Design Of GPS Positioning Information Processing System Based On FPGA

With the development of GPS (Global Positioning System) technology, the number of GPS receiver users is increasing rapidly as GPS has the global, all-weather, low-cost features. However, there are various error sources in positioning, which affected stand-alone positioning accuracy. Two methods of reducing the errors are often considered to improve accuracy: one is increasing the cost of hardware such as using high-precision phase antenna, differential technology; The other is using the different filter algorithms for the different error sources from the software aspect. Comparing with the former, the latter can save the cost in the requirement of the accuracy, and convenient for the system fusion, it is a more promising approach in the application of the GPS. But the system takes much time to locate using filter algorithm, the traditional CPU doesn't meet with the real time requirement, and the FPGA becomes more and more popular as its parallel computing.This paper uses the simultaneous and top-down and bottom-up hybrid design method to implement the system design according the "timing-circuit" mind based on FPGA. The implement of the GPS system which is a rapid calculation, high precision and real-time location system based on FPGA is a new attempt and lays the foundation for the development of the embedded system. The system achieves receiving information from the GPS-OEM, transforming the coordinate and reducing the location errors using kalman filter. Specific tasks are as follow:The system implements to receive the GPS data correctly and display the data on the LCD, and transforms the latitude and longitude information to the data in the WGS-84 coordinate. According the standards and formats of the GPS output, the module of serial achieves receiving the latitude and longitude information and sending the data to real-time display on LCD. It changes the data from ASCII codes to decimal integer data while extracting the information, and using shift and addition to replace the multiplication attain the aim that reducing resource utilization. In the coordinate conversion process, firstly, the paper uses the lookup table to find transformation parameters required, secondly, changes the parameters into double-precision floating point data, and then converses the coordinate. According to the rule of Gauss formula, the formula is simplified to which involves only addition and multiplication, to reduce computations and save resources.This paper implement kalman filter. Firstly, the various errors are considered as the total of first-order Markov process by analyzing the errors impacted of position accuracy, and system equations, observation equations and filtering equations are established. Then, the kalman filter is designed based on decentralized kalman filtering and simulated in Matlab. The results show that the design of the kalman filter has good convergence, high positioning accuracy and small estimation error. On the basis of simulation, kalman filter is implemented based on FPGA. The paper uses kalman filter to improve the accuracy of data, and adopts time-sharing multiplexing technology in IP core and modules which has designed and tree fabric to save resources when the system has met the real-time requirement.The design is based on XC5VLX110-FF676 of Virtex-5 of Xilinx, and implemented using Verilog HDL. And then, it is placed and routed using the Xilinx tool ISE 10.1. It consumed 44 438 logic resources in total and the clock frequency is able to reach 100MHZ or more. Under premise of the requirement of real-time signal processing and accuracy, the design resources are optimal. Modelsim simulation verifies the correctness of the design.