Decoupled Prescribed Performance and Safe Formation Control of Multi-Agent Systems Under Input Constraints

Reliable formation control in real-world multi-agent systems is challenging due to the concurrent need to meet performance specifications, enforce safety constraints, and respect actuator limitations. While prescribed performance control (PPC) ensures bounded error evolution via predefined performance functions, incorporating safety and input constraints within this framework remains nontrivial. This paper develops a modular decoupled control architecture that integrates PPC and control barrier function (CBF) to enforce performance and safety. The fixed performance bound limitation in PPC control is overcome through the introduction of an auxiliary system that adaptively adjusts performance functions, thereby effectively enabling the quantification of performance degradation due to constraints while avoiding control singularities. To further mitigate conflicts between safety and performance, an online trajectory optimization module is designed to generate smooth and collision-free reference trajectories. The proposed approach is validated on a team of Crazyflie quadrotors navigating obstacle environments, demonstrating safe and accurate formation tracking under stringent constraints.