Distributed Collaboration in Multimanipulator Systems With Switching and Weight-Unbalanced Topologies

In multimanipulator systems (M
$^{2}$Ss), one of the main challenges for the collaboration of manipulators is posed by the switching and weight-unbalanced topologies (SWUTs) due to the different communication power and unpredictable connection failures, which has not been investigated and can cause instability and performance degradation. In this article, following the distributed principle, the collaboration in M$^{2}$Ss with SWUTs is constructed for problems in the forms of optimization with a universal convex performance index and an inequality constraint for joint velocities, in which an algorithm for weight-unbalanced topologies to estimate their required left eigenvector is introduced, and a distributed formula with topology switching compensation is proposed as well. In the next step, a neural dynamics controller is employed to solve the aforementioned optimization problems, with its stability and convergence proved strictly. Afterwards, the efficacy of the proposed controller is validated through both simulations and physical experiments, and comparisons are provided to demonstrate its superiority in handling the abovementioned challenges in the distributed collaboration in M$^{2}$
Ss with SWUTs.