Research Area:  Machine Learning

Abstract:

This paper addresses an effective deep learning-based technique for detection of robotic manipulator’s failure execution. The problem is based on the control strategy of robotic manipulators subjected to uncertain dynamics. The main contribution is to detect the failures at each different position and instance of robotic manipulators with a certain control strategy. An efficient deep belief neural network-based model is developed with an effective distribution of features at each layer of the network to demonstrate the accurate detection of failures at each instance. With the help of various suitable learning parameters at different stages of network and contrastive divergence operation, the proposed method is able to be an emergent solution for the failure detection. The performance of the proposed DBN is compared with other seven standard machine learning-based classifiers and the results are evident toward the significant impact on the high detection rate as well as the robustness of the proposed method.

Keywords:  

Author(s) Name:  Pandit Byomakesha Dash, Bighnaraj Naik, Janmenjoy Nayak, S. Vimal

Publisher name:  Springer

DOI:  10.1007/s00500-021-05572-0

Volume Information:  Volume 27, pages 363–375, (2023)

Paper Link:   https://link.springer.com/article/10.1007/s00500-021-05572-0