基于微形变分析的电容式MEMS加速度计温漂误差精密估计方法

doi:10.12305/j.issn.1001-506X.2024.07.25

摘要/Abstract

摘要：

针对传统的电容式微机电系统(micro-electro-mechanical system, MEMS)加速度计(capacitive MEMS accelerometers, CMA)温漂误差(temperature drift error, TDE)补偿方法存在非精准溯源TDE致使TDE估计精度低、反复尝试估计模型构型导致构建过程复杂繁琐的问题, 提出基于微形变分析的CMA TDE精密补偿方法。首先, 通过微形变分析内部结构精准溯源TDE, 基于径向基函数神经网络(radial basis function neural network, RBFNN)构建改进型TDE精密估计模型; 其次, 基于专家经验和模糊理论提出Expert-Fuzzy辅助决策下TDE估计模型辨识方法, 为改进模型提供有效的构型指导; 然后, 设计升温试验测试CMA, 构建传统模型和改进模型并通过对比其输出偏置稳定性评估TDE估计性能。实验结果表明, 改进模型构建过程大大简化, 补偿后CMA偏置稳定性提升约35%。本方法能够更加精准地估计TDE, 有效解耦硅基材料的温度依赖性并提升CMA的环境适应性。

关键词: 微机电系统, 温漂误差估计, 微形变分析, Expert-Fuzzy辅助决策, 径向基函数神经网络

Abstract:

Owing that the conventional temperature drift error (TDE) estimation method for capacitive micro-electro-mechanical system (MEMS) accelerometers (CMA) has demerits of inaccurate TDE traceability and tedious model establishing process, low accuracy and high complexity are introduced inevitably. A precise TDE estimation method of CMA based on micro-deformation analysis is proposed. Firstly, TDE is traced accurately with micro-deformation analysis, and a modified TDE estimation model is established with radial basis function neural network (RBFNN). Secondly, an accurate model identification method is proposed under Expert-Fuzzy assisted decision-making based on expert experience and fuzzy theory, which offers effective structural configuration guidance to the modified model. Then, temperature experiment is designed to test CMA, and the conventional and modified models are evaluated in performance by comparing their bias stability. Experimental results show that the establishing process for the modified model is greatly simplified, and its bias stability is improved by 35%. It ensures that TDE of CMA can be estimated much more precisely, which decouples temperature dependency of Si-based material and improves the environmental adaptability of CMA.

Key words: micro-electro-mechanical system (MEMS), temperature drift error (TDE) estimation, micro-deformation analysis, Expert-Fuzzy aided decision-making, radial basis function neural network (RBFNN)

中图分类号:

TH714
TN409

齐兵, 程建华, 赵砚驰, 汪籽粒. 基于微形变分析的电容式MEMS加速度计温漂误差精密估计方法[J]. 系统工程与电子技术, 2024, 46(7): 2437-2445.

Bing QI, Jianhua CHENG, Yanchi ZHAO, Zili WANG. Precise temperature drift error estimation method for capacitive MEMS accelerometers based on micro-deformation analysis[J]. Systems Engineering and Electronics, 2024, 46(7): 2437-2445.

图/表 18

图1

图2

图3

图4

图5

表1

表2

表3

图6

图7

表4

表5

表6

图8

图9

表7

图10

表8

参考文献 21

1	CAI Q , ZHAO F J . A novel parallel processing model for noise reduction and temperature compensation of MEMS gyroscope[J]. Micromachines, 2021, 12 (11): 1285. doi: 10.3390/mi12111285
2	CAI J , CHENG J H , LIU J X , et al. A polar rapid transfer alignment assisted by the im-proved polarized-light navigation[J]. IEEE Sensors Journal, 2020, 22 (3): 2508- 2517.
3	SHEN X W , YUAN D , CHANG R H , et al. A nonlinear observer for attitude estimation of vehicle-mounted satcom-on-the-move[J]. IEEE Sensors Journal, 2019, 19 (18): 8057- 8066. doi: 10.1109/JSEN.2019.2918381
4	杜俊伟, 陈起金, 牛小骥, 等. MEMS惯性高度尺测量精度分析[J]. 系统工程与电子技术, 2019, 41 (11): 2605- 2610.
	DU J W , CHEN Q J , NIU X J , et al. Analysis on measurement acuracy of MEMS inertial height gauge[J]. Systems Engineering and Electronics, 2019, 41 (11): 2605- 2610.
5	CHANG H W , GEORGY J . Improved cycling navigation using inertial sensors measurements from portable devices with arbitrary orientation[J]. IEEE Trans.on Instrumentation and Measurement, 2015, 64 (7): 2012- 2019. doi: 10.1109/TIM.2014.2381356
6	ISOBE A , KAMADA Y , TAKUBO C , et al. Design of perforated membrane for low-noise capacitive MEMS accelerometers[J]. IEEE Sensors Journal, 2020, 20 (3): 1184- 1190. doi: 10.1109/JSEN.2019.2948172
7	蒙涛, 王昊, 李辉, 等. MEMS陀螺误差建模与滤波方法[J]. 系统工程与电子技术, 2009, 31 (8): 1944- 1948. doi: 10.3321/j.issn:1001-506X.2009.08.037
	MENG T , WANG H , LI H , et al. Error modeling and filtering method for MEMS gyroscope[J]. Systems Engineering and Electronics, 2009, 31 (8): 1944- 1948. doi: 10.3321/j.issn:1001-506X.2009.08.037
8	WANG X , BLEIKER S J . Wafer-level vacuum sealing by transfer bonding of silicon caps for small footprint and ultra-thin MEMS packages[J]. Journal Microelectromechanical Systems, 2019, 28 (3): 460- 471. doi: 10.1109/JMEMS.2019.2910985
9	ZHANG H , WER X Y , GAO Y , et al. Analytical study and thermal compensation for capacitive MEMS accelerometer with anti-spring structure[J]. Journal of Microelectromechanical Systems, 2020, 29 (5): 1389- 1400. doi: 10.1109/JMEMS.2020.3011949
10	YIN Y G, FANG Z X. A temperature-insensitive micromachined resonant accelerometer with thermal stress isolation[C]//Proc. of the IEEE Sensors, 2018: 1656-1659.
11	MO Y M , YANG J , PENG B , et al. Design and verification of a structure for isolating stress in sandwich MEMS accelerometer[J]. Microsystem Technology, 2021, 27 (5): 1943- 1950. doi: 10.1007/s00542-020-04980-w
12	BASARAB M , GIANI A . Thermal accelerometer simulation by the R-functions method[J]. Applied Sciences-Basel, 2020, 10 (23): 8373. doi: 10.3390/app10238373
13	马龙, 代超璠, 裴昕. MEMS陀螺仪冗余配置故障检测方法[J]. 系统工程与电子技术, 2017, 39 (10): 2298- 2304.
	MA L , DAI C F , PEI X . Fault detection method for redundant configuration of MEMS gyroscope[J]. Systems Engineering and Electronics, 2017, 39 (10): 2298- 2304.
14	KIM B , HOPCROFT M A . Temperature dependence of quality factor in MEMS resonators[J]. Journal of Microelectromecha-nical Systems, 2008, 17 (3): 755- 766. doi: 10.1109/JMEMS.2008.924253
15	QI B , SHI S S , ZHAO L , et al. A novel temperature drift error precise estimation model for mems accelerometers using microstructure thermal analysis[J]. Micromachines, 2022, 13 (6): 835. doi: 10.3390/mi13060835
16	DAI G , LI M , HE X P , et al. Thermal drift analysis using a multiphysics model of bulk silicon MEMS capacitive accelerometer[J]. Sensors and Actuators Physical, 2011, 172 (2): 369- 378. doi: 10.1016/j.sna.2011.09.016
17	CHENG J C , FANG J C , WU W R , et al. Temperature drift modeling and compensation of RLG based on PSO tuning SVM[J]. Measurement, 2014, 55, 246- 254. doi: 10.1016/j.measurement.2014.05.013
18	PAN Y J , LI L L , REN C H . Study on the compensation for a quartz accelerometer based on a wavelet neural network[J]. Measurement Science & Technology, 2010, 21 (10): 105202.
19	XU D C, YANG Z M, ZHAO H M, et al. A temperature compensation method for MEMS acelerometer based on LM_BP neural network[C]//Proc. of the IEEE Sensors, 2016.
20	WANG S D , ZHU W L , SHEN Y J , et al. Temperature compensation for MEMS resonant accelerometer based on genetic algorithm optimized backpropagation neural network[J]. Sensors and Actuators Physical, 2020, 316, 112393. doi: 10.1016/j.sna.2020.112393
21	MA P , JIA Y D , PRASHANTH K G , et al. Effect of Si content on the microstructure and properties of Al-Si alloys fabricated using hot extrusion[J]. Journal of Materials Research, 2017, 32 (11): 2210- 2217. doi: 10.1557/jmr.2017.97

序号	名称	表达式	神经层
1	径向基函数	$\phi_{j}(x)=\mathrm{e}^{-\left\\|x-c_{j}\right\\|^{2} / 2 \sigma_{j}^{2}}$	隐藏层
2	对称S形函数	$\tan \operatorname{sig}(x)=\frac{2}{1+\mathrm{e}^{-2 x}}-1$	隐藏层
3	对数S形函数	$\log \operatorname{sig}(x)=\frac{1}{1+\mathrm{e}^{-x}}$	隐藏层
4	线性函数	$y=x$	隐藏层输出层
5	对称线性饱和函数	$y=\left\{\begin{array}{l}-1, x <-1 \\x, x \in[-1,1] \\1, x>1\end{array}\right.$	隐藏层输出层

提升倍数 P	输入模糊集合1	计算时间 T_P/s	输入模糊集合2	论域
指标1	PG	指标1	PG	(0, 1)
指标2	PB	指标2	PB	(1, 2)
指标3	PM	指标3	PM	(2, 3)
指标4	PS	指标4	PS	(3, 4)
指标5	PZ	指标5	PZ	(4, 5)
指标6	NZ	指标6	NZ	(5, 6)
指标7	NS	指标7	NS	(6, 7)
指标8	NM	指标8	NM	(7, 8)
指标9	NB	指标9	NB	(8, 9)
指标10	NG	指标10	NG	(9, 10)

输入模糊集合1	输入模糊集合2	输出模糊集合 (隐含层/输出层函数)	论域
元素1	元素1	径向基/线性	(0, 5)
元素2	元素2	对称S形/线性	(5, 15)
元素3	元素3	对数S形/线性	(15, 25)
元素4	元素4	均为线性	(25, 35)
元素5	元素5	对称线性饱和/线性	(35, 45)
元素6	元素6	径向基/对称线性饱和	(45, 55)
元素7	元素7	对称S形/对称线性饱和	(55, 65)
元素8	元素8	对数S形/对称线性饱和	(65, 75)
元素9	元素9	线性/对称线性饱和	(75, 85)
元素10	元素10	均为对称线性饱和	(85, 90)

输出层	隐含层	提升倍数P	计算时间T_P/s
线性函数	径向基函数	5.22×10³	4.22×10^-4
	对称S形函数	4.07×10¹¹	3.82×10^-4
	对数S形函数	5.29×10¹⁰	3.38×10^-4
	线性函数	1.65×10¹⁷	3.13×10^-4
	对称线性饱和函数	4.29×10¹¹	3.07×10^-4
对称线性饱和函数	径向基函数	68.53	3.67×10^-4
	对称S形函数	97.02	3.47×10^-4
	对数S形函数	99.35	3.58×10^-4
	线性函数	100.45	2.96×10^-4
	对称线性饱和函数	96.89	3.08×10^-4

P	模糊集合1	T_P(s)	模糊集合2
1.65×10¹⁷	PB	4.22×10^-4 3.82 ×10^-4 3.67×10^-4 3.58 ×10^-4	PB
5.29×10¹⁰ 4.07×10¹¹ 4.29 ×10¹¹	PM	-	-
5.22×10³	PS	3.38×10^-4 3.47×10^-4	PS
100.45	NS	3.13×10^-4 3.07 ×10^-4 2.96×10^-4 3.08 ×10^-4	NB
97.02 99.35 96.89	NM	-	-
68.53	NB	-	-