Research On Optimal Design And Height Fitting Method Of GPS Control Network In Long And Narrow Area

With the rapid development of GPS technology,the use of GPS technology to measure high-speed railways can play a significant role in the construction of the railway and the subsequent operation and maintenance.At present,many railways under construction have a harsh geographic environment,long and narrow,with large height differences,which have a lot of influence on the measurement.This also gives us the opportunity to conduct research on certain aspects.This article will discuss two aspects in depth.One is that before the measurement,the measurement control network is usually designed,but it is generally designed based on the experience of the technicians.This form is subjective,not popularized and rigorous,and may sometimes lead to the network The structural accuracy is insufficient or too cumbersome,resulting in a waste of manpower,material resources,and financial resources.The second is that the normal height cannot be obtained directly from GPS measurement,and it must be obtained through leveling measurement,which will greatly affect the results in harsh operating conditions.This paper takes the measurement of the Mengla section of the Yumo Railway as the engineering background,and optimizes the design and research of the height fitting method for this narrow and long belt-shaped high-difference control network,which has high engineering significance.The main work is as follows:(1)Analyze several indicators for optimization design.Consider the use of reliability indicators for optimization for special measurement areas such as the Yumo Railway.The redundant observation component of the reliability indicators is mainly considered to ensure accuracy and reliability.Under the premise,multiple simulation experiments are performed to obtain the optimal network structure,eliminate more redundant observations,reduce the workload of measurement,and save the cost of the project.(2)The optimized design of the control network is measured.After the measurement data is obtained,the data is first preprocessed,and then the data is calculated by the Huatest static data processing software.After the closed loop of the unqualified baseline is processed,the control network performs baseline calculation and network adjustment,and the final solution results meet the requirements of the specification.Qualified settlement results indirectly indicate the success of the optimized design and meet the design requirements.(3)Several methods of height fitting are analyzed and researched.For this special area,considering the advantages of the wide application range and high accuracy of the neural network fitting method,it is considered to use the LM-BP neural network method to try and study it,and Compared with the traditional BP fitting method,it analyzes and compares the number of hidden layers.(4)This paper proposes a cuckoo search algorithm to improve the fitting algorithm of the LS-SVM model,which perfectly solves the problem of difficult selection of regularization parameters and kernel width in the LS-SVM model.And using this method to analyze and compare the different training samples,confirming the unique advantages of CS-LS-SVM in small sample data.(5)Using the measurement data of Yumo Railway,the LM-BP,CS-LS-SVM fitting method and the traditional polynomial fitting method are compared and analyzed,and it is found that these two fitting methods are very suitable for measuring with this kind of narrow and long height difference.Area,can get very high fitting accuracy,when there are few known points,CS-LS-SVM can give play to its unique advantages,which has very important engineering significance.