针对选煤厂巡检自动化的需求,提出一种基于声音增强技术的智能巡检机器人设计方案.该方案采用麦克风阵列采集环境声音,通过卡尔曼滤波、梅尔频率倒谱系数(Mel Frequency Cepstrum Coefficient,MFCC)特征提取、长短期记忆网络(Long Short-Term Memory,LSTM)诊断设备异常.实验结果表明,该机器人在复杂环境下表现出优异的故障诊断与运动控制性能,故障检出率大于93％,定位误差小于5 cm,为建设智能选煤厂提供了可行的解决方案.
Design and Research of Intelligent Inspection Robot for Coal Preparation Plant Based on Sound Enhancement
A design scheme for an intelligent inspection robot based on sound enhancement technology is proposed to meet the demand for automated inspection in coal preparation plants. This scheme uses a microphone array to collect environmental sound,and diagnoses equipment abnormalities through Kalman filtering,Mel Frequency Cepstrum Coefficient (MFCC) feature extraction,and Long Short-Term Memory (LSTM) network. The experimental results show that the robot exhibits excellent fault diagnosis and motion control performance in complex environments,with a fault detection rate of over 93％ and a positioning error of less than 5 cm,providing a feasible solution for the construction of intelligent coal preparation plants.
万方数据
维普
