This article is cited in 1 paper

System analysis, management and information processing

Intelligent system for testing robotic complexes using sigma-pi neural networks

R. A. Zhilov

Institute of Applied Mathematics and Automation – branch of Kabardino-Balkarian Scientific Center of the Russian Academy of Sciences, 360000, Russia, Nalchik, 89 A Shortanov street

Abstract: The paper considers the problem of developing an intelligent testing system for robotic systems based on sigma-pi neural networks. On production lines where industrial robots are used, the task of testing them for performance is urgent. There are two main ways to solve this problem: routine checks of robotic systems or constant observation of the operator at the robotic line. This paper presents an intelligent system built on the basis of sigma-pi neural networks, which will be able to solve a similar problem using readings from sensors located at different nodes of the robot. A neural network trained according to the algorithm considered in the work can continuously monitor the state of robots on the production line and make a decision to stop the line in case of suspicion of a breakdown. As an example of the operation of a sigma-pi neural network in this work, an example is provided based on 5 input data, that is, data from 5 sensors, normalized according to the principle "there is a signal" or "there is no signal".

Keywords: :sigma-pi neural networks, control problem, intelligent testing, robotic systems, neurocontrol.

UDC: 004.8

MSC: 68T07

Received: 28.10.2021
Accepted: 12.11.2021

DOI: 10.35330/1991-6639-2021-6-104-43-49

Bibliographic databases:

•