Multi-robot competitive tracking based on k-WTA neural network with one single neuron

A k-winners-take-all (k-WTA) neural network is designed and applied to a task assignment problem in a multi-robot competitive target tracking scenario in this paper. The proposed neural network features a single neuron and a non-hard-limiting activation function, which greatly simplifies the model structure and reduces the computation cost. This neural network has finite-time convergence property and can be applied to real-time situations. The stability and convergence property of the neural network is theoretically analyzed. Simulations of handling a situation that a target moves at a higher speed than tracking robots are conducted to demonstrate the effectiveness of the designed scheme. (c) 2021 Elsevier B.V. All rights reserved.

Automated summary

In this paper, a k-winners-take-all (k-WTA) neural network is designed and applied to a task assignment problem in a multi-robot competitive target tracking scenario. The proposed neural network features a single neuron and a non-hard-limiting activation function, which greatly simplifies the model structure and reduces the computation cost. The stability and convergence property of the neural network is theoretically analyzed, and simulations demonstrate its effectiveness in tracking targets moving at higher speeds than the tracking robots.