Abstract:This paper focuses on excitation collection and composite learning adaptive control design for uncertain nonlinear systems. By adopting the spectral decomposition technique, a linear regression equation is constructed to collect previously appeared excitation information, establishing a relationship between unknown parameters and the system's historical data. A composite learning term, developed using the linear regression equation, is incorporating into the Lyapunov-based parameter update law. In comparison to the existing results, all spectrums of previously appeared excitation information are collected, with the matrices in linear regression equation guaranteed to be bounded. This paper introduces concepts of excited and unexcited subspaces for analyzing the parameter estimation errors, and a novel Lyapunov function is developed for stability analysis. It is demonstrated that, without imposing any excitation condition, the state and excited parameter estimation error component converge to zero, while the unexcited component remains unchanged. Simulation results are provided to validate the theoretical findings.

What problem does this paper attempt to address?

The paper primarily addresses the issue of adaptive control in uncertain nonlinear systems, specifically focusing on how to design a composite learning adaptive controller without the need for Persistent Excitation (PE) conditions. Its core contribution lies in proposing a spectral decomposition-based excitation collection method and developing a new composite learning adaptive control scheme based on this method. Specifically, the objectives of the paper can be summarized as follows: 1. **Application of Spectral Decomposition Technology in Excitation Collection**: By adopting spectral decomposition technology, the paper designs different forgetting factors to handle previously collected excitation information in different directions. This ensures that all previously encountered excitation information can be fully collected to correct parameter estimates. Additionally, this method guarantees that the matrices involved in the linear regression equations remain bounded and do not weaken the information in weakly excited directions. 2. **Concept of Excited and Non-Excited Subspaces**: The concept of excited and non-excited subspaces is proposed to analyze the parameter estimation error vector. Using these concepts, a novel Lyapunov function is constructed to analyze the stability of the system under insufficient excitation. For first-order uncertain systems, the proposed adaptive control scheme ensures that the system state and the excited parameter estimation error components exponentially converge to zero, while the non-excited parameter estimation error components remain unchanged. 3. **Application to Higher-Order Systems**: Furthermore, the modified composite learning adaptive control scheme is applied to higher-order systems with mismatched uncertainties by introducing the backstepping technique. Even without imposing any excitation conditions, it ensures that the system state and the excited parameter estimation error components asymptotically converge to zero, while the non-excited components remain unchanged. The paper validates the effectiveness of the theoretical findings through simulation results and provides a summary in the conclusion section. This research offers a new approach to adaptive control of uncertain nonlinear systems, particularly suitable for practical applications where traditional excitation conditions are difficult to meet.

Composite Learning Adaptive Control without Excitation Condition

A Composite Adaptive Robust Control for Nonlinear Systems with Model Uncertainties

Model reference composite learning control without persistency of excitation

Composite Learning Fuzzy Control of Stochastic Nonlinear Strict-Feedback Systems

Composite Learning from Adaptive Backstepping Neural Network Control

Composite learning control with application to inverted pendulums

Composite Consensus Control And Cooperative Adaptive Learning

Composite Intelligent Learning Control of Strict-Feedback Systems with Disturbance.

Command-filtered backstepping neural network control with composite learning

Composite learning backstepping control with guaranteed exponential stability and robustness

Composite Learning Fuzzy Control of Uncertain Nonlinear Systems

Composite Energy Function-Based Spatial Iterative Learning Control in Motion Systems

Event-Triggered Adaptive Control of Uncertain Nonlinear Systems With Composite Condition

Composite Adaptive Control for Bilateral Teleoperation Systems without Persistency of Excitation

Adaptive control of A class of nonlinear systems with guaranteed parameter estimation: A concurrent learning based approach

Composite Neural Dynamic Surface Control of a Class of Uncertain Nonlinear Systems in Strict-Feedback Form

Globally Composite-Learning-Based Intelligent Fast Finite-Time Control for Uncertain Strict-Feedback Systems with Nonlinearly Periodic Disturbances

Composite Adaptive Control for Anti-Unwinding Attitude Maneuvers: An Exponential Stability Result Without Persistent Excitation

Least-squares Learning Control with Guaranteed Parameter Convergence

Optimal Robust Control of Nonlinear Systems with Unknown Dynamics via NN Learning with Relaxed Excitation

Online Recorded Data-Based Composite Neural Control of Strict-Feedback Systems With Application to Hypersonic Flight Dynamics