Design principles for miniature rotary-wing hybrid-locomotion robots

Abstract

Hybrid-locomotion robots that utilize wheeled ground locomotion and rotary-wing flight hold great promise for increasing the mobility of miniature robots. Such robots improve upon ground-only robots in that a flight mode is available for use when otherwise impassable terrain or obstacles are encountered. This thesis presents designs for such robots and formulates principles for the design of similar robots. The design discussed herein is based on a two-wheeled ground robot. The robot tips itself on-end, allowing rotors to unfold from along the length of the robot's body. While the design principles were developed as a result of work on this robot, they are nevertheless applicable to other hybrid-locomotion robots utilizing a rotary-wing flight mode, regardless of the form they take. A predecessor utilized twin coaxial rotors, while the novel V2 design presented here is equipped with one main rotor and a tail rotor. Both designs use a set of arms, hinged near one side of the robot, to push on the ground, creating a moment about one of the wheels and lifting the other off the ground, ultimately positioning the robot's long axis vertically. Mechanisms for engaging drivetrains, folding/unfolding rotors, and unfolding a tail with a tail rotor (in the V2 design only), complete the transformation; the orientation-adjusting arms then serve as landing gear for the flight mode. Performance of the novel design improves upon the predecessor, but still leaves much to be desired. With a top speed of 0.3 m/s on flat terrain, a climbing capability of 19mm, and a lifetime of up to 2.5 hours on the ground, terrestrial performance is satisfactory. The flight mode has not been successfully demonstrated, but related testing shows that battery life in flight would be less than 5 minutes. Such performance metrics lead to the primary design principle: hybrid-locomotion robots with rotary-wing flight modes should have ground-focused designs to enable the terrestrial navigation of all expected environments. Such designs will reserve the flight mode for only unexpected obstacles or terrain, while spending the majority of their time in the much more energy-efficient ground mode. Four other principles have been formulated:Parts should perform as many functions as possible to minimize the quantity and weight of components. Drive systems for the two locomotion modes should be separate due to widely disparate requirements. The robot's flight hardware will always be much larger than is practical in the ground mode in environments of reasonable complexity, and thus the robot should collapse to a smaller size in the ground mode. The robot's flight mode components must be protected from the environment while in the ground mode. With design principles formulated, design changes can be explored. Impact tolerance of any meaningful amount proves to be infeasible due to the structural requirements of impact-tolerant components. Scaling the designs up is feasible, but significantly reduces the utility of the robot in cluttered and indoor environments. Scaling the designs down is largely impractical due to lack of actuator availability and the complexity of the design. Also, it would render the robot largely useless in outdoor environments, as terrain would be difficult in the ground mode, and wind makes flight at such scales infeasible. Alternate rotor configurations are possible, and a quad-rotor system may be superior to the designs discussed herein. The work in this thesis represents a first step - a proof of concept. Useful robots will require significantly more mechanical work. In addition, development would be required in both electronics and controls, and the robots could benefit greatly from the development of autonomous behaviors. The results to date have certainly been promising and merit further work.