Key Technologies Of Multi-level 3D Virtual Monitoring System With Multiple Information Fusion For Automobile Assembly Line

In recent years, the car market competition is becoming increasingly f ierce, the user personalized needs is increasing, and automobile production mode is changing from mass production to mass production, small batch, multispecies, fast rhythm of the flexibility production, which makes the automobile manufacturing inevitable irreversible to the digital, intelligent, visual development. In the paper, according to the automobile assembly line has the characteristics with multi-layer system structure, variety of equipment, complex and changeable condition environment, a large number of multi-source heterogeneous information and holographic, real-time, and transparent production control requirements, a multi-level 3D virtual monitoring system with multiple information fusion for automobile assembly line is developed, through the combination of network technology and computer graphics technology, to have a visual management of automobile assembly production depth and high efficiency in a friendly way of 3D human computer interaction, to improve the transparency and convenience of automobile production and assembly process management and then provide support for sustainable production and process optimization dec ision making. The specific research work of this paper is as follows:First, based on the requirement analysis and system target design analysis of visual monitoring of automobile assembly line, the architecture of multi-level 3D virtual monitoring system with multip le information fusion for automobile assembly line is built, and the system structure, operation mode, function structure and network topology are designed for 3D real time visual management and control system. Then according to the design of system architecture, aiming at the two key technology, the entire line of manufacturing information integrated processing technology and manufacturing scene 3D visualization synchronization technology were studied.In the aspect of manufacturing information integration and processing, according to the characteristics of automobile assembly line and the demand of multi dimension information, the collection strategy of multi-source heterogeneous information is put forward, and the scheme of information integration based on mult i- level monitoring is designed; according to the correlation of multi-level and multi dimension state information, state feature recognition method of basic state and compound state is proposed; In view of the complex state fusion recognition problem, a multi information fusion technology is put forward which is a combination of support vector machine and evidence theory, and its effectiveness is verified by a numerical example.In the realization of the 3D visual synchronization of the automobile assembly line manufacturing scene, the techniques of simultaneous mapping and optimization about static and dynamic scenes of real monitor are studied separately. Static mapping studies 3D static display of complex manufacturing scenes optimization through making the object 3D modeling, virtual scene construction, static scene illuminant optimization and static model optimization; considering the real time dynamic change characteristics and the control requirement of the manufacturing process, in order to solve the problem of synchronous display of dynamic scenes, the model behavior control technology which is parallel driven by the dynamic characteristics of equipment system and the logic of process behavior and scene display feed-forward control method are proposed in order to realize the dynamic synchronous monitoring system based on real time monitoring data.Finally, by taking the study on the transmission line and AGV auxiliary material transportation system of the automobile assembly with friction drive as an example, the application object 3D virtual monitoring system is developed, the system information collection, integration, processing method and the visualization of 3D virtual monitor ing implementation steps are described in detail, the application results of scene display and condition monitoring ver ify the feasibility and effectiveness of the proposed technique.