Cooperative Acquisition And Deep Mining Of Location Information In Intelligent Transportation Systems

With the continuous development of China’s economy,the number of vehicles is increasing,and the urban environment is becoming more and more complex.Intelligent Transportation Systems(ITS)is developing rapidly nowadays,and location information is the basis of ITS.In complex urban environment,satellite navigation faces many problems such as signal shielding and multipath effect,and the positioning accuracy is seriously degraded.Cooperative localization is one of the effective methods to improve the positioning accuracy in urban environments,but it also faces the problems caused by Non-Line-of-Sight(NLOS)path observation.At present,urban traffic is also facing traffic congestion,environmental pollution and other issues,these problems are what ITS to solve.The accuracy of location information is increasing,and it should assist ITS to solve these urban traffic problems in a more intelligent way.This paper designs a novel ITS processing framework,which includes localization and deep mining.Firstly,this paper proposes a Geographic Information Enhanced Cooperative Localization algorithm(GIE-CL).This algorithm innovatively designes a geographic information based Region Sampling Method(RSM)to handle the NLOS link observation in urban environment,and then the iterative Generate Approximate Message Passing(GAMP)algorithm is actived to improve the cooperative localization accuracy.Experiments show that GIE-CL is superior to other algorithms in urban environments with NLOS paths.Moreover,GIE-CL handles the high-speed changes of the vehicular network topology by the invariance and controllability of the geographic information.Based on cooperative positioning and the popularity of other smart devices,the ITS can periodically acquire the geographical position information of all moving objects in the system,and then splicing the position information of each moving object to its motion trajectory.In the deep mining part,this paper proposes a vehicle recognition algorithm based on Recurrent Neural Network(RNN),which can preprocess the motion trajectory,propose feature vectors,and use artificial intelligence to accurately identify the transpotation type.By identifying the transpotation type,we can form a real-time urban traffic heat map,traffic police and traffic lights can also make real-time adjustments to guide the movement of the traffic,thus solving the traffic congestion problem.From a more macro perspective,the proportion of travel tools and regional distribution can also reflect the development of urban economy and assist the future planning and construction of the city.In future research work,we will try to apply NLOS path observations to GIE-CL instead of simply eliminating them.How to infer the movement pattern and living habits of moving objects from the trajectory and predict the position is also one of the topics worthy of future research.