I. Introduction
In the real world physical systems are generally constrained, and the constraints range from their own limitations to the workspace and task requirements. The violation of these constraints has great impacts on the safety of physical systems; see [1] and references therein for more examples. In this respect, safety constraints are essential for physical systems, and impose strict requirements on system states/inputs to avoid system damages and economic losses. With the safety constraints in the priority place, physical systems are expected to accomplish some desired tasks, including stabilization, tracking and even temporal logic tasks [2], [3], [4]. From the control perspective, numerous approaches have been proposed and applied in the literature to deal with different tasks under safety constraints [1], [5], [6]. For instance, model predictive control (MPC) approach has been implemented to deal with motion planning and control problems of marine vehicles [6]; different energy-based functions, such as artificial potential functions, control Lyapunov and barrier functions, have been combined and mixed up to address formation, tracking and reach-avoid tasks of mobile robots [2], [5], [7].