Immersion and Invariance Disturbance Observer-Based Fault-Tolerant Formation Control of Multi-UAVs under Safety Constraints

This paper investigates safety-constrained control problem in multi-UAV formation systems considering both unknown time-varying disturbances and actuator faults. To enhance the disturbance rejection capability of the system, an immersion and invariance disturbance observer is developed to accurately estimate disturbances in real-time. Then, an adaptive sliding mode fault-tolerant formation controller is proposed to ensure consensus in multi-UAV systems, which incorporates a neural network to compensate for system uncertainties. Building on the fault-tolerant formation controller, high-order control barrier functions are incorporated into a locally formulated quadratic programming (QP) for each UAV to ensure inter-UAV and obstacle avoidance safety, which minimally adjusts the fault-tolerant formation control input only when safety constraints become active. Numerical simulations and real-time experiments demonstrate the effectiveness and safety performance of the proposed control strategy.