Multi-robot Trajectory Generation for an Aerial Payload Transport System

Abstract

In this work, we consider the problem of planning safe, feasible trajectories for a team of quadrotors with slung-loads operating in an obstacle-free, three-dimensional workspace. We are particularly interested in generating dynamic trajectories—trajectories where robots’ payloads are allowed to swing in accordance with the system’s natural dynamics—for fast, agile, coordinated payload transportation. This capability is applicable to tasks such as construction, where a single crane performing sequential tasks could be replaced by multiple quadrotors performing tasks in parallel for increased efficiency. We model this problem as a labeled multi-robot planning problem, where robots must navigate payloads from given start positions to fixed, non-interchangeable goal positions. Our system presents three novel challenges: (1.) Each vehicle has eight degrees-of-freedom, significantly increasing the size of the team’s joint state space. (2.) Each vehicle is a nonlinear, order dynamical system with four degrees of under-actuation. (3.) Each vehicle is a multi-body system. We present a safe and complete Quadratic Programming solution and validate its practicality with experiments containing up to nine quadrotors.