Study Of The Identification For Temperature Model And The Method For Compensating Of Accelerometer Based On DSP

The accelerometer is the indispensable apparatus in the Inertia Navigation System (INS), the INS's precision affected by the accelerometer's. But the accelerometer's precision affected by the temperature. The identification of the accelerometer's static temperature model and the research on the error compensation are the important subjects in the present field of INS. Based on studying the different methods what researched on the world, we pointed out the shortage of some methods and gave a new method. We pay more attention to the identification for temperature model of the quartz flexible accelerometer and the research on the error–compensation. The purpose of the system realizes the real time soft compensation. The content of this paper are following:1.The works what have done on reducing even eliminating the effects of temperature were discussed.2.The Accelerometer's Static Rolling Experiment in 1g-Gravitational field is proposed in the paper, according to the System-identification Theory. Above all, based on the data from the Accelerometer's Static Rolling Experiment in 1g-Gravitational field, by using the method of Least-square Procedure, we got the value of parameter of the model of accelerometer, and then by using the cubic spline interpolation we got the temperature function of parameter. Finally we can build the static temperature model of this accelerometer during +20℃～+50℃. Then, through the model, we can finish easily the accelerometer's temperature error compensation on the software design. In this paper some3.The design and realize of the hardware system and software system were introduced. The programme is made up of different blocks, and we provide the simple program of C language. Finally, discusses how to improve the accuracy of data collection, and give the corresponding improvement measures from hardware and software.The experimental result proves that the accelerometer's output precision has been improved greatly by the error-compensation. And this method is feasible.