Design of Vector Field for Different Subphases of Gait and Regeneration of Gait Pattern

In this paper, we have designed the vector fields (VFs) for all the six joints (hip, knee, and ankle) of a bipedal walking model. The bipedal gait is the manifestation of temporal changes in the six joints angles, two each for hip, knee, and ankle values and it is a combination of seven different discrete subphases. Developing the correct joint trajectories for all the six joints was difficult from a purely mechanics-based model due to its inherent complexities. To get the correct and exact joint trajectories, it is very essential for a modern bipedal robot to walk stably. By designing the VF correctly, we are able to get the stable joint trajectory ranges and able to reproduce angle ranges from theses designed VFs. This is purely a data driven computational modeling approach, which is based on the hypothesis that morphologically similar structure (human-robot) can adopt similar gait patterns. To validate the correctness of the design, we have applied all the possible combination of joint trajectories to HOAP-2 bipedal robot, which could walk successfully maintaining its stability. The VF provides joint trajectories for a particular joint. The results show that our data driven computational model is able to provide the correct joints angle ranges, which are stable.

I. Introduction

The modern robot industries have given much needed boost to bipedal robot development, which can work outside constrained laboratory environment [1]. The bipedal locomotion is considered to be a complex hybrid system [2]. It is the manifestation of changes in the six joints, two each for hip, knee, and ankle [3]. The whole bipedal walk is divided into seven different discrete subphases [4]. The earlier developed bipedal models are not perfect as they had flat foot and bending knees, which obstruct the bipedal not to perform all the seven phases of walk [5]. To mimic the human similar walk, we have designed the vector field (VF) from the human gait collected data.