A novel shearer cutting pattern recognition model with chaotic gravitational search optimization. (October 2019)

• Record Type: Journal Article
• Authors: Jiang, Yaping; Xu, Zhipeng; Zhang, Zeyin; Liu, Xinggao
• Is Part Of: Measurement. Volume 144(2019)
• Page Start: 225
• Page End: 233
• Publication Date: 2019-10
• Subjects: Cutting pattern recognition -- Relevance Vector Machine (RVM) -- Gravitational Search Algorithm (GSA) -- Chaotic mapping; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2019.05.019
• Languages: English
• ISSNs: 0263-2241

Abstract

Graphical abstract: Highlights: Cutting pattern recognition based on the voltage, current and rotating speed of the shearer motor. No additional feature extraction or data processing operations on the raw data. Gravitational Search Algorithm is introduced to optimize the hyperparameters in Relevance Vector Machine. Chaotic mapping increases the diversity of the optimization algorithm. Abstract: The accurate recognition of the shearer cutting pattern is the focus in fully mechanized coal mining. Hence, a new cutting pattern recognition model based on the combination of Relevance Vector Machine (RVM) and Chaotic Gravitational Search Algorithm (CGSA) is proposed. Initially, the motor operation data, including voltage, current and motor speed, are collected as the detection signal and the RVM classifier based on Bayesian framework is chosen for pattern recognition. In order to optimize the parameters in RVM, which has a great influence on the performance of RVM, the optimization algorithm Gravitational Search Algorithm (GSA) is introduced. Finally, the basic GSA is modified into CGSA with the chaotic mapping for increasing the search diversity of the algorithm. The experimental study demonstrates the advantageous performance of the proposed model even without any feature extraction operations.