Conrad P, Graichen K (2024)
Publication Type: Journal article
Publication year: 2024
Book Volume: 12
Pages Range: 153243-153252
DOI: 10.1109/ACCESS.2024.3480522
Self-triggered control aims to reduce resource utilization, particularly in networked systems, by sampling only when necessary to ensure a specific level of control performance. However, its potential in general nonlinear control applications, where computational efficiency is of importance, has not been fully explored. This paper introduces a self-triggering mechanism for nonlinear Model Predictive Control (MPC), extending its application beyond networked systems. The method maximizes the sampling time while accounting for the current sensitivity of the MPC cost function. A higher sensitivity in terms of a larger cost gradients, indicates that more frequent sampling is required due to increased susceptibility to disturbances and uncertainties. Asymptotic stability of the closed-loop system is guaranteed by imposing a decrease condition on the future optimal cost. Additionally, theoretical bounds on the sampling time are established, relating the cost reduction rate to the prediction horizon. To reduce the numerical complexity for the evaluation of this stability condition, a computationally efficient Taylor approximation of the future optimal cost is presented. This practical approach requires no user-defined triggering functions, uncertainty bounds, or additional parameters such as Lipschitz constants of the system or terminal regions. Two examples illustrate the method's ability to adapt to varying system dynamics while ensuring stability and preserving control performance.
APA:
Conrad, P., & Graichen, K. (2024). A sensitivity-based approach to self-triggered nonlinear model predictive control. IEEE Access, 12, 153243-153252. https://doi.org/10.1109/ACCESS.2024.3480522
MLA:
Conrad, Paulina, and Knut Graichen. "A sensitivity-based approach to self-triggered nonlinear model predictive control." IEEE Access 12 (2024): 153243-153252.
BibTeX: Download