Dynamic Legged Robots for Use in Multiple Regimes

Animals have demonstrated that legged locomotion can provide an efficient, rapid and robust means for traversing natural and artificial terrains and obstacles while employing several distinct locomotion modalities. This has led researchers to develop biologically-inspired, legged systems that improve the mobility of robotic platforms on rough terrain as compared to traditional wheeled and tracked systems. While several effective, dynamical legged robots have been developed, they still are a step behind their biological counterparts in terms of speed, efficiency, and, in particular, versatility. A contributing factor is that these platforms are designed and optimized to utilize a single locomotion modality and are not capable of traversing the variety of terrains found in the natural world, whereas many biological creatures are capable of multi-modal locomotion. This motivates the exploration of the principles that will enable the development of biologically-inspired, dynamical legged robots capable of utilizing multiple locomotion modalities. This dissertation focuses on several topics related to the development of biologically-inspired, multi-modal platforms. In particular, a focus is made towards the extension of a dynamic scaling method for dynamical legged systems, the assessment of dynamic stability metrics for use on experimental platforms and the design and characterization of the first legged, robotic platform capable of dynamical, biologically-inspired locomotion in multiple domains. The resulting insights from this work will foster an improved understanding of multi-modal robotic locomotion and provide a set of tools for designing and characterizing future robotic platforms.