基于小波最优重构尺度的AUV推进器故障检测方法

针对采用传统小波方法检测外部干扰下自主式水下机器人(AUV)推进器故障时存在的故障检测灵敏度较低问题,提出一种基于小波最优重构尺度确定的AUV推进器故障检测方法,基于小波Shannon熵的小波最优重构尺度确定方法确定离散多层小波分解后细节系数的最优重构尺度,目的是滤除外部干扰等与故障无关信号,并选择故障信息含量最多的最佳重构尺度进行小波单支重构以识别AUV推进器故障特征.AUV实验样机水池实验结果表明,与传统小波方法相比较,所提方法故障检测灵敏度提高了27.78%.

Thruster Fault Detection for Autonomous Underwater Vehicle Based on the Optimal Wavelet Reconstruction Scale Determination

Aiming at the unsatisfactory fault detection sensitivity based on the conventional wavelet decomposition for autonomous underwater vehicle (AUV) subject to the external disturbance, a thruster fault detection method was proposed based on optimal wavelet reconstruction scale determination. In order to filter the irrelevant information, e.g. external disturbance, the optimal reconstruction scale was determined based on Shannon entropy theory for the detailed coefficients obtained from discrete wavelet decomposition. Therefore, fault feature extraction could be achieved by the determined optimal reconstruction scale. In the proposed method, the fault detection was to be achieved by selecting the modulus maximum of the detailed coefficients obtained from multi-resolution wavelet decomposition. Pool-experiments were performed on the Beaver 2 AUV. The experiment results showed the sensitivity of fault detection was improved 27.78% in comparison with the conventional wavelet decomposition.