作者：朱菊香1, 谷卫2, 罗丹悦2, 潘斐2, 张赵良1
作者单位：1. 无锡学院轨道交通学院，江苏 无锡 214105;
2. 南京信息工程大学自动化学院，江苏 南京 210000
关键词：多传感器数据融合;卡尔曼滤波;环境监测;粒子群;bp神经网络
摘要：
针对室内环境监测中单一传感器测量数据精度低、可靠性差的问题，提出一种基于粒子群（particle swarm optimization，pso）优化bp神经网络多传感器数据融合算法。首先使用防脉冲干扰平均滤波算法来消除检测数据中的异常数据和噪声数据。其次，利用卡尔曼滤波算法对多同类传感器进行数据级融合，有效地降低因噪声干扰导致的测量误差，为异质传感器进行决策级融合提供最佳数据。最后，采用pso优化bp神经网络算法进行决策级融合。实验结果表明，基于pso优化bp神经网络多传感器数据融合算法对测试样本的平均绝对百分比误差（mean absolute percentage error, mape）和拟合度（r2）均优于bp神经网络和自适应加权（adaptive weighted, aw）优化bp神经网络，且运行时间比bp神经网络以及aw-bp神经网络分别短69.31%、50.36%。经验证，基于pso优化bp神经网络多传感器数据融合算法具有更高的融合精度，同时缩短了算法的运行时间。
multi-sensor data fusion based on pso optimized bp neural network
zhu juxiang1, gu wei2, luo danyue2, pan fei2, zhang zhaoliang1
1. school of rail transit, wuxi university, wuxi 214105, china;
2. school of automation, nanjing university of information technology, nanjing 210000, china
abstract: aiming at the problems of low precision and poor reliability of single sensor measurement data in indoor environmental monitoring, a multi-sensor data fusion algorithm based on particle swarm optimization (pso) optimization was proposed. firstly, the anti-pulse interference averaging filtering algorithm is used to eliminate abnormal data and noise data in the detected data. secondly, the kalman filter algorithm is used to perform data-level fusion of multiple similar sensors, which can effectively reduce the measurement error caused by noise interference and provide the best data for decision-level fusion of heterogeneous sensors. finally, the pso optimization bp neural network algorithm is used for decision-level fusion. the experimental results show that the mean absolute percentage error (mape) and fit (r2) of the test samples based on the pso optimized bp neural network multi-sensor data fusion algorithm are better than those of the bp neural network and adaptive weighed (aw) optimizes the bp neural network, and the running time is 69.31% and 50.36% shorter than the bp neural network and the aw-bp neural network, respectively. it has been verified that the multi-sensor data fusion algorithm based on pso optimized bp neural network has higher fusion accuracy and shortens the running time of the algorithm.
keywords: multi-sensor data fusion;kalman filter;environmental monitoring;particle swarm;bp neural network
2022, 48(8):94-100 收稿日期: 2022-03-13;收到修改稿日期: 2022-05-18
基金项目: 江苏省高等学校自然科学研究项目（21kjb460005）；南京信息工程大学滨江学院人才启动经费资助项目（2019r018）；南京信息工程大学滨江学院自然科学类项目（2019bjyng002）
作者简介: 朱菊香（1979-），女，江苏常州市人，副教授，硕士生导师，研究方向为自动化及控制技术、检测技术
参考文献
[1] deng z, wang j. multi-sensor data fusion based on improved analytic hierarchy process[j]. ieee access, 2020, 8: 9875-9895
[2] xu y, chen b z, hu z c. research for multi-sensor data fusion based on huffman tree clustering algorithm in greenhouses[j]. international journal of embedded systems, 2016, 8(1): 34-38
[3] zhang b c, lin j q, chang z c, et al. the application of multi sensor data fusion based on the improved bp neural network algorithm[c]// 28th chinese control and conference (2016 ccdc) . ieee, 2016: 1150-1154.
[4] zhang k, wang z, guo l, et al. an asynchronous data fusion algorithm for target detection based on multi-sensor networks[j]. ieee access, 2020, 8: 59511-59523
[5] ferrer-cid p, barcelo-ordinas j m, garcia-vidal j, et al. multisensor data fusion calibration in iot air pollution platforms[j]. ieee internet of things journal, 2020, 7(4): 3124-3132
[6] guo l, wang g, yu x. design for indoor environment monitoring system based on embedded system and multi-sensor data fusion algorithm[j]. international journal of smart home, 2016, 10(1): 31-40
[7] 高尚, 演明, 张洪泉. 基于gd-fnn算法的危化品气体传感器阵列数据融合[j]. 传感器与微系统, 2020, 39(9): 146-149 153
[8] 王玉钟, 王焕钦, 胡俊涛, 等. 基于多传感器融合技术的微型空气质量监测系统优化设计(英文)[j]. 大气与环境光学学报, 2021, 16(4): 349-357
[9] 叶瑾, 许枫, 杨娟, 等. 一种基于多传感器的复合量测imm-ekf数据融合算法[j]. 电子学报, 2020, 48(12): 2326-2330
[10] wu j, yang s x. intelligent control of bulk tobacco curing schedule using ls-svm-and anfis-based multi-sensor data fusion approaches[j]. sensors, 2019, 19(8): 1778
[11] 聂珲, 陈海峰. 基于nb-iot环境监测的多传感器数据融合技术[j]. 传感技术学报, 2020, 33(1): 144-152
[12] 辛宇, 陈兴, 许素安, 等. 基于ga-pso-bp的车辆动态称重算法研究[j]. 中国测试, 2021, 47(7): 26-30
[13] 杨涛锋, 彭艺. 基于改进pso的arima-svm空气质量预测研究[j]. 云南大学学报(自然科学版), 2020, 42(5): 854-862
[14] 崔萌洁, 卢文科, 左锋. 基于pso-bp模型的扩散硅压力传感器温度补偿[j]. 中国测试, 2019, 45(11): 95-100 125
