LocoGear: Locomotion Analysis of Robotic Landing Gear for Multicopters

摘要

Nowadays, there are few unmanned aerial vehicles (UAVs) that are capable of landing on uneven surfaces or moving on the complex ground. A drone having both of these functionalities can significantly improve its performance in accomplishing difficult tasks, such as monitoring and exploring various types of terrain, searching for survivors, and delivering medical aid or repair kits in an unstructured dynamic environment, e.g., rescue operation following an earthquake. This article presents a motion analysis of the gait of an adaptive landing platform for multicopters. This landing gear has four robotic legs facilitated with torque sensors in the knee joints. Though it had been developed to provide adaptive landing on uneven surfaces, we decided to exploit the system for terrain locomotion purposes. For this purpose, we have analyzed the center of mass trajectory to evaluate the possible trajectory of the robot. Based on the result of this experimental analysis, we propose a LocoGear, a novel algorithm for locomotion of UAV equipped with the robotic landing gear. The servomotor torques were estimated using Lagrangian dynamic formulation and kinetostatic methods. We conducted experiments to verify the LocoGear approach based on feedforward control that proves the capability of landing gear to move along the desired trajectory. We achieved the robot motion along the straight line with the standard deviation of length in each step of 9.8 mm and standard deviation in yaw of 9°.