Hybrid RRT*-Gauss Pseudospectral Trajectory Planning Method for eVTOL Application

The complexity of trajectory design for electric vertical takeoff and landing aircraft (eVTOL) originates the nonlinear aircraft dynamics, path constraints, and coupling characteristics between translational and rotational motions, which in turn slows down the solving process. To this end, this paper proposes a hybrid trajectory planning method composed of two steps. Firstly, the Rapidly-exploring Random Tree* (RRT*) planner is used to generate a feasible flight route in the presence of obstacles. Thereafter, the cube B-spline curve and differential flatness are employed to generate a smooth trajectory based on the waypoints from RRT*, which serves as the initial guess of Gauss pseudo-spectral method. As such, the advantages of global and numerical trajectory methods are combined, which simultaneously ensures that the route conforms to the aircraft dynamics and control input conditions, and the optimality of the trajectories. A test case is presented to demonstrate the effectiveness of the approach for designing trajectory of a multi-rotor eVTOL, which shows the hybrid RRT*-Gauss pseudo-spectral algorithm performs well in computation time, maximum relative error and control resource consumption.