Research on Temperature Compensation Technology of Fiber Optic Gyroscope based on ISCSO-BP Neural Network Model

doi:10.16182/j.issn1004731x.joss.25-0073

Abstract

Abstract:

To address the issue that changes in ambient temperature significantly affect the output accuracy of the fiber optic gyro (FOG), which causes zero bias drift, increases measurement errors, and limits their application accuracy in complex environments, a temperature compensation model based on BP neural networks was proposed. To improve the performance of neural networks, the sand cat swarm optimization (SCSO) was improved, and the improved SCSO (ISCSO) was used to optimize the weights and thresholds of BP neural networks. Experimental results show that using the ISCSO-BPNN temperature compensation model to compensate for the gyro's temperature errors significantly improves the zero bias stability and overall compensation accuracy compared with other comparative algorithms.

Key words: fiber optic gyroscope, temperature error, temperature compensation, BP neural network, sand cat swarm optimization(SCSO)

CLC Number:

TP391.9

Zhang Zhili, Liu Jin, Zhou Zhaofa, Liang Zhe, Zhang Yunhao. Research on Temperature Compensation Technology of Fiber Optic Gyroscope based on ISCSO-BP Neural Network Model[J]. Journal of System Simulation, 2025, 37(11): 2904-2917.

Figures/Tables 18

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Table 1

Information on test functions

函数	函数名称	基准测试函数	维度	搜索空间	最优值
F1	Rosenbrock	$f 1 (x) = 100 (x 12 - x 2) 2 + (1 - x 1) 2$	2	[-2.048, 2.048]²	0
F2	Michalewicz	$f 2 (x) = 21.5 + x 1 s i n (4 π x 1) + x 2 s i n (20 π x 2)$	2	[-3.0, 12.1; 4.1, 5.8]	-38.85
F3	Shaffer F6	$f 3 (x) = 0.5 - s i n 2 x 12 + x 22 - 0.5 1 + 0.001 (x 12 + x 22) 2$	2	[-100, 100]²	1
F4	Camel	$f 4 (x) = (4 - 2.1 x 12 + x 14 3) x 12 + x 1 x 2 + (- 4 + 4 x 22) x 22$	2	[-3, 3; -2, 2]	-1.031 6
F5	Quartic	$f 5 (x) = ∑ i = 1 30 i x 4 + G a u s s (0,1)$	30	[-1.28, 1.28]³⁰	0
F6	Schwefel's Problem2.22	$f 6 (x) = ∑ i = 1 30 \| x i \| + ∏ i = 1 30 \| x i \|$	30	[-10, 10]³⁰	0
F7	Rastrigin	$f 7 (x) = 100 - ∑ i = 1 10 [10 c o s (2 π x i) - x i 2]$	10	[-5.12, 5.12]¹⁰	0

Table 1

Table 2

Fig. 7

Fig. 8

Fig. 9

Fig. 10

Fig. 11

Table 3

Table 4

Table 5

Table 6

Table 7

References 15

[1]	Shupe D M. Thermally Induced Nonreciprocity in the Fiber-optic Interferometer[J]. Applied Optics, 1980, 19(5): 654-655.
[2]	Shupe D M. Fiber Resonator Gyroscope: Sensitivity and Thermal Nonreciprocity[J]. Applied Optics, 1981, 20(2): 286-289.
[3]	冯丽爽, 南书志, 金靖. 光纤陀螺温度建模及补偿技术研究[J]. 宇航学报, 2006, 27(5): 939-941, 1049.
	Feng Lishuang, Shuzhi Nan, Jin Jing. Research on Modeling and Compensation Technology for Temperature Errors of FOG[J]. Journal of Astronautics, 2006, 27(5): 939-941, 1049.
[4]	韩冰, 林玉荣, 邓正隆. 光纤陀螺温度漂移误差的建模与补偿综述[J]. 中国惯性技术学报, 2009, 17(2): 218-224.
	Han Bing, Lin Yurong, Deng Zhenglong. Overview 0n Modeling and Compensation of FOG Temperature Drift[J]. Journal of Chinese Inertial Technology, 2009, 17(2): 218-224.
[5]	赵深, 何巍, 辛璟焘, 等. 基于CSAPSO-BP神经网络的光纤陀螺温度补偿研究[J]. 压电与声光, 2023, 45(4): 589-594.
	Zhao Shen, He Wei, Xin Jingtao, et al. Research on Temperature Compensation of Fiber Optic Gyroscope Based on CSAPSO-BP Neural Network[J]. Piezoelectrics & Acoustooptics, 2023, 45(4): 589-594.
[6]	Jiang Yufei, He Kunpeng, Tan Panlong, et al. Enhancing FOG Temperature Compensation Using LSTM Method[C]//2023 42nd Chinese Control Conference (CCC). Piscataway: IEEE, 2023: 3878-3882.
[7]	Song Yuhuan, Lu Lichuan, Wang Gui, et al. Study of Modeling and Temperature Compensation Technology of Airborne FOG Components[C]///International Conference on Mechanisms and Robotics (ICMAR 2022). Bellingham: SPIE, 2022: 123310X.
[8]	姚成敏, 朱节中, 杨再强. 蜣螂优化算法在Canny边缘检测算法中的应用[J]. 国外电子测量技术, 2024, 43(4): 143-151.
	Yao Chengmin, Zhu Jiezhong, Yang Zaiqiang. Application of Dung Beetle Optimization Algorithm in Canny Edge Detection Algorithm[J]. Foreign Electronic Measurement Technology, 2024, 43(4): 143-151.
[9]	佟林, 覃方君, 冯卡力, 等. 基于粒子群优化算法的光纤陀螺温度误差分段补偿方法[J]. 中国惯性技术学报, 2019, 27(4): 505-509.
	Tong Lin, Qin Fangjun, Feng Kali, et al. Segmentation Compensation Method for FOG Temperature Error Based on Particle Swarm Optimization Algorithm[J]. Journal of Chinese Inertial Technology, 2019, 27(4): 505-509.
[10]	王瑞, 郑百东, 李飞, 等. ELM-AdaBoost模型在光纤陀螺温度误差补偿中的应用[J]. 兵工自动化, 2024, 43(2): 63-68.
	Wang Rui, Zheng Baidong, Li Fei, et al. Application of ELM-AdaBoost Model in Temperature Error Compensation of Fiber Optic Gyroscope[J]. Ordnance Industry Automation, 2024, 43(2): 63-68.
[11]	贾朔, 车昊, 李洋, 等. 基于BP神经网络的光纤陀螺误差补偿方法[J]. 物理与工程, 2020, 30(4): 116-120.
	Jia Shuo, Che Hao, Li Yang, et al. The Fiber Optic Gyro Error Compensation by Using BP Neural Network[J]. Physics and Engineering, 2020, 30(4): 116-120.
[12]	Seyyedabbasi Amir, Kiani Farzad. Sand Cat Swarm Optimization: A Nature-inspired Algorithm to Solve Global Optimization Problems[J]. Engineering With Computers, 2023, 39(4): 2627-2651.
[13]	蒋开正, 吕丽平. 改进沙猫群优化算法优化堆叠降噪自动编码器的发动机故障诊断[J]. 机械设计, 2023, 40(8): 56-62.
	Jiang Kaizheng, Liping Lü. Optimization of Fault Diagnosis for Stacked Denoising Auto Encoder's Engine by Improved Sand Cat Swarm Optimization Algorithm[J]. Journal of Machine Design, 2023, 40(8): 56-62.
[14]	张志利, 刘瑾, 周召发, 等. 基于沙猫群优化算法和BP模型的光纤陀螺温度补偿研究[J]. 火箭军工程大学学报, 2024, 38(4): 54-60, 67.
	Zhang Zhili, Liu Jin, Zhou Zhaofa, et al. A Fiber Optic Gyroscope Temperature Compensation Method Based on SCSO-BP Model[J]. Journal of Rocket Force University of Engineering, 2024, 38(4): 54-60, 67.
[15]	中国人民解放军总装备部. 光纤陀螺仪测试方法: [S]. 北京: 总装备部军标出版发行部, 2015: 1-35.
	General Equipment Department of the Chinese People's Liberation Army. Measurement Methods for Fiber Optic Gyroscope: [S]. Beijing: General Equipment Department Military Standard Publishing and Distribution Department, 2015: 1-35.

函数	指标	SCSO	DBO	WOA	ZOA	ISCSO
F1	Mean	5.854 5×10^-4	5.967 8×10^-4	5.656 2×10^-4	6.462 0×10^-4	5.204 3×10^-4
	Std	3.049 87×10^-4	3.095 7×10^-4	3.223 8×10^-4	2.738 3×10^-4	2.934 8×10^-4
	Cr/%	100	100	100	99	100
	Iter	59.24	69.17	35.60	77.12	40.06
F2	Mean	-38.797 2	-38.735 3	-38.735 7	-38.425 1	-38.804 4
	Std	0.058 2	0.046 26	0.045 0	0.752 3	0.047 0
	Cr/%	54	5	4	50	58
	Iter	471.30	480.91	487.37	420.57	457.83
F3	Mean	0.997 5	0.997 5	0.996 7	0.997 5	0.997 7
	Std	1.395 4×10^-11	1.276 5×10^-11	3.744 3×10^-3	6.291 8×10^-12	7.148 6×10^-12
	Cr/%	0	0	0	0	0
	Iter	500	500	500	500	500
F4	Mean	-1.031 1	-1.031 1	-1.031 1	-1.031 63	-1.031 3
	Std	2.955 5×10^-4	2.844 3×10^-4	2.904 0×10^-4	2.781 0×10^-4	2.700 3×10^-4
	Cr/%	100	100	100	100	100
	Iter	8.69	8.61	13.37	6.92	7.28
F5	Mean	6.388 8×10^-4	1.223 3×10^-3	3.884 2×10^-3	5.932 3×10^-4	5.462 6×10^-4
	Std	3.750 5×10^-4	7.078 2×10^-4	3.916 8×10^-3	3.134 9×10^-4	2.840 8×10^-4
	Cr/%	97	46	31	100	100
	Iter	142.92	411.73	433.61	66.59	60.17
F6	Mean	5.083 4×10^-4	6.523 1×10^-4	7.947 3×10^-4	7.498 2×10^-4	4.576 9×10^-4
	Std	3.001 9×10^-4	2.485 0×10^-4	2.444 8×10^-4	2.717 6×10^-4	2.825 8×10^-4
	Cr/%	100	100	100	100	100
	Iter	20.61	47.67	103	20.31	8.04
F7	Mean	3.902 1×10^-4	3.953 3×10^-4	3.863 2×10^-4	2.654 8×10^-4	2.414 8×10^-4
	Std	3.297 4×10^-4	3.472 4 ×10^-4	3.196 4×10^-4	2.830 3×10^-4	2.520 9×10^-4
	Cr/%	100	89	94	95	100
	Iter	13.86	113.07	163.62	69.44	4.96

模型	补偿前	补偿后
BP	0.008 4	0.018 500 0
DBO-BP	0.008 4	0.005 219 0
SCSO-BP	0.008 4	0.004 298 5
ISCSO-BP	0.008 4	0.003 090 9

模型	补偿前	补偿后
BP	0.017 4	0.031 100 0
DBO-BP	0.017 4	0.006 608 9
SCSO-BP	0.017 4	0.005 313 2
ISCSO-BP	0.017 4	0.003 717 7

模型	补偿前	补偿后
BP	0.010 8	0.012 800 0
DBO-BP	0.010 8	0.005 521 2
SCSO-BP	0.010 8	0.004 875 4
ISCSO-BP	0.010 8	0.003 236 3

模型	补偿前	补偿后
BP	0.013 9	0.018 000 0
DBO-BP	0.013 9	0.006 188 3
SCSO-BP	0.013 9	0.005 642 1
ISCSO-BP	0.013 9	0.004 236 3