Multi-robot object tracking and docking systems based on networked control frames

Fan Wen,Zhenshen Qu,Changhong Wang

DOI: https://doi.org/10.1108/03684921011063592

IF: 2.352

2013-01-01

Kybernetes

Abstract:Purpose - The purpose of this paper is to describe how, in order to fulfill the specific missions under some special environments without people participating, a multi-robot object tracking and docking systems are designed based on networked control frames. Design/methodology/approach - In the process of target recognition and tracking, the tracking robot obtains the target robot's position and poses information by means of multi-sensors, and tracking the target robot uses a data fusion algorithm based on network-delay. In the phase of docking, the exterior parameters of the CCD camera installed on the tracking robot can be calculated in-phase by recognizing the coded target in a place on the target robot. Finally, the relative position and pose parameters between the tracking robot and the target robot can be derived using the coordinate rotation parameters. Findings - The experiment results indicated that the relative position measure error is less than 1.5 percent, and the relative pose measure error less than 1 degrees within 1.5-10 m. The research results show that the system can actualize object tracing and docking missions accurately and timely. Research limitations/implications - This paper is devoted to multi-robot object tracking and docking systems. Practical implications - The main applications are in the exploration in the seabed, consignment in the workshop, formation of spacecrafts, docking of spacecrafts, and so on. Originality/value - The system can actualize object tracing and docking missions accurately, and the system is of reliable, real-time, and robust capabilities. This will aid all developers and researchers to enhance their technicality.