Cloud-Based Computational Data-Enabled Predictive Control
Li Dai,Teng Huang,Runze Gao,Yuan Zhang,Yuanqing Xia
DOI: https://doi.org/10.1109/jiot.2022.3194945
IF: 10.6
2022-01-01
IEEE Internet of Things Journal
Abstract:This article considers the data-driven optimal control problem for unknown linear systems subject to system constraints from a computation efficiency improvement perspective. We propose a novel Computational Data-enabled Predictive Control (CDeePC) algorithm in a cloud environment, built on a recent work DeePC [1] . First, massive input–output data samples are precollected to describe the system input/output behavior through a behavioral systems theory approach, which might result in a large-scale online optimization problem with high dimensional decision variables. To solve it, in CDeePC, a multiblock alternating direction method of multipliers (ADMMs) algorithm is then employed, in which a solution to the original large problem can be calculated by solving iteratively a set of small subproblems. Next, to further improve the robustness of the algorithm, an online feedback CDeePC (OCDeePC) algorithm is proposed by utilizing real-time data samples to better capture the system characterization. Two analytic inversion formulas are derived for fast computation of time-varying controller parameters based on the result at the previous time. After that, we discuss the computational complexity and provide the convergence proof of proposed algorithms. Finally, after designing a cloud-based control architecture and formulating the iterative scheme to compute control actions as a workflow, we construct and deploy the proposed controllers as a service in the cloud. A case study of tracking control of wheeled mobile robots is provided to illustrate the efficacy of the proposed algorithms.