncertainties, distinguishing attacks from inherent non-Gaussian noises, and ensuring robust estimation accuracy. To tackle the challenges posed by FDI attacks and non-Gaussian noises, this paper first reconstructs the fractional-order measurement model using Bernoulli random variables. We then propose a robust maximum correntropy fractional-order central difference Kalman filter (RMCFCDKF) based on the maximum correntropy criterion (MCC). Additionally, by leveraging the contraction mapping theorem, we establish sufficient conditions for the convergence of the RMCFCDKF algorithm, ensuring the existence and uniqueness of the fixed-point iterative solution. Furthermore, an adaptive kernel width strategy is proposed and integrated into the RMCFCDKF algorithm, leading to the development of the adaptive RMCFCDKF (ARMCFCDKF) with no requirement for manual selection of kernel widths. Finally, evaluations on both a synthetic and a public real-world battery dataset, under FDI attacks and non-Gaussian noises, demonstrate the effectiveness of the proposed algorithms, with ARMCFCDKF reducing the average root mean square error (ARMSE) compared to MCFCDKF by 54.9% and 43.4%, respectively.
- 29 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 32 (10709908) , 1995-1999
Steady-State Performance Analysis of the Nearest Kronecker Product Decomposition Based LMS Adaptive Algorithm
In order to address issues such as convergence rate, stability, and computational complexity caused by the identification of long length impulse response systems, an effective nearest Kronecker product (NKP) decomposition strategy has been introduced and extended to various adaptive filters in recent years. However, the theoretical performance of the NKP decomposition-based adaptive filtering algorithms has not been thoroughly analyzed in these studies. In this letter, we focus on analyzing the steady-state performance of the NKP-based least mean square (NKP-LMS) algorithm and presents the theoretical upper bound of the step-size. Finally, simulation results confirm the precision of the theoretical assessment of the NKP-LMS algorithm and highlight its benefits in low-rank system identification.
- 29 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 32 (10709908) , 1995-1999
Steady-State Performance Analysis of the Nearest Kronecker Product Decomposition Based LMS Adaptive Algorithm
In order to address issues such as convergence rate, stability, and computational complexity caused by the identification of long length impulse response systems, an effective nearest Kronecker product (NKP) decomposition strategy has been introduced and extended to various adaptive filters in recent years. However, the theoretical performance of the NKP decomposition-based adaptive filtering algorithms has not been thoroughly analyzed in these studies. In this letter, we focus on analyzing the steady-state performance of the NKP-based least mean square (NKP-LMS) algorithm and presents the theoretical upper bound of the step-size. Finally, simulation results confirm the precision of the theoretical assessment of the NKP-LMS algorithm and highlight its benefits in low-rank system identification.
- 18 November 2022
- Published by Elsevier in Neurocomputing
- Vol. 520, 331-341
A class of augmented complex-value FLANN adaptive algorithms for nonlinear systems
- 18 November 2022
- Published by Elsevier in Neurocomputing
- Vol. 520, 331-341
A class of augmented complex-value FLANN adaptive algorithms for nonlinear systems
- 26 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 29 (10709908) , 1853-1857
Performance Analysis of Fractional-Order Adaptive Filtering Algorithm and Its Improvement
Considering the $ \alpha$ -stable input signals and noises, several robust fractional-order adaptive filtering algorithms have been proposed in recent years. However, the mean-square performance analysis of the FoAF algorithm has not been well studied in the literature. Toward that end, this letter provide a theoretical mean-square performance analysis of the FoAF algorithm, involving transient and steady-state behavior. Furthermore, an improved fractional-order adaptive filtering algorithm is derived by utilizing the M-estimate technique. Simulation experiments verify the results of the theoretical analysis and demonstrate the advantages of the proposed algorithm in heavy-tailed non-Gaussian environments.
- 26 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 29 (10709908) , 1853-1857
Performance Analysis of Fractional-Order Adaptive Filtering Algorithm and Its Improvement
Considering the $ \alpha$ -stable input signals and noises, several robust fractional-order adaptive filtering algorithms have been proposed in recent years. However, the mean-square performance analysis of the FoAF algorithm has not been well studied in the literature. Toward that end, this letter provide a theoretical mean-square performance analysis of the FoAF algorithm, involving transient and steady-state behavior. Furthermore, an improved fractional-order adaptive filtering algorithm is derived by utilizing the M-estimate technique. Simulation experiments verify the results of the theoretical analysis and demonstrate the advantages of the proposed algorithm in heavy-tailed non-Gaussian environments.
- 17 June 2022
- Published by Springer Nature in Circuits, Systems, and Signal Processing
- Vol. 41 (10) , 5785-5806
Widely Linear Complex-Valued Least Mean M-Estimate Algorithms: Design and Performance Analysis
- 17 June 2022
- Published by Springer Nature in Circuits, Systems, and Signal Processing
- Vol. 41 (10) , 5785-5806
Widely Linear Complex-Valued Least Mean M-Estimate Algorithms: Design and Performance Analysis