Modeling, Simulation and Control of a Propeller-aided legged robot

Abstract

Legged robots are very common for harsh terrain navigation, but struggle with big obstacles, having to inefficiently move around them. flying vehicles, in particular multirotors, don’t have this problem thanks to their 6-DOF motion capabilities, but are less energy ef- ficient having to counteract the force of gravity, leading to shorter mission durations. This Thesis introduces a design approach towards the combination of the two families of robot. The modeling, control strategy and state estimation of a legged robot aided by propellers are introduced. The propellers provide jumping and air control capa- bilities to the robot. An Extended Kalman Filter is implemented to address the noisy Inertial Measurement Units and Global Positioning System sensors used for the robot. A simple estimation and control loop of the robot is described. The robot’s jumping capabilities are demonstrated inside of a simulation.

Legged robots are very common for harsh terrain navigation, but struggle with big obstacles, having to inefficiently move around them. flying vehicles, in particular multirotors, don’t have this problem thanks to their 6-DOF motion capabilities, but are less energy ef- ficient having to counteract the force of gravity, leading to shorter mission durations. This Thesis introduces a design approach towards the combination of the two families of robot. The modeling, control strategy and state estimation of a legged robot aided by propellers are introduced. The propellers provide jumping and air control capa- bilities to the robot. An Extended Kalman Filter is implemented to address the noisy Inertial Measurement Units and Global Positioning System sensors used for the robot. A simple estimation and control loop of the robot is described. The robot’s jumping capabilities are demonstrated inside of a simulation.