Research on Multi-robot Intelligent Fusion Technology Based on Multi-mode Deep Learning

Yang Sun ； Shoulin Yin ； Lin Teng

《International Journal of Electronics and Information Engineering》 12卷3期 (2020/09) Pp. 119-127

https://doi.org/10.6636/IJEIE.202009_12(3).03

摘要

In this paper, we focus on the multi-robot intelligent fusion technology based on multi-mode deep learning method. And we briefly propose three new schemes in the multi-robot intelligent system. 1) A grab classifier based on deep learning is established. This paper constructs a grab classifier based on multi-mode deep learning algorithm, which realizes the precise classification of grab rectangles. 2) A multi-robot distributed self-organizing task allocation algorithm based on multi-mode fusion is proposed to solve the problem of resource competition conflict in distributed task allocation and the problem that the solution quality and solving cost are difficult to coordinate. 3) Aiming at the path planning problem of multi-robot system without communication under un- known dynamic environment, a multi-robot path planning and dynamic collision avoidance system based on two-layer fuzzy logic is designed. Finally, experiments show that the proposed methods have better results than other approaches.

關鍵字

Grab Classifier ； Multi-robot Distributed Self-organizing Task Allocation ； Multi-robot Intelligent Fusion ； Path Planning