Autonomous Bipedal Humanoid Grasping with Base Repositioning and Whole-Body Control

Abstract

Autonomous behaviors in humanoid robots are generally implemented by considering the robot as two separate parts, using the lower body for locomotion and balancing, and the upper body for manipulation actions. This paper provides a unified framework for autonomous grasping with bipedal robots using a compliant whole-body controller. The grasping action is based on parametric grasp planning for unknown objects using shape primitives, which allows a generation of multiple grasp poses on the object. A reachability analysis is used to select the final grasp, and also for triggering a base repositioning behavior that locates the robot on a better position for grasping the desired object more confidently, considering all grasps and the uncertainty in reaching the desired position. The whole-body controller accounts for perturbations at any level and ensures a successful execution of the intended task. The approach is implemented in the humanoid robot TORO, and different experiments demonstrate its robustness and flexibility.

Whole-Body Control for Multi-Contact Balancing of Humanoid Robots

“…Demonstration of the performance and versatility of the whole-body control framework in various applications including planning [Werner et al, 2016, Sundaram et al, 2018, teleoperation [Porges et al, 2019, Abi-Farraj et al, 2018, and industrial manufacturing [Kheddar et al, 2019].…”

“…Chapter 9 [Werner et al, 2016] [ Sundaram et al, 2018] [ [ Porges et al, 2019] [Abi-Farraj et al, 2018 [ Kheddar et al, 2019] Whole-Body Control Generalization Demonstration Round and square brackets are used to combine symbols into vectors and matrices, respectively. For instance, a vector is given by f = ( fx fy fz ) T and a matrix by J = [ Ad J ].…”

“…We have demonstrated several times that the framework for whole-body balancing on multiple contacts can be interfaced with various external components such as algorithms for path planning [Werner et al, 2016, Sundaram et al, 2018 or input devices for teleoperation [Abi-Farraj et al, 2019, Porges et al, 2019. Of course, the variants of the multi-contact balancer presented in Sections 6.1 to 6.5 require a different interface depending on the provided features and internal architecture.…”

“…To employ humanoid robots for disaster scenarios or in the field of industrial manufacturing, it is essential that the robots show a certain level of autonomy, which can be achieved by combining whole-body control with planning and perception. This section presents the works of Sundaram et al [2018] and Werner et al [2016], where the MCB controller from Section 6.1 was combined with planning and perception algorithms for autonomous grasping and multi-contact stair climbing, respectively. The MCB controller was extended in to increase the achievable stair height from 5 cm to 18 cm, which corresponds to an ordinary staircase in Germany [DIN e.V., 2016].…”

“…The MCB controller was extended in to increase the achievable stair height from 5 cm to 18 cm, which corresponds to an ordinary staircase in Germany [DIN e.V., 2016]. 9.1.1 Autonomous Grasping Sundaram et al [2018] presented an integrated pipeline for generating autonomous grasping behaviors with humanoid robots. The pipeline comprises several components for processing visual information, planning, and control, as shown in the system overview in Fig.…”