Research On Motion Distortion And Fault Diagnosis Method Of Serial Robot Based On Pose Information

In industrial production,tandem robots have been widely used,but during the use of industrial robots,due to the wear and tear of parts and components and the external force during work,the robot is extremely prone to failure.Once a failure occurs,the end of the robot The actuator will produce motion distortion,and the position and attitude during processing will be abnormally changed,which will affect the processing accuracy and cause serious losses in severe cases.Therefore,this paper mainly tests the acceleration and angular velocity information of the robot during motion for fault diagnosis research.This article mainly completed the following aspects:(1)Designed and developed a signal acquisition system based on STM32 and MPU9250.The system mainly includes the following modules:STM32 microprocessor module,inertial sensor module,data external storage module,serial communication module and GPRS network communication module.The system can realize the collection,transmission and storage of three-axis acceleration and three-axis angular velocity.(2)Developed a software acquisition system based on C and Visual C#.This software system can send and receive character commands to and from the upper and lower computers.It can also draw waveforms on the received data.At the same time,it has also developed signal analysis and processing functions Including spectrum analysis,time domain characteristic parameter statistics and filtering processing functions.(3)This paper proposes a method for extracting transient impact signals based on saliency models.The model is divided by Gammatone band pass filtering,then envelope extraction and non-linear compression processing.Finally,the obtained signal is multiscale The two-dimensional filtering process obtains the time-domain saliency map and global saliency map of the signal,and extracts the shock signal from it.Compared with the traditional model,the shock signal extracted by this method is more clear and can effectively suppress the background interference signal.(4)The shock signal extracted by the auditory saliency model is used for fault diagnosis,and the extracted shock signal is used for feature parameter calculation.Then,the KPCA kernel principal component analysis method is used to reduce the multidimensional feature parameters,and finally a three-dimensional cluster display under different fault samples is generated Result graph,to achieve the identification of different faults.