轻量级深度感知引导的模块化微型无人机主动对接OA

Aiming at the constraints of limited payload capacity and limited onboard computational resources in modular micro air vehicles(MAVs),a visual servoing-based active docking control method was proposed by using a multi-zone time-of-flight(ToF)sensor.Firstly,based on the decoupled dynamic model of a quadrotor UAV,a position-attitude decoupled four-loop cascaded PID low-level controller was designed.Secondly,the ranging principle and characteristics of the VL53L5CX multi-zone ToF sensor were systematically analyzed and tested.Thirdly,a lightweight target recognition algorithm based on region-growing clustering and a decision-tree control strategy based on region partitioning were designed.Finally,the proposed method was validated in both a PyBullet simulation environment and a physical platform based on the Crazyflie.The results show that in simulation,the docking success rate exceeds 90%within an inter-module distance of less than 0.5 m.In real flight experiments,the docking success rate reaches over 80%within the same distance range.It verifies the effectiveness and feasibility of the proposed method on micro UAV platforms with limited computational resources and payload capacity.