Event-based walking control — From neurobiology to biped robots

Abstract

In this paper we describe an event-based walking control system for biped robots. Classically, the control of biped robots has been separated into walking pattern generation and stabilizing control. While this is an effective strategy in wellknown environments, walking in rough, unmodelled terrain can easily destabilize the system. Findings from neurobiology suggest that gait generation in animals and humans is strongly linked to sensory signals indicating certain events in the gait cycle. We therefore propose an event-based controller that triggers phase transitions based on the sensed walking state instead of only relying on time-based reference trajectories. We have implemented the proposed approach on our humanoid robot Lola. We present simulations and experimental results demonstrating the effectiveness of the approach when walking over unknown obstacles.

Real-time pattern generation among obstacles for biped robots

“…For our work we want to emphasize the redundant kinematic structure of the legs with 7 DoF and the pelvis with 2 DoF. The recognition of the enviroment is done using the depth information obtained by an Asus Xtion PRO LIVE RGB-D camera 2 , which is mounted on a pan/tilt unit on the head of the robot. The vision processing software runs on an onboard computer with an Intel Core i7-4770S @ 3.1 GHz (4x) processor and 8GB RAM, and communicates with our main computer via Ethernet using UDP protocol.…”

“…Recent research even enables humanoids to compensate for disturbances [1], [2], [3]. However, current humanoid robots are still unable to match human performance in geometrically complex unknown environments.…”

“…This is why we rely exclusively on on-board sensing. (2) Enabling the robot to react to changes in the sensed environment or to user inputs in real-time without requiring slow movements to account for long processing times. Based on our work presented in [4] we developed a framework which closes the gap between obstacle recognition, collision-free walking in 2D and whole-body 3D collision avoidance.…”

An estimation model for footstep modifications of biped robots

“…The calculation of compensating motions is controlled by the finite state machine of the pattern generator. Note that the method is also triggered by the early contact event which is part of the sensor feedback of the finite state machine [24]. If the vertical contact force of the current swing leg exceeds a certain threshold, the next double support phase is initiated.…”

Ground Contact Detection with Shortstroke Buttons for Biped Robots

“…Simulation results with our robot Lola suggest that impact forces in case of an unexpected ground contact can be reduced significantly by integrating these sensors. 1 We currently use a PT1-filter with a cutoff frequency of 100 Hz. 2 We currently use a threshold of 50 N.…”