Research On Key Technologies Of Image Data Processing And Wireless Communication For UHV Substation Robot Inspection

Substation power equipment has slowly transformed from the traditional human monitoring and camera remote monitoring mode into an intelligent management mode.The intelligent identification of the equipment is realized through robotics and the realtime interaction with the back-end server through wireless communication technology is used to realize all-round monitoring of the substation.Monitoring,however,the accuracy of target recognition and the reliability and real-time of information interaction are still important issues for the realization of intelligent substation management.With the rapid development of deep learning and wireless communication technology,the accuracy of target recognition has been greatly improved.In communication with the background server,the wireless communication network is very suitable for the complex environment of substations due to its flexible networking and low cost.in.In order to ensure the safety of equipment,substations have higher requirements for the reliability of wireless communication.By increasing the power of the transmitting node,the reliability can be improved,and it can also cause mutual interference between other node equipment and affect the overall communication performance.Therefore,the transmission power is limited.Control is very important.In response to the above problems,this article uses the Jetson Nano development board to build a YOLOv5 framework to identify and locate the target,and realizes the control of the wireless communication transmission power based on the linear quadratic Gaussian control(LQG)algorithm.The main work of this thesis is as follows:(1)According to the requirements of substation target recognition accuracy,select the embedded Jetson Nano development board,and build a software framework based on the NVIDA platform,use the improved YOLOv5 algorithm to train the pictures on the Personal Computer,and transplant the trained model files to Jetson Nano Realize target detection and recognition in the development board.(2)By analyzing the wireless communication reliability performance index and path loss model,the signal-to-noise ratio is selected as the evaluation index,and based on the logarithmic path loss model,the system space equation with the signal-to-noise ratio as the state variable is constructed.(3)The Kalman filter is used to separate the signal-to-noise ratio interference and measurement noise in the system,avoiding the influence of measurement noise on the control performance,and using linear quadratic regulator(LQR)control and multiparameter planning algorithms to Obtain the optimal transmission power of the wireless communication link.(4)The statistical confidence interval is introduced to compensate the expected signalto-noise ratio to ensure that the lower limit of the signal-to-noise ratio still meets the reliability requirements,and simulation comparisons and experiments are performed to verify the feasible performance of the proposed LQG control method.