基于卡尔曼滤波器和多层感知器的大麦幼苗最优生长参数预测

doi:10.11959/j.issn.2096-3750.2021.00218

摘要/Abstract

摘要：

为了提高生长舱大麦幼苗的质量和种植效率，首先利用卡尔曼滤波算法对传感器采集的数据进行处理，有效降低了环境因素和传感器本身误差的影响，提高了采集数据的精度，保证生长舱的精确控制和准确的实验数据，然后利用多元非线性回归、径向基函数和多层感知器神经网络对不同条件下，大麦种子萌发生长约160 h后的平均生长高度、麦苗重量和种子重量干燥比进行分析比较，结果表明，多层感知器网络模型对数据的拟合效果最好。利用该模型预测最优环境时的大麦幼苗平均高度和麦苗种子重量比与实际种植效果基本一致，为生长舱大麦幼苗的种植提供一定参考。

关键词: 麦苗萌发, 卡尔曼滤波, 径向基函数, 多层感知器, 传感器

Abstract:

In order to improve the quality and planting efficiency of barley seedlings in the growth chamber, the Kalman filter algorithm was firstly used to process the data collected by the sensor, which effectively reduced the influence of environmental factors and the error of the sensor itself, improved the accuracy of the collected data, and ensured the precise control in the growth chamber and accurate test data.Then multiple nonlinear regression, radial basis function and multilayer perceptron neural network were used to analyze the average growth height, seedling weight and seed weight of barley seeds about 160 hours after germination under different conditions.The drying ratio was analyzed and compared.The results show that the multi-layer perceptron network model fits the data best.Using this model to predict the average height of barley seedlings and the ratio of seedling weight of barley seedlings in the optimal environment is basically consistent with the actual planting effect, which provides a certain reference for the planting of barley seedlings in the growth chamber.

Key words: barley seedling germination, Kalman filter, radial basis function, multilayer perceptron, sensor

中图分类号:

TP271

黄云龙, 李正权, 孙煜嘉. 基于卡尔曼滤波器和多层感知器的大麦幼苗最优生长参数预测[J]. 物联网学报, 2021, 5(4): 90-98.

Yunlong HUANG, Zhengquan LI, Yujia SUN. Prediction of optimal growth parameters of barley seedling based on Kalman filter and multilayer perceptron[J]. Chinese Journal on Internet of Things, 2021, 5(4): 90-98.

图/表 6

图1

图2

图3

图4

图5

表1

参考文献 26

[1]	梅玉芹, 俸继红, 杨建菊 ,等. 温度和盐对大麦种子萌发及幼苗生长的影响[J]. 种子世界, 2014(2): 36-37.
	MEI Y Q , FENG J H , YANG J J ,et al. Effects of temperature and salt on Seed Germination and seedling growth of barley[J]. Seed World, 2014(2): 36-37.
[2]	张敬云, 谢作如 . 利用 Arduino 探究种子萌发的最宜土壤湿度[J]. 中国信息技术教育, 2016(Z2): 89-91.
	ZHANG J Y , XIE Z R . Using Arduino to explore the optimal soil moisture for seed germination[J]. China Information Technology Edu-cation, 2016(Z2): 89-91.
[3]	WU Y B , YAN Z M , WU D F ,et al. Effect of NaCl stress on seed germination and seedling growth in wheat[C]// 2010 4th International Conference on Bioinformatics and Biomedical Engineering. IEEE, 2010: 1-3.
[4]	YI Z H , CUI J J , FU Y M ,et al. Effect of different light intensity on physiology,antioxidant capacity and photosynthetic characteristics on wheat seedlings under high CO2 concentration in a closed artificial ecosystem[J]. Photosynthesis Research, 2020,144(1): 23-34.
[5]	DOEHLERT D C , KU M S B , EDWARDS G E . Dependence of the post-illumination burst of CO2 on temperature,light,CO2,and O2 concentration in wheat (triticum aestivum)[J]. Physiologia Plantarum, 1979,46(4): 299-306.
[6]	武闯 . 仓库温湿度监控系统实现及测点定位精度研究[D]. 芜湖:安徽师范大学, 2016.
	WU C . Research on the realization of warehouse temperature and hu-midity monitoring system and the positioning accuracy of measuring points[D]. Wuhu:Anhui Normal University, 2016.
[7]	赵国强, 贾鹤鸣, 张森 ,等. 微型植物工厂温湿度场分析与传感器优化布局[J]. 森林工程, 2019,35(2): 61-68.
	ZHAO G Q , JIA H M , ZHANG S ,et al. Analysis of temperature and humidity field and layout optimization of sensor in micro plant facto-ry[J]. Forest Engineering, 2019,35(2): 61-68.
[8]	弓倩 . 基于IMM 的多传感器融合跟踪算法研究[D]. 西安:西安工业大学, 2012.
	GONG Q . A research on tracking algorithm of multi-sensor data fusion based on interactive multiple models (IMM)[D]. Xi'an:Xi'an Tech-nological University, 2012.
[9]	李海艳, 李维嘉, 黄运保 . 基于卡尔曼滤波的多传感器测量数据融合[J]. 武汉大学学报(工学版), 2011,44(4): 521-525,529.
	LI H Y , LI W J , HUANG Y B . Multisensor measured data fusion based on Kalman filtering[J]. Engineering Journal of Wuhan University, 2011,44(4): 521-525,529.
[10]	KOVVALI N , BANAVAR M , SPANIAS A . An introduction to Kalman filtering with MATLAB examples[C]// Synthesis Lectures on Signal Processing. Morgan ＆ Claypool, 2013: 79-81.
[11]	彭丁聪 . 卡尔曼滤波的基本原理及应用[J]. 软件导刊, 2009,8(11): 32-34.
	PENG D C . Basic principle and application of Kalman filter[J]. Soft-ware Guide, 2009,8(11): 32-34.
[12]	张谦, 景占荣 . 一种过程噪声自适应调节的卡尔曼滤波算法[J]. 电子测量技术, 2007,30(5): 18-20,31.
	ZHANG Q , JING Z R . Kalman filtering algorithm with process noise adjusting adaptively[J]. Electronic Measurement Technology, 2007,30(5): 18-20,31.
[13]	韩璐, 景占荣, 段哲民 . 一种基于预测滤波器的自适应卡尔曼滤波算法[J]. 火力与指挥控制, 2010,35(3): 110-113.
	HAN L , JING Z R , DUAN Z M . Application of adaptive Kalman filter to missile fixed SINS/GPS integrated navigation system based on pre-dictive filtering[J]. Fire Control ＆ Command Control, 2010,35(3): 110-113.
[14]	曾康华 . 计量经济学[M]. 北京: 清华大学出版社, 2016.
	ZENG K H . Econometrics[M]. Beijing: Tsinghua University Press, 2016.
[15]	赵坷 . 非线性模型解算及其应用研究[D]. 青岛:山东科技大学, 2014.
	ZHAO K . Nonlinear model solution and its application[D]. Qingdao:Shandong University of Science and Technology, 2014.
[16]	石国春 . 关于序列二次规划(SQP)算法求解非线性规划问题的研究[D]. 兰州:兰州大学, 2009.
	SHI G C . Research on algorithm of sequential quadratic programming for nonlinear programming problems[D]. Lanzhou:Lanzhou Univer-sity, 2009.
[17]	CHERRAK O , GHENNIOUI H , ABARKAN E ,et al. Levenberg-Marquardt algorithm[J]. Tutorial on LM Algorithm, 2004,11(1): 101-110.
[18]	何叶丹, 马昌凤, 范斌 . 求解非线性方程组的一种新的L-M方法[J]. 福建师范大学学报(自然科学版), 2014,30(2): 20-25,41.
	HE Y D , MA C F , FAN B . A new L-M method for the systems of non-linear equalities[J]. Journal of Fujian Normal University (Natural Science Edition), 2014,30(2): 20-25,41.
[19]	李艳 . 径向基函数及其应用[D]. 大连:大连理工大学, 2009.
	LI Y . Radial basis function and its application[D]. Dalian:Dalian Uni-versity of Technology, 2009.
[20]	朱明星, 张德龙 . RBF网络基函数中心选取算法的研究[J]. 安徽大学学报(自然科学版), 2000,24(1): 72-78.
	ZHU M X , ZHANG D L . Study on the algorithms of selecting the radial basis function center[J]. Journal of Anhui University (Natural Sciences), 2000,24(1): 72-78.
[21]	潘琪 . 径向基神经网络基函数中心确定方法改进研究[D]. 哈尔滨:东北农业大学, 2017.
	PAN Q . Research on establishing an improved method to determine the radial basis function center of radial basis function neural net-work[D]. Harbin:Northeast Agricultural University, 2017.
[22]	DU Y L , . Prediction of consumer price index based on RBF neural network[C]// 2018 Chinese Control And Decision Conference (CCDC). IEEE, 2018: 2858-2861.
[23]	ROSLI N , IBRAHIM R , ISMAIL I ,et al. Neural network architecture selection for efficient prediction model of gas metering system[C]// 2016 2nd IEEE International Symposium on Robotics and Manufacturing Automation (ROMA). IEEE, 2016: 1-5.
[24]	KASHANINEJAD M , DEHGHANI A A , KASHIRI M . Modeling of wheat soaking using two artificial neural networks (MLP and RBF)[J]. Journal of Food Engineering, 2009,91(4): 602-607.
[25]	NANDI A , JANA N D , DAS S . Improving the performance of neural networks with an ensemble of activation functions[C]// 2020 International Joint Conference on Neural Networks (IJCNN). IEEE, 2020: 1-7.
[26]	IZZELDIN H , ASIRVADAM V S , SAAD N . Enhanced conjugate gradient methods for training MLP-networks[C]// 2010 IEEE Student Conference on Research and Development (SCOReD). IEEE, 2010: 139-143.

误差参数		麦苗平均高度		麦苗种子重量干燥比
误差参数		均方误差MSE	平均相对误差MRE	均方误差MSE	平均相对误差MRE
	NLR	0.004 8	0.003 8	0.005 7	0.009 3
训练集	RBF	0.003 2	0.001 6	0.009 2	0.010 5
	MLP	0.004 6	0.003 8	0.004 7	0.008 9
	NLR	0.004 7	0.004 1	0.018 6	0.016 0
测试集	RBF	0.011 2	0.003 6	0.009 6	0.011 4
	MLP	0.004 9	0.003 7	0.005 8	0.010 1