Adaptive Backstepping Multi-Agent Control of Multiple UAVs Conveying Load: A Graph Rigidity Approach

Abstract

Quadrotor Unmanned Aerial Vehicles (UAVs) are highly nonlinear and underactuated in nature as they possess fewer control inputs than control variables. A combination multiple UAVs conveying a single load is more complex but offers some advantages due to it's complimentary nature, as such, heavier loads can be conveyed compared to using single UAVs. Consequently, the need to avoid inter-agent collision necessitates the use of multi-agent control strategies which takes the presence of load disturbances into account. Therefore, in this work a Multi-agent UAV-Load system comprising dual-UAV with a single load is broken into altitude/position and attitude subsystem dynamics. To ensure that formation is maintained in a desired path, using rigid graph theory, an Adaptive Backstepping Multi-agent Control, based on Lyapunov theory is designed for the position subsystem in the presence of load disturbances. Furthermore, another Backstepping Finite-time Sliding Mode Control is designed for the attitude subsystem which is verified by simulation.