The Research Of GPS/DR Fusion Simulation Based On UKF

Container transportation is always an important representative of the global freight industry, and constantly to the large-scale, intelligent and green development direction.Automation is the core of the transformation operation equipment upgrade, China is the first global container throughput, which didn’t has a automation terminal or AGV. Now many countries have set up automatic loading dock with AGV, such as:Singapore port, Rotterdam port, DongJing pot etc.China initially formed four goods professional, efficient transport system around the coal, oil, iron ore and container.The research about the port AGV will has important strategic significance and broad market prospect for china.AGV control system has three major technology:Navigation、Layout designing、Guidance.This article mainly discuss AGV navigation technology.Since1953the began of first AGV in United States, many research institutions, companies research on AGV guidance technology, and made some progress.Now main utdoor AGV navigation means has magnetic navigation, laser navigation, visual navigation.This kand of AGV navigation technology cost high, and the path not easy to change, so there still have way from the real mature practical port AGV system.In order to study the suitable way for AGV navigation, this paper focuses on the AGV control system.Federal kalman filter respects code disc and GPS as separate filter, and two filter don’t affecte each other. At last combination the optimal estimate of the two filter to get the ultimate value, and feedback it to the code disc and GPS in next filtering.In the research, we finds that the physical error of AGV right wheels code disk causes the AGV car path always left. Because of the experimental conditions limited, we have no way to get the exact value of AGV code disc. This paper puts forward system compensation coefficient G, and applies it to the federal kalman filter. Simulation results demonstrate that the compensation coefficient G solved the problem of code disk error uncertain, make the effect of filtering easier for manual control, and the precision of the system is improved by0.4m.This paper has designed a unit kalman filter for GPS/DR integrated navigation system base on AGV, built a MATLAB simulation model. In50m, GPS navigation error range is6m, signal fluctuates greatly. DR navigation maximum error is1.8m, the result don’t convergence. GPS/DR integrated navigation system error is0.2m, the signal wave smoothly and result convergence. GPS/DR integrated navigation system overcomes the GPS signal shelter problem, and makes full use of the advantages just as short-term、high precision、independent positioning of DR. The simulation results shows that this method ensure the accuracy of the system premise and effectively improved the system fault tolerance and working reliability.