Abstract:The standard l(1)-norm regularization is recently introduced for impact force identification, but generally underestimates the peak force. Compared to l(1)-norm regularization, l(p)-norm (0 <= p < 1) regularization, with a nonconvex penalty function, has some promising properties such as enforcing sparsity. In the framework of sparse regularization, if the desired solution is sparse in the time domain or other domains, the under-determined problem with fewer measurements than candidate excitations may obtain the unique solution, i.e., the sparsest solution. Considering the joint sparse structure of impact force in temporal and spatial domains, we propose a general l(p)-norm (0 <= p < 1) regularization methodology for simultaneous identification of the impact location and force time-history from highly incomplete measurements. Firstly, a nonconvex optimization model based on l p-norm penalty is developed for regularizing the highly under-determined problem of impact force identification. Secondly, an iteratively reweighed l(1)-norm algorithm is introduced to solve such an under-determined and unconditioned regularization model through transforming it into a series of l(1)-norm regularization problems. Finally, numerical simulation and experimental validation including single-source and two-source cases of impact force identification are conducted on plate structures to evaluate the performance of l(p)-norm (0 <= p < 1) regularization. Both numerical and experimental results demonstrate that the proposed l(p)-norm regularization method, merely using a single accelerometer, can locate the actual impacts from nine fixed candidate sources and simultaneously reconstruct the impact force time-history; compared to the stateof-the-art l(1)-norm regularization, l(p)-norm (0 <= p < 1) regularization procures sufficiently sparse and more accurate estimates; although the peak relative error of the identified impact force using l(p)-norm regularization has a decreasing tendency as p is approaching 0, the results of l(p)-norm regularization with 0 <= p <= 1/2 have no significant differences.

Regularized nonlinear regression for simultaneously selecting and estimating key model parameters: Application to head-neck position tracking

Separable Nonlinear Least Squares Algorithm for Robust Kinematic Calibration of Serial Robots.

Regularized Nonlinear Regression with Dependent Errors and its Application to a Biomechanical Model

Identifiability Implies Robust, Globally Exponentially Convergent On-line Parameter Estimation: Application to Model Reference Adaptive Control

An L-BFGS-B approach for linear and nonlinear system identification under $\ell_1$- and group-Lasso regularization

Regularized Least-Mean-Square Algorithms

A Improve Direct Path Seeking Algorithm for L1/2 Regularization, with Application to Biological Feature Selection

Inertial Parameter Identification for Closed-Loop Mechanisms: Adaptation of Linear Regression for Coordinate Partitioning

Linear Programming-Based Sparse Kernel Regression with L1-Norm Minimization for Nonlinear System Modeling

L1 regularization facilitates detection of cell type-specific parameters in dynamical systems

A Weighted Least-Squares Method for Non-Asymptotic Identification of Markov Parameters from Multiple Trajectories

Parameter estimation of two classes of nonlinear systems with non-separable nonlinear parameterizations

Online Learning-Based Inertial Parameter Identification of Unknown Object for Model-Based Control of Wheeled Humanoids

Parameter Identification for Nonlinear State-Space Models of A Biological Network Via Linearization and Robust State Estimation

Two Stage Non-penalized Corrected Least Squares for High Dimensional Linear Models with Measurement error or Missing Covariates

Lp-Norm Regularization for Impact Force Identification from Highly Incomplete Measurements

Improving the Robustness of Variable Selection and Predictive Performance of Regularized Generalized Linear Models and Cox Proportional Hazard Models

A robust sparse identification method for nonlinear dynamic systems affected by non-stationary noise

Input Design for Regularized System Identification: Stationary Conditions and Sphere Preserving Algorithm

Joint State and Parameter Estimation of Nonlinear Dynamical Systems Through Alternating Optimization

Joint Track-to-track Association and Sensor Registration at the Track Level