|SUBJECT:||M.S. Thesis Presentation|
|TIME:||Friday, November 9, 2012, 9:00 a.m.|
|TITLE:||Exploring Lift-Off Dynamics In a Jumping Robot|
|COMMITTEE:||Dr. Harvey Lipkin, Co-Chair (ME)
Dr. Daniel I. Goldman, Co-Chair (PHYS)
Dr. Alexander Alexeev (ME)
Terrestrial organisms and robots are tasked with traversing complex environments in ways that conventional wheeled vehicles are unable to accomplish, and do so by effectively deforming the appendages of their bodies. Jumping is a prevalent mode of locomotion for various animals. It is used to reach higher places and as survival and predatory behavior. And with the advent of robots designed with inspiration from nature’s excellent jumpers, there is an imperative to understand the fundamental factors and mechanics that optimize jumping performance. Animals are known to amplify their jumping power with effective use of compliant structures by performing catapult jumps, squat jumps and countermovements. Certain animals even utilize a variant of counter-movement known as the stutter jump, where the jump is preceded by a small initial hop. While biologically inspired robots have taken a cue from nature to produce hopping gaits, catapults and even basic squat jumps, systematic studies of which mathematically defined movement trajectories maximize jumping performance are relatively scarce. This thesis presents a systematic study with a robotic jumping template based on a 1D variant of the spring-loaded inverted pendulum (SLIP) model of jumping to uncover the movement strategies for maximum height jumping when a catapult mechanism is unavailable. The robot is a mass-spring arrangement with an actuated mass which we actuate sinusoidally. In concert with simulation, we systematically vary sine wave parameters of the command trajectory for robotic mass actuation to uncover which values provide the best jump performance. A countermovement produces a stutter jump that is optimal at a frequency lower than the natural frequency, f0, and a squat jump produces a single jump that is comparably optimal but a faster frequency, which requires more internal power. An analysis of the dynamical model reveals how optimal lift-off results from non-resonant transient dynamics.