CN103743435A

CN103743435A - Multi-sensor data fusion method

Info

Publication number: CN103743435A
Application number: CN201310721964.0A
Authority: CN
Inventors: 罗文广; 张晓亮
Original assignee: Guangxi University of Science and Technology
Current assignee: Guangxi University of Science and Technology
Priority date: 2013-12-23
Filing date: 2013-12-23
Publication date: 2014-04-23

Abstract

The invention discloses a multi-sensor data fusion method. Multi-sensors acquire data signals and the data signals are converted through an A/D converter so as to obtain digital signals; the digital signals undergo data filtering and pretreatment; and finally, feature extraction and algorithm fusion are carried out to obtain a result after multi-sensor data fusion. During the data-level multi-sensor data fusion, two same sensors simultaneously measure a same indoor environment quality factor. Thus, real-time performance and measurement accuracy of sensors of the same type are simultaneously raised. By the method, original measurement data of sensors is fully utilized; and fusion of information, such as error of mean square, measurement accuracy and the like, of sensors is carried out. It is not required to know any experiment knowledge about sensor measurement data. Drifting and noise of sensors are inhibited to some extent, and measurement accuracy of a system is raised.

Description

A multi-sensor data fusion method

技术领域technical field

本发明涉及一种多传感器数据融合方法。The invention relates to a multi-sensor data fusion method.

背景技术Background technique

目前，随着多传感器数据融合技术的发展和待解决问题的复杂程度的提高，单一融合算法的局限性日益暴露，将多种数据融合算法进行结合成为多传感器数据融合技术的发展趋势。而目前的自适应加权融合算法是对同类传感器数据融合的常用算法。在室内环境品质各个因素的测量过程中，得到的实际测量值存在噪声误差，经常用均方误差来评判实际测量数据的精确程度。At present, with the development of multi-sensor data fusion technology and the increasing complexity of the problems to be solved, the limitations of a single fusion algorithm are increasingly exposed, and the combination of multiple data fusion algorithms has become the development trend of multi-sensor data fusion technology. The current adaptive weighted fusion algorithm is a commonly used algorithm for the fusion of similar sensor data. In the process of measuring various factors of indoor environmental quality, the actual measurement values obtained have noise errors, and the mean square error is often used to judge the accuracy of the actual measurement data.

发明内容Contents of the invention

本发明目的是针对现有技术存在的缺陷提供一种多传感器数据融合方法。The purpose of the present invention is to provide a multi-sensor data fusion method aiming at the defects existing in the prior art.

本发明为实现上述目的，采用如下技术方案：一种多传感器数据融合方法，多传感器获取数据信号，然后通过A/D转换器进行转换,得到数字信号；然后进行数据滤波和预处理，然后再经过特征提取和算法融合得到多传感器数据融合后的结果。In order to achieve the above object, the present invention adopts the following technical solutions: a multi-sensor data fusion method, multi-sensors obtain data signals, and then convert them through A/D converters to obtain digital signals; then perform data filtering and preprocessing, and then After feature extraction and algorithm fusion, the result of multi-sensor data fusion is obtained.

进一步的，所述多传感器为采用两个传感器，则两个传感器的方差分别为σ₁ ²、σ₂ ²，所要估计的真值为X，传感器的测量值分别为x₁、x₂，彼此相互独立，并且

是X的无偏估计，加权因子分别为w₁、w₂，融合后的加权因子和值分别满足：Further, the multi-sensor adopts two sensors, and the variances of the two sensors are σ ₁ ² and σ ₂ ² respectively, the true value to be estimated is X, and the measured values of the sensors are x ₁ and x ₂ , respectively. independent of each other, and

is the unbiased estimate of X, the weighting factors are w ₁ and w ₂ , and the weighting factors and values after fusion satisfy:

w₁+w₂=1 （2-1）w ₁ +w ₂ =1 (2-1)

$\overset{^^}{x x} = = [\begin{matrix} {w w}_{11} & {w w}_{22} \end{matrix}] {[\begin{matrix} {x x}_{11} & {x x}_{22} \end{matrix}]}^{T T} = = {Σ Σ}_{i i = = 11}^{22} {w w}_{i i} {x x}_{i i} - - - - - - ((22 - - 22))$

由式（2-2）可知

为X的无偏估计，并且融合值是各个传感器测量值的线性函数；It can be seen from formula (2-2)

is an unbiased estimate of X, and the fusion value is a linear function of the individual sensor measurements;

各个传感器方差为：The variance of each sensor is:

${σ σ}^{22} = = E E. [[\frac{11}{44} {Σ Σ}_{j j = = 11}^{44} {(({x x}_{j j} - - \overset{&OverBar; &OverBar;}{x x}))}^{22}]] - - - - - - ((22 - - 33))$

式（2-3）中：i为测量次数；In formula (2-3): i is the number of measurements;

总均方差为：The total mean square error is:

${σ σ}^{22} = = E E. [[{((x x - - \overset{^^}{x x}))}^{22}]] = = E E. [[{Σ Σ}_{i i = = 11}^{22} {w w}_{i i}^{22} {((x x - - {x x}_{i i}))}^{22} + + 22 {Σ Σ}_{\underset{i i &NotEqual; &NotEqual; j j}{i i = = 11,, j j = = 11}}^{22} {w w}_{i i} {x x}_{j j} ((x x - - {x x}_{i i})) ((x x - - {x x}_{j j})) - - - - - - ((22 - - 44))$

由式（2-4）可知，总均方差σ²是关于各个加权因子的多元二次函，必然存在最小值，根据拉格朗日定理极值理论可知，在总均方差σ²最小时对应的加权因子为：It can be seen from formula (2-4) that the total mean square error σ ² is a multivariate quadratic function about each weighting factor, and there must be a minimum value. According to the extreme value theory of Lagrangian theorem, when the total mean square error σ ² is the smallest, it corresponds to The weighting factor for is:

${w w}_{i i}^{' '} = = \frac{11}{{σ σ}_{i i}^{22} {Σ Σ}_{i i = = 11}^{22} \frac{11}{{σ σ}_{i i}^{22}}} - - - - - - ((22 - - 55))$

此时所对应的最小均方差为：The corresponding minimum mean square error at this time is:

${σ σ}_{min min}^{22} = = \frac{11}{{Σ Σ}_{i i = = 11}^{22} \frac{11}{{σ σ}_{i i}^{22}}} - - - - - - ((22 - - 66)) . .$

本发明的有益效果：本发明在数据级多传感器数据融合中采用两个相同的传感器同时测量同一个室内环境品质因数，这样同时提高了同类传感器的实时性和测量精度。该方法充分利用了传感器的原始测量数据，讲传感器的均方误差、测量精度等信息进行融合，不要求知道传感器测量数据的任何实验知识，它在一定程度上抑制了传感器的漂移和噪声，提高了系统的测量精度。Beneficial effects of the present invention: the present invention uses two identical sensors to simultaneously measure the same indoor environmental quality factor in data-level multi-sensor data fusion, thus simultaneously improving the real-time performance and measurement accuracy of similar sensors. This method makes full use of the original measurement data of the sensor, and fuses information such as the mean square error and measurement accuracy of the sensor. It does not require any experimental knowledge of the sensor measurement data. It suppresses the drift and noise of the sensor to a certain extent and improves the measurement accuracy of the system.

具体实施方式Detailed ways

本发明涉及一种多传感器数据融合方法，多传感器获取数据信号，然后通过A/D转换器进行转换,得到数字信号；然后进行数据滤波和预处理，然后再经过特征提取和算法融合得到多传感器数据融合后的结果。The invention relates to a multi-sensor data fusion method. Multi-sensors acquire data signals, and then convert them through an A/D converter to obtain digital signals; then perform data filtering and preprocessing, and then obtain multi-sensors through feature extraction and algorithm fusion. The result of data fusion.

进一步的，所述多传感器为采用两个传感器，则两个传感器的方差分别为σ₁ ²、σ₂ ²，所要估计的真值为X，传感器的测量值分别为x₁、x₂，彼此相互独立，并且x?是X的无偏估计，加权因子分别为w₁、w₂，融合后的加权因子和值分别满足：Further, the multi-sensor adopts two sensors, and the variances of the two sensors are σ ₁ ² and σ ₂ ² respectively, the true value to be estimated is X, and the measured values of the sensors are x ₁ and x ₂ , respectively. are independent of each other, and x? is an unbiased estimate of X, the weighting factors are w ₁ and w ₂ respectively, and the weighting factors and values after fusion satisfy respectively:

w₁+w₂=1 （2-1）w ₁ +w ₂ =1 (2-1)

由式（2-2）可知

各个传感器方差为：The variance of each sensor is:

总均方差为：The total mean square error is:

以上所述仅为本发明的较佳实施例，并不用以限制本发明，凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection of the present invention. within range.

Claims

1. A multi-sensor data fusion method is characterized in that multi-sensors acquire data signals, then convert them through an A/D converter to obtain digital signals; then perform data filtering and preprocessing, and then undergo feature extraction and algorithm fusion The result of multi-sensor data fusion is obtained.

2. A multi-sensor data fusion method according to claim 1, wherein the multi-sensor uses two sensors, and the variances of the two sensors are σ ₁ ² and σ ₂ ² respectively, and the estimated The true value is X, the measured values of the sensor are x ₁ , x ₂ , which are independent of each other, and x? is an unbiased estimate of X, the weighting factors are w ₁ , w ₂ , and the fused weighting factors and values respectively satisfy :

w ₁ +w ₂ =1 (2-1)

\overset{^^}{x x} = = [\begin{matrix} {w w}_{11} & {w w}_{22} \end{matrix}] {[\begin{matrix} {x x}_{11} & {x x}_{22} \end{matrix}]}^{T T} = = {Σ Σ}_{i i = = 11}^{22} {w w}_{i i} {x x}_{i i} - - - - - - ((22 - - 22))

It can be seen from formula (2-2) is an unbiased estimate of X, and the fusion value is a linear function of the individual sensor measurements;

The variance of each sensor is:

{σ σ}^{22} = = E E. [[\frac{11}{44} {Σ Σ}_{j j = = 11}^{44} {(({x x}_{j j} - - \overset{&OverBar; &OverBar;}{x x}))}^{22}]] - - - - - - ((22 - - 33))

In formula (2-3): i is the number of measurements;

The total mean square error is:

{σ σ}^{22} = = E E. [[{((x x - - \overset{^^}{x x}))}^{22}]] = = E E. [[{Σ Σ}_{i i = = 11}^{22} {w w}_{i i}^{22} {((x x - - {x x}_{i i}))}^{22} + + 22 {Σ Σ}_{\underset{i i &NotEqual; &NotEqual; j j}{i i = = 11,, j j = = 11}}^{22} {w w}_{i i} {x x}_{j j} ((x x - - {x x}_{i i})) ((x x - - {x x}_{j j})) - - - - - - ((22 - - 44))

It can be seen from formula (2-4) that the total mean square error σ ² is a multivariate quadratic function about each weighting factor, and there must be a minimum value. According to the extreme value theory of Lagrangian theorem, when the total mean square error σ ² is the smallest, it corresponds to The weighting factor for is:

{w w}_{i i}^{' '} = = \frac{11}{{σ σ}_{i i}^{22} {Σ Σ}_{i i = = 11}^{22} \frac{11}{{σ σ}_{i i}^{22}}} - - - - - - ((22 - - 55))

The corresponding minimum mean square error at this time is:

{σ σ}_{min min}^{22} = = \frac{11}{{Σ Σ}_{i i = = 11}^{22} \frac{11}{{σ σ}_{i i}^{22}}} - - - - - - ((22 - - 66)) . .