Auxiliary generative mutual adversarial networks for class-imbalanced fault diagnosis under small samples

The effect of intelligent fault diagnosis of mechanical equipment based on data-driven is often premised on big data and class-balance.However,due to the limitation of working environment,operating conditions and equipment status,the fault data collected by mechanical equipment are often small and imbalanced with normal samples.Therefore,in order to solve the abovementioned dilemma faced by the fault diagnosis of practical mechanical equipment,an auxiliary generative mutual adversarial network(AGMAN)is proposed.Firstly,the generator combined with the auto-encoder(AE)constructs the decoder reconstruction feature loss to assist it to complete the accurate mapping between noise distribution and real data distribution,generate highquality fake samples,supplement the imbalanced dataset to improve the accuracy of small sample class-imbalanced fault diagnosis.Secondly,the discriminator introduces a structure with unshared dual discriminators.Realize the mutual adversarial between the dual discriminator by setting the scoring criteria that the dual discriminator are completely opposite to the real and fake samples,thus improving the quality and diversity of generated samples to avoid mode collapse.Finally,the auxiliary generator and the dual discriminator are updated alternately.The auxiliary generator can generate fake samples that deceive both discriminators at the same time.Meanwhile,the dual discriminator cannot give correct scores to the real and fake samples according to their respective scoring criteria,so as to achieve Nash equilibrium.Using three different test-bed datasets for verification,the experimental results show that the proposed method can explicitly generate highquality fake samples,which greatly improves the accuracy of class-unbalanced fault diagnosis under small sample,especially when it is extremely imbalanced,after using this method to supplement fake samples,the fault diagnosis accuracy of DCNN and SAE are relatively big improvements.So,the proposed method provides an effective solution for small sample class-unbalanced fault diagnosis.

Intelligent navigation algorithm of plant phenotype detection robot based on dynamic credibility evaluation

Due to the non-standardization and complexity of the farmland environment,Global Navigation Satellite System(GNSS)navigation signal may be affected by the tree shade,and visual navigation is susceptible to winged insect and mud,which makes the navigation information of the plant phenotype detection robot unreliable.To solve this problem,this study proposed a multi-sensor information fusion intelligent navigation algorithm based on dynamic credibility evaluation.First,three navigation methods were studied:GNSS and Inertial Navigation System(INS)deep coupling navigation,depth image-based visual navigation,and maize image sequence navigation.Then a credibility evaluation model based on a deep belief network was established,which used dynamically updated credibility to intelligently fuse navigation results to reduce data fusion errors and enhance navigation reliability.At last,the algorithm was loaded on the plant phenotype detection robot for experimental testing in the field.The result shows that the navigation error is 2.7 cm and the navigation effect of the multi-sensor information fusion method is better than that of the single navigation method in the case of multiple disturbances.The multi-sensor information fusion method proposed in this study uses the credibility model of the deep belief network to perform navigation information fusion,which can effectively solve the problem of reliable navigation of the plant phenotype detection robot in the complex environment of farmland,and has important application prospects.