A Robust Low-Complexity Control Framework for Uncertain Nonlinear Systems With Hard and Soft Output Constraints

• 成果类型: Article
• 署名作者: Zuo, Huang; Tao, Jie; Lin, Ming; Zhao, Yao; Liu, Yong-Hua; Lu, Renquan; Su, Chun-Yi
• 署名单位: Guangdong University of Technology; Guangdong University of Technology
• 刊物名称: IEEE TRANSACTIONS ON AUTOMATIC CONTROL
• ISSN/ISSBN: 0018-9286
• DOI: 10.1109/TAC.2026.3672070
• 发表日期: 2026
• 关键词: PRESCRIBED PERFORMANCE CONTROL; funnel control; Barrier functions; adaptive-control; tracking

摘要

This article investigates the control problem of uncertain nonlinear systems subject to both hard and soft output constraints. The hard safety constraint is treated as strictly inviolable, whereas the prescribed tracking performance is formulated as an adjustable soft constraint. When conflicts arise between these constraints, two direct real-time soft-boundary modification strategies are proposed to ensure strict compatibility with the hard constraint, in contrast to existing approaches that rely on indirect adjustments via auxiliary dynamics. The first strategy employs LogSumExp smooth max/min operators to enforce an approximate violation tolerance that remains strictly above the minimum level required by the hard constraint, yielding a concise and effective conflict resolution mechanism. The second strategy introduces a multievent-triggered scheme equipped with smooth transition functions that achieves finite-time decoupling of the hard and soft constraints during conflict intervals, ensuring exact soft-boundary regulation while rigorously preserving hard-constraint priority. A modular initialization mechanism is further designed to eliminate compatibility requirements commonly imposed in prior works, such as the need for simultaneous satisfaction of both constraints at the initial moment. Building on these direct modification strategies, a robust and low-complexity controller is constructed using a bilateral barrier function to enforce both the hard constraint and the modified soft constraint. Simulation results validate the effectiveness and advantages of the proposed framework.