Generating real-time trajectories for a planar biped robot crossing a wide ditch with landing uncertainties

Abstract

SUMMARYDitch crossing is one of the essential capabilities required for a biped robot in disaster management and search and rescue operations. Present work focuses on crossing a wide ditch with landing uncertainties by an under-actuated planar biped robot with five degrees of freedom. We consider a ditch as wide for a robot when the ankle to ankle stretch required to cross it is at least equal to the leg length of the robot. Since locomotion in uncertain environments requires real-time planning, in this paper, we present a new approach for generating real-time joint trajectories using control constraints not explicitly dependent on time, considering impact, dynamic balance, and friction. As part of the approach, we introduce a novel concept called the point of feasibility for bringing the biped robot to complete rest at the end of ditch crossing. We present a study on the influence of initial posture on landing impact and net energy consumption. Through simulations, we found the best initial postures to efficiently cross a wide ditch of width 1.05 m, with less impact and without singularities. Finally, we demonstrate the advantage of the proposed approach to cross a wide ditch when the surface friction is not same on both sides of the ditch.

SLIP-Based Control of Bipedal Walking Based on Two-Level Control Strategy

“…In the single-support phase, in order to ensure smooth movement of leg 2 from takeoff to touchdown we define a parabolic curve for point E (toe of swing leg) as (22).…”

“…Due to possible saturation of the ATRIAS motors, we put a ±13 N m saturation block to the output of control law. To ensure the swing toe clears the ground, we assume its height in mid-stance in (22) as y E mid = 0.06 m. For swing leg control (23), PID coefficients resulted by Newtonian optimization (24) using fmincon have the values of K P = 1033, K I = −66.2, K D = 69.5. The controller coefficients values in (25) to minimize torso angle error at the end of single-support phase are K F p = 0.6, K F d = 0.3 using a search algorithm.…”

“…In some recent works, Gupta et al 21 optimized walking gait variables and utilized artificial neural networks for generating walking gaits on uneven surfaces. Janardhan et al 22 investigated biped trajectory generation for crossing large ditches and proposed point of feasibility to bring the robot to rest at the end of ditch crossing. Kajita et al 23 controlled HRP-2Kai walking by discretizing the continuous system by a constant unit length along the walking direction.…”

Dynamic Optimization of Mechanism Parameters of Bipedal Robot Considering Full-Range Walking Energy Efficiency

“…With a body structure resembling that of humans, bipedal robots are well-suited to adapt to human living and working environments [5]. They can engage in tasks alongside humans or collaborate with them [6], for example, in operating production stations, assisting in driving, conducting rescue operations in hazardous areas [7,8], and navigating human environments, including stairs and obstacles. Humanoid bipedal robots are expected to emerge as a prominent robot form in the future.…”