Virtual mechanism approach for dual-arm manipulation

Abstract

SUMMARYWe propose a novel control approach for cooperative dual-arm manipulation tasks. Our scheme has three typical features: (1) the task performed by two robots is represented as a motion of a virtual mechanism and the task execution is accomplished by controlling the virtual mechanism; (2) the two arms and the task form a joined kinematic chain; (3) the scheme allows a cooperative dual-arm system to perform the task also when robot base is moving. The calculation of the Jacobian matrix of a chained two-arm mechanisms is based on a methodology which is using the Jacobian matrices of particular robot mechanisms and their end-effector positions and orientations. The proposed algorithm for dual-arm manipulation is verified by simulations of two cooperating planar robots and by experiments on a dual-arm robot consisting of two KUKA LWR arms.

Cooperative control of a serial-to-parallel structure using a virtual kinematic chain in a mobile dual-arm manipulation application

“…In the virtual model control [9], the legs of a robot are programmed to mimic different virtual mechanical structures in order to control the dynamics along the gravity force direction. In the virtual mechanism approach [10], a VKC was used to chain serial mechanisms together. Our approach is different from [9] in that the VKC is part of a kinematic model, and different from [10] in that we use the VKC to specify the common motion shared by the two parallel structures.…”

Coordinated multi-arm motion planning: Reaching for moving objects in the face of uncertainty

“…Zhu [34] proposed a motion synchronization controller to coordinate the end-effectors of two robots when they are rigidly or flexibly holding an object without payload. Likar et al [14] proposed a velocity level motion synchronization algorithm for controlling a cooperative dual-arm system. They introduced an augmented kinematic chain which is a representation of two arms and the object.…”

“…5). Due to (14), successful tracking and coordination with the object and with each other is achieved.…”

“…The fingers close and the box is grabbed from the human. The initial values of γ p and γ o in (14) are set to 0. An example of the desired robot trajectory and the box trajectory are shown in Fig.6.…”

“…The new feasible intercept point is chosen in the vicinity of the previous one to minimize the convergence time. As the motion of the object is fast and the predicted reaching points are not accurate, the initial values of γ p and γ o in (14) are set to 0.5 and 0.0, respectively. This decreases the convergence duration of the robots to the real object.…”