As a master of automation technology, industrial robot is known as "the Pearl on the crown of manufacturing industry". It integrates the latest achievements in the fields of machinery, electronics, computers, sensors, automatic control and artificial intelligence. It is increasingly replacing operators to complete monotonous, frequent and repetitive long-time work. In the traditional manufacturing field, industrial robot has become an indispensable core automation equipment after its birth, growth and maturity. The industrial robot industry is a new industry rising rapidly after the automobile industry. In the 21st century, with the rapid development of science and technology, the industrial robot industry has thrived in this "hotbed". The continuous progress and improvement of industrial robot technology has improved social productivity and promoted social development.

Remote monitoring of industrial robots using the Internet can realize remote condition monitoring. It is the premise of remote fault diagnosis and maintenance. It is conducive to improve the maintenance response speed and advance processing ability, reduce the failure rate, and has high practical application value to ensure the efficient operation of industrial robots. In industry, network communication technology has been more and more used in foreign industrial robot remote monitoring service system. ABB Sweden has developed the "remote service" platform. Its core device is a hardware service box with communication unit, which can realize the functions of remote real-time monitoring, data analysis and processing, fault prediction and so on, and can effectively reduce the use and after-sales cost of industrial robots.

Fanuc Japan (FANUC) also put forward its own remote monitoring service scheme. Different from ABB's scheme, FANUC industrial robot itself supports remote monitoring and remote diagnosis and maintenance operation without adding equipment similar to the service box. FANUC's remote service system can carry out remote communication and fault analysis, and timely notify the on-site operators and Fanuc's service engineers to enter Perform troubleshooting.

In academic circles, Jay Lee of the University of Cincinnati and others applied remote prediction analysis to the health management of industrial robots of Nissan company, and successfully predicted the early fault characteristics of industrial robot manipulator through torque and temperature signals [4]. Ken Taylor of the University of Western Australia remotely monitors an ABB industrial robot with six degrees of freedom when connected to the Internet. He can successfully grasp and carry objects to build complex models. Carter et al. Studied the design and implementation of the controller wireless access network of FANUC industrial robot r-j3ib, and successfully designed and developed a prototype network for remote monitoring of industrial robots in Nissan training institution in Mississippi.

Bjerkeng et al. Studied the industrial robot with active camera control based on the weighted pseudo inverse redundancy solution, proposed a systematic remote monitoring solution, and obtained successful experimental verification on KUKA kr-16 manipulator [7]. Sallinen et al. Proposed a framework for remote monitoring and maintenance of industrial robots based on Web user interface, which plays a key role in the implementation of safety monitoring and remote maintenance of industrial robots.

Leutert et al. Developed a remote information control system for remote monitoring of industrial robots. The system allows rapid analysis of the current work cycle and data processing of industrial robots, and makes the data more intuitive. Suzuki et al. Proposed using the tactile interface of tactile information to evaluate the remote monitoring of industrial robot, and completed the experiment of vibration tactile information of industrial robot. In China, Chen Yong of Southeast University proposed a remote monitoring method of industrial robot based on Ethernet network, processed multi-sensor signals through information fusion technology, and was applied to the industrial robot of automobile production line of NAC group. Li Changfeng of Harbin University of technology proposed a network control system of parallel robot based on TCP / IP protocol, which solved the remote control problem of parallel robot in inertial nuclear confinement fusion device, and realized the remote real-time control of parallel robot.

Lu Lu of Fudan University adopts the three-tier network model structure and puts forward a scheme of remote monitoring industrial robot system based on WWW. the scheme carries out real-time remote monitoring through Internet / Intranet and can be connected with other automation systems to realize the whole Bureau Management of the whole production process.

Liu Lei of Dalian University of technology studied the remote service system developed by ABB in Sweden and Fanuc in Japan. After analyzing their respective advantages and disadvantages, a remote monitoring system with both advantages is proposed and successfully applied to Xinsong sr6c industrial robot. Aiming at the establishment of socket and the development of multithreading technology, Zhang Aimin and others of Guangzhou CNC Equipment Co., Ltd. proposed the implementation process of industrial robot remote monitoring and diagnosis system based on TCP / IP protocol. Through this system, the position, running state and other relevant parameter information of on-site industrial robot can be viewed or modified anytime and anywhere.

On the basis of studying the traditional web-based industrial robot monitoring scheme, Zhou Li of Central South University proposed a robot remote monitoring system scheme based on CORBA and JSP / JavaBean technology, realizing the application of web technology and CORBA Technology in the field of industrial control [16]. Xu Qian of Jiangsu University has studied the integration of web server and industrial robot network control platform on embedded system. Operators can remotely monitor industrial robots through web browser.

Mei Xuhong of Hangzhou University of Electronic Science and technology proposed C / s for remote monitoring of industrial robots (client and server) mode. Yin Hongli of Central South University proposed an architecture based on Internet and sensor drive, so as to ensure no distortion of control information, reduce the action time difference between local simulated virtual industrial robot and remote real-time industrial robot, and combine remote monitoring and control friendly.

At present, industrial robot remote monitoring mainly realizes the remote monitoring function with the help of Internet-based remote monitoring technology, Internet communication technology, computer visualization technology and virtual reality technology. Industrial robot remote monitoring technology has a strong theoretical foundation and good technical support. However, in real life, the confidential data in remote monitoring and fault diagnosis may have potential security risks of leakage, tampering or loss. It is imperative to strengthen the data security research in industrial robot remote monitoring and fault diagnosis.

The collection, processing and fusion of industrial robot field signals is the basis of realizing industrial robot remote monitoring and fault diagnosis. After the on-site signal acquisition of industrial robot, the signals obtained from multi-sensor can not be directly used to evaluate, diagnose and predict the operation state of industrial robot. Because the amount of data of multi-sensor signals is large, there may be redundancy and contradiction between information, so it is difficult to analyze and apply directly. These signals must be processed, transformed and fused, Extract the signal eigenvalues that can accurately reflect the state of industrial robot. By establishing the mapping relationship between these signal eigenvalues and performance, life and fault, we can realize the reliable identification of the specific state of industrial robot. In terms of signal acquisition, David Alejandro et al. Proposed a new method to complete the signal acquisition of industrial robot vibration phenomenon by using accelerometer and gyroscope sensors.

Zhang Xingwu of China University of science and technology and others have designed an industrial robot intelligent signal acquisition and processing module with wide adaptability by using the signal acquisition and information processing technology with digital signal processing (DSP) as the core and combined with the design concept of modern intelligent sensor technology. Hu Xudong of Zhejiang University discussed the case of remote signal acquisition of industrial robot based on network. In terms of signal processing, Chen Xijun, Institute of automation, Chinese Academy of Sciences, and others proposed a method of using multiple DSPs to control and process all kinds of non visual sensors, and gave the principle and specific implementation process of sensor signal processing. Wang Jun of Huazhong University of technology and others introduced a high-precision multifunctional six dimensional wrist force sensor and its signal processing system for real-time detection of force / torque signals of mechanical wrist. Ma Yulong, State Key Laboratory of robotics, Shenyang Institute of automation, Chinese Academy of Sciences, discussed the application of colorless Kalman filter based on acceleration signal enhancement in robotics.

Yu along of Huaiyin Normal University applies multi wavelet transform to the signal denoising of industrial robot wrist force sensor. The floating threshold method is used to eliminate the noise. The wavelet transform method can process the signal more effectively, which has incomparable advantages over the traditional methods. Fu Yili of CIMS research center of Tongji University proposed a signal processing system based on digital signal processor (DSP) and complex programmable logic device (CPLD), which is suitable for real-time signal processing.

Based on the analysis of the signal characteristics of the chip, Cheng Li and others of Central South University for Nationalities proposed a signal decoding method, designed the corresponding algorithm, and implemented it in the specific industrial robot control system. In terms of signal fusion, scholars at home and abroad have conducted extensive research on multi-sensor signal fusion technology. At present, the representative methods in the field of multi-sensor signal fusion include weighted average, Bayesian estimation, Kalman filter, S-D reasoning, fuzzy logic and artificial neural network.

Patrik Axelsson et al. Developed a sensor signal fusion method for state estimation of flexible industrial robot. Guan Tianyun of Zhejiang University proposed a multi-sensor signal intelligent fusion system based on fuzzy integral theory, which was successfully applied to the fusion of non visual sensor signals of industrial robots.

It can be seen from the above literature that the methods of industrial robot signal acquisition, processing and fusion are gradually developing from the traditional single sensor to the digital signal processing system and more intelligent. The academic circles at home and abroad are in the stage of active thinking and exploration.