On hybrid modeling and control of a multi-propeller multifunction aerial robot with flying-walking locomotion

Flying-walking locomotion is widely adopted by birds. It increases the agility and activities of birds, but has not been reported in the robotics research area. This paper is trying to address the problem of flying-walking locomotion with a multi-propeller multifunction aerial robot (MMAR). The dynamics of hybrid flying-walking locomotion is rather complex since it needs to consider the modeling of aerial robots contacting with the environment. By dividing the flying-walking locomotion into several motion modes, hybrid modeling framework is employed to model the dynamics of the overall flying-walking locomotion maneuver. Contact dynamics between the robot and the ground in the overall maneuver is derived from the constrained Lagrangian. Furthermore, the models of different modes are analyzed for the control purposes. Based on the dynamic model, an optimal planning algorithm is proposed to minimize the interaction between the legs and main-body of MMAR during the motion. Several composite controllers are designed to stabilize the motion of the main-body and the motion of the legs in different modes. Such controllers are designed using trajectory linearization control approach and computed-torque method. Simulation tests are presented to show the feasibility of proposed flying-walking locomotion.