基于改进的RBF神经网络倾角传感器温度补偿方法研究

Research on temperature compensation method of inclination sensor based on improved RBF neural network

针对MEMS倾角传感器零位温度漂移问题,提出了粒子群优化(PSO)算法和遗传算法(GA)相结合优化径向基函数(RBF)神经网络的补偿方法,克服了RBF神经网络收敛慢、泛用性低的缺陷。结果表明:该方法能够有效地消除温度对MEMS倾角传感器输出的影响。相较于RBF神经网络模型,最大相对误差(MRE)减小了21.03%,均方根误差(RMSE)减小了23.54%,温度漂移得到明显改善,提高了倾角传感器的稳定性与准确性。

Aiming at the problem of zero position temperature drift of MEMS tilt sensor,a compensation method of optimizing radial basis function(RBF)neural network by combining particle swarm optimization(PSO)algorithm and genetic algorithm(GA)is proposed,which overcomes the defects of slow convergence and low universality of RBF neural network.The results show that this method can effectively eliminate the influence of temperature on the output of MEMS tilt sensor.Compared with RBF neural network model,the maximum relative error(MRE)is reduced by 21.03%,the root mean square error(RMSE)is reduced by 23.54%,the temperature drift is significantly improved,and the stability and accuracy of the tilt sensor are improved.