Software Optimisation of Digital Readout Electronics for Micro-Machined Gyroscopes

摘要

This paper report about a new amplitude demodulation algorithm by the example of the micro-machined angular rate sensor DAVED (DECOUPLED ANGULAR VELOCITY DETECTOR) developed at HSG-IMIT. The basis for this demodulation algorithm is digital readout electronics for gyroscope. The simulation and measurement results will be presented for the first time. The resonant frequencies of gyroscopes are normally in the range from 1...50 kHz. Due to the higher requirement, the higher speed of DSP is needed. For high performance applications, for which the DAVED~R-RR is developed, robust and time consumption of algorithm should be carefully considered. The output signal of the sensor is an amplitude modulated (AM) signal. Standard Demodulation methods are based on Muitiplication Demodulation. Least Mean Square Demodulation (LMSD) and Recursive Least Square algorithm (RLSD) for fixed-point DSP system are developed to enhance system performance and reduce the time consumption of demodulation algorithm. Furthermore, Kalman Filter Demodulation (KFD) is also built up to compare the performance. Four demodulation methods are compared in simulation results. Simulations show that LMSD can slightly increase system performance. On the other hand, due to the speed of DSP, only Multiplication Demodulation and LMSD methods are implemented in real system. Experimental results are shown: In fix point DSP system, Least Mean Square Demodulation (LMSD) can decrease RMS noise level about 13～30% and reduce time consumption 10～20% (due to cancel some filters) relative to Multiplication Demodulation. Through simulation and experimental results, LMSD method is preferred for demodulation in fix point system. LMSD algorithm can be applied to other demodulation applications also.