Design Of Integrated GPS/DR Navigation System Based On DSP

After analyzing the key features and situation of GPS (Global Positioning System) and DR (Dead Reckoning) navigation technique, this paper discusses a design of integrated GPS/DR navigation system, in which Federated Kalman filter based on theory of optimal filter fuses data from local sensors, GPS and DR. This thesis also presents a system design including both hardware and software with a low power consumption DSP (Digital Signal Processor) to perform the filter algorithm and distribution of information. Furthermore, we achieve the federated Kalman filter algorithm for vehicle navigation application.The work of this paper is generalized as following:Firstly, after analyzing a long term GPS navigation information including Longitude and Latitude, we conclude that GPS positioning information is equivalent to a sum of actual position and a random noise, which means that GPS result is closer to its real position after a Kalman filter.Then, a federated filter is designed to combine the dispersed filters' outputs. We specify a state vector and local filter models for vehicle position and raise a method to decide the coefficient for information distributioaIt's also important to modify the models and algorithm to make it more practical and efficient when executed on the DSP platform. After the optimization of design, data collection and fusing are preformed by the DSP data fusing system with a TMS320VC5410, a fix point, low power consumption and high speed DSP.Author analyzes the method to initialize the states of DR and its validity in this paper. Error model of gyroscope and odometer and measures to reduce positioning error are also discusses.At the end of the paper, we discuss the error model of GPS/DR navigation system based on federated filter theory, and present some new ideals about terrestrial mobile position location.