Shape-induced obstacle attraction and repulsion during dynamic locomotion

Abstract

Robots still struggle to dynamically traverse complex 3D terrain with many large obstacles, an ability required for many critical applications. Body–obstacle interaction is often inevitable and induces perturbation and uncertainty in motion that challenges closed-form dynamic modeling. Here, inspired by recent discovery of a terradynamic streamlined shape, we studied how two body shapes interacting with obstacles affect turning and pitching motions of an open-loop multi-legged robot and cockroaches during dynamic locomotion. With a common cuboidal body, the robot was attracted towards obstacles, resulting in pitching up and flipping-over. By contrast, with an elliptical body, the robot was repelled by obstacles and readily traversed. The animal displayed qualitatively similar turning and pitching motions induced by these two body shapes. However, unlike the cuboidal robot, the cuboidal animal was capable of escaping obstacle attraction and subsequent high pitching and flipping over, which inspired us to develop an empirical pitch-and-turn strategy for cuboidal robots. Considering the similarity of our self-propelled body–obstacle interaction with part–feeder interaction in robotic part manipulation, we developed a quasi-static potential energy landscape model to explain the dependence of dynamic locomotion on body shape. Our experimental and modeling results also demonstrated that obstacle attraction or repulsion is an inherent property of locomotor body shape and insensitive to obstacle geometry and size. Our study expands the concept and usefulness of terradynamic shapes for passive control of robot locomotion to traverse large obstacles using physical interaction. Our study is also a step in establishing an energy landscape approach to locomotor transitions.

Locomotor transitions in the potential energy landscape-dominated regime

“…To begin to understand complex physical interaction during locomotion in nature (figure 1d), we abstracted complex threedimensional terrain as a composition of diverse large obstacles (figure 1e) that present distinct locomotor challenges. These include compliant beams [50,51], rigid pillars [52], gaps [53] and bumps [54]. To enable systematic experiments (as in a wind or water tunnel), for each model terrain, we created a testbed that allowed controlled, systematic variation of obstacle properties such as stiffness [50], geometry [52] and size [53,54] (figure 3b).…”

“…These include compliant beams [50,51], rigid pillars [52], gaps [53] and bumps [54]. To enable systematic experiments (as in a wind or water tunnel), for each model terrain, we created a testbed that allowed controlled, systematic variation of obstacle properties such as stiffness [50], geometry [52] and size [53,54] (figure 3b). In addition, because animals and robots often flip over when traversing large obstacles [4,52,55], we studied strenuous ground self-righting in which existing appendages must be co-opted [55][56][57][58][59].…”

“…These physical principles operating on a rugged landscape leads to multi-pathway protein folding transitions. Although our locomotor-terrain interaction systems are macroscopic, self-propelled and far-fromequilibrium, their locomotor transitions display similar features, including stochasticity, multi-pathway transitions, kinetic energy fluctuation (from oscillatory self-propulsion) and favourability of some modes over others [4,[51][52][53][54][56][57][58][59], but with the addition of intelligence.…”

“…In addition, escape from a basin is more likely in directions on the landscape along which the barriers are lower [50,57]. Finally, during a transition, the system tends to transition to more favourable modes attracted to lower basins [50,52,57]. The animal's locomotor transitions also largely followed these principles during rapid, bandwidth-limited escape or emergency self-righting response [50][51][52][53][54]56,57].…”

“…In the potential energy landscape-dominated regime, physical principles and strategies that we discovered (figure 4c; electronic supplementary material, table S1) are applicable to a broad range of the parameter space of model systems. For example, obstacle attraction or repulsion is an inherent property of the locomotor shape and insensitive to pillar size and geometry [52]. Strategies that favour bump or gap traversal are applicable to a large range of bump heights [54] or gaps widths [53].…”