CN114531729A - Positioning method, system, storage medium and device based on channel state information - Google Patents
Positioning method, system, storage medium and device based on channel state information Download PDFInfo
- Publication number
- CN114531729A CN114531729A CN202210432885.7A CN202210432885A CN114531729A CN 114531729 A CN114531729 A CN 114531729A CN 202210432885 A CN202210432885 A CN 202210432885A CN 114531729 A CN114531729 A CN 114531729A
- Authority
- CN
- China
- Prior art keywords
- phase data
- training
- neural network
- output
- network model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000012549 training Methods 0.000 claims abstract description 133
- 238000003062 neural network model Methods 0.000 claims abstract description 47
- 239000011159 matrix material Substances 0.000 claims description 37
- 230000006870 function Effects 0.000 claims description 29
- 238000004422 calculation algorithm Methods 0.000 claims description 20
- 210000002569 neuron Anatomy 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 10
- 230000009466 transformation Effects 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 abstract description 48
- 238000001514 detection method Methods 0.000 abstract description 31
- 230000008569 process Effects 0.000 description 10
- 230000007613 environmental effect Effects 0.000 description 7
- 230000000694 effects Effects 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000033001 locomotion Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000011478 gradient descent method Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000035508 accumulation Effects 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
- H04W64/006—Locating users or terminals or network equipment for network management purposes, e.g. mobility management with additional information processing, e.g. for direction or speed determination
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
- G01S5/0252—Radio frequency fingerprinting
- G01S5/02521—Radio frequency fingerprinting using a radio-map
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Remote Sensing (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Signal Processing (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Radar, Positioning & Navigation (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Image Analysis (AREA)
Abstract
本发明提供一种基于信道状态信息的定位方法、系统、存储介质及设备，包括获取与多个位置一一对应的多组子载波相位数据，根据校准后的子载波相位数据构建训练集；将训练集输入深度神经网络模型中进行训练，得到最优深度神经网络模型及对应目标位置的指纹数据，根据指纹数据建立指纹库；将移动目标的子载波相位数据与指纹库进行匹配，确定移动目标的位置。本发明中的基于信道状态信息的定位方法、系统、存储介质及设备，通过深度神经网络训练减少子载波相位数据在低信噪比环境中因环境等因素所受到的的影响，提高了子载波相位数据用于目标位置检测的准确性和可靠性。
The present invention provides a positioning method, system, storage medium and device based on channel state information, including acquiring multiple sets of sub-carrier phase data corresponding to multiple locations one-to-one, and constructing a training set according to the calibrated sub-carrier phase data; The training set is input into the deep neural network model for training, and the optimal deep neural network model and the fingerprint data corresponding to the target position are obtained, and the fingerprint database is established according to the fingerprint data; the sub-carrier phase data of the moving target is matched with the fingerprint database to determine the moving target. s position. The positioning method, system, storage medium and device based on the channel state information in the present invention reduce the influence of the sub-carrier phase data in the low signal-to-noise ratio environment due to factors such as the environment through deep neural network training, and improve the sub-carrier phase data. Phase data is used for accuracy and reliability of target position detection.
Description
技术领域technical field
本发明涉及无线信号及信息处理技术领域，特别涉及一种基于信道状态信息的定位方法、系统、存储介质及设备。The present invention relates to the technical field of wireless signal and information processing, and in particular, to a positioning method, system, storage medium and device based on channel state information.
背景技术Background technique
随着环境检测技术研究的发展，CSI(ChannelStationInformation，信道状态信息)成为环境检测的典型应用。CSI是描述无线信道状态的信息集合，其中包括子载波信号强度、振幅、相位和信号延迟等信息。与以往基于RSSI(ReceivedSignalStrengthIndicator)的环境定位和测距技术相比，采用CSI会拥有更高的细粒度，即对目标环境可做出更精细的描述，同时对于外界环境因素的干扰，基于CSI的定位和测距技术稳定性也会更好，不会因为受到多径效应而导致精确度不高等问题，具有更好的定位、测距效果和服务能力。With the development of environmental detection technology research, CSI (ChannelStationInformation, channel state information) has become a typical application of environmental detection. CSI is a collection of information describing the state of a wireless channel, including sub-carrier signal strength, amplitude, phase, and signal delay. Compared with the previous environment positioning and ranging technology based on RSSI (Received Signal Strength Indicator), the use of CSI will have higher fine-grainedness, that is, a more detailed description of the target environment can be made. The stability of positioning and ranging technology will also be better, and there will be no problems such as low accuracy due to multipath effects, and it will have better positioning, ranging effects and service capabilities.
应用CSI进行环境检测时，由于在CSI发射接收过程中存在设备接收时频率不能准确同步等问题，不能直接使用接收到的CSI，通过对其进行合理的变换和处理，能够消除接收的CSI的信号误差，在环境检测等应用上获得更高的检测精度。When using CSI for environmental detection, the received CSI cannot be directly used due to the problem that the frequency cannot be accurately synchronized during the CSI transmission and reception process, and the received CSI signal can be eliminated by reasonable transformation and processing. error, and obtain higher detection accuracy in applications such as environmental detection.
但在实际应用中，除去接收的误差，信号在多障碍多路径的复杂环境进行传输过程中，还会受到阻挡、干扰、噪声等因素影响，不仅会造成CSI发射和接收存在一定的偏差，同时对于后续的相位校准、建立指纹库、指纹匹配都有着一定的影响。因此，利用直接测量、直接使用接收数据这样的传统方法建立指纹库，在复杂环境下，对移动目标的位置检测的可靠性和准确性会变差。However, in practical applications, except for the reception error, the signal will be affected by factors such as blocking, interference, noise, etc. during the transmission process in the complex environment with multiple obstacles and multiple paths, which will not only cause a certain deviation between CSI transmission and reception, but also It has a certain impact on the subsequent phase calibration, establishment of fingerprint database, and fingerprint matching. Therefore, using the traditional methods such as direct measurement and direct use of received data to establish a fingerprint database, in a complex environment, the reliability and accuracy of the position detection of the moving target will be deteriorated.
发明内容SUMMARY OF THE INVENTION
基于此，本发明的目的是提供一种基于信道状态信息的定位方法、系统、存储介质及设备，解决背景技术中在复杂环境下，移动目标位置检测的可靠性和准确性会变差的问题。Based on this, the purpose of the present invention is to provide a positioning method, system, storage medium and device based on channel state information, so as to solve the problem of poor reliability and accuracy of moving target position detection in complex environments in the background art .
本发明一方面提供一种基于信道状态信息的定位方法，方法包括：One aspect of the present invention provides a positioning method based on channel state information, the method comprising:
将目标位置预先划分为多个位置，获取与多个位置一一对应的多组子载波相位数据，对多组子载波相位数据进行校准，根据校准后的子载波相位数据构建训练集；Divide the target position into multiple positions in advance, obtain multiple sets of sub-carrier phase data corresponding to the multiple positions, calibrate the multiple sets of sub-carrier phase data, and construct a training set according to the calibrated sub-carrier phase data;
构建深度神经网络模型，将训练集输入深度神经网络模型中进行训练，得到最优深度神经网络模型及对应目标位置的指纹数据；Build a deep neural network model, input the training set into the deep neural network model for training, and obtain the optimal deep neural network model and fingerprint data corresponding to the target position;
根据指纹数据建立指纹库，获取移动目标的子载波相位数据，将移动目标的子载波相位数据与指纹库进行匹配，根据匹配结果确定移动目标的位置。The fingerprint database is established according to the fingerprint data, the sub-carrier phase data of the moving target is obtained, the sub-carrier phase data of the moving target is matched with the fingerprint database, and the position of the moving target is determined according to the matching result.
本发明中的基于信道状态信息的定位方法，通过构建深度神经网络训练模型和预先划分的多个位置的子载波相位数据，通过深度神经网络训练，得到最优深度神经网络模型，通过深度神经网络训练能够减少子载波相位数据在低信噪比环境中因环境等因素所受到的的影响，再根据训练后校准的子载波相位数据建立对应目标位置的指纹库，将移动目标的子载波数据与指纹库进行比对，得到移动目标的位置信息，通过深度神经网络训练，提高了接收的子载波相位数据的准确性和可靠性，进而提高了对移动目标位置检测的可靠性和准确性，解决了背景技术中在复杂环境下，移动目标位置检测的可靠性和准确性会变差的问题。The positioning method based on the channel state information in the present invention obtains the optimal deep neural network model by constructing a deep neural network training model and pre-divided sub-carrier phase data of multiple positions, and through deep neural network training. Training can reduce the influence of sub-carrier phase data due to factors such as the environment in a low SNR environment, and then establish a fingerprint database corresponding to the target position according to the sub-carrier phase data calibrated after training, and compare the sub-carrier data of the moving target with the target position. The fingerprint database is compared to obtain the position information of the moving target. Through the training of the deep neural network, the accuracy and reliability of the received sub-carrier phase data are improved, and the reliability and accuracy of the position detection of the moving target are further improved. This solves the problem that the reliability and accuracy of moving target position detection will deteriorate under complex environments in the background art.
另外，根据本发明上述的基于信道状态信息的定位方法，还可以具有如下附加的技术特征：In addition, according to the above-mentioned channel state information-based positioning method of the present invention, it may also have the following additional technical features:
进一步的，深度神经网络模型包括输入层、多个隐藏层和输出层，输入层用于输入多组子载波相位数据建立的输入矩阵，输出层用于输出多组子载波相位数据经多个隐藏层训练后得到的输出矩阵，输出矩阵包括最优深度神经网络模型及对应目标位置的指纹数据。Further, the deep neural network model includes an input layer, a plurality of hidden layers and an output layer, the input layer is used to input the input matrix established by the multiple groups of subcarrier phase data, and the output layer is used to output the multiple groups of subcarrier phase data through multiple hidden layers. The output matrix obtained after layer training, the output matrix includes the optimal deep neural network model and the fingerprint data corresponding to the target position.
进一步的，构建深度神经网络模型，将训练集输入深度神经网络模型中进行训练，得到最优深度神经网络模型及对应目标位置的指纹数据的步骤包括：Further, the steps of constructing a deep neural network model, inputting the training set into the deep neural network model for training, and obtaining the optimal deep neural network model and fingerprint data corresponding to the target position include:
构建模型损失函数，基于模型损失函数的深度神经网络模型对训练集进行训练，当模型损失函数输出值达到预设值的网络模型参数，即是深度神经网络模型最优模型参数，输出深度神经网络模型的最优模型参数及对应目标位置的输出指纹数据。Build the model loss function, and train the training set based on the deep neural network model of the model loss function. When the output value of the model loss function reaches the preset network model parameters, that is, the optimal model parameters of the deep neural network model, the output of the deep neural network The optimal model parameters of the model and the output fingerprint data corresponding to the target position.
进一步的，基于模型损失函数的深度神经网络模型对训练集进行训练的步骤包括：Further, the steps of training the training set based on the deep neural network model of the model loss function include:
将多组子载波相位数据经输入层后输入隐藏层中，根据对比散度算法对子载波相位数据的每个隐藏层的训练结果进行估计，根据每个隐藏层的训练结果估计不同隐藏层之间的网络模型参数。After inputting multiple sets of subcarrier phase data into the hidden layer through the input layer, the training results of each hidden layer of the subcarrier phase data are estimated according to the contrast divergence algorithm, and the difference between different hidden layers is estimated according to the training results of each hidden layer. between the network model parameters.
进一步的，隐藏层包括第一隐藏层、第二隐藏层和第三隐藏层；根据对比散度算法对子载波相位数据的每个隐藏层的训练结果进行估计，根据每个隐藏层的训练结果估计不同隐藏层之间的网络模型参数的步骤包括：Further, the hidden layer includes a first hidden layer, a second hidden layer and a third hidden layer; according to the contrast divergence algorithm, the training result of each hidden layer of the subcarrier phase data is estimated, according to the training result of each hidden layer. The steps of estimating network model parameters between different hidden layers include:
将输入层的输入与第一隐藏层的各个神经元的输出状态进行联合概率分布相乘，以估计第一隐藏层的输出；Multiply the input of the input layer with the output state of each neuron of the first hidden layer by a joint probability distribution to estimate the output of the first hidden layer;
再将第一隐藏层的输出与第二隐藏层的各个神经元的输出状态进行联合概率分布相乘，以估计第二隐藏层的输出；Then, the output of the first hidden layer is multiplied by the joint probability distribution of the output state of each neuron in the second hidden layer to estimate the output of the second hidden layer;
将第二隐藏层的输出与第三隐藏层的各个神经元的输出状态进行联合概率分布相乘，以估计第三隐藏层的输出；Multiply the output of the second hidden layer with the output state of each neuron of the third hidden layer by a joint probability distribution to estimate the output of the third hidden layer;
根据每个隐藏层的输出反向推算出各个隐藏层的输入，进而根据各个隐藏层的输入得到不同隐藏层之间的网络模型参数。According to the output of each hidden layer, the input of each hidden layer is reversely calculated, and then the network model parameters between different hidden layers are obtained according to the input of each hidden layer.
进一步的，根据对比散度算法对子载波相位数据的每个隐藏层的训练结果进行估计的步骤后还包括：Further, the step of estimating the training result of each hidden layer of the subcarrier phase data according to the contrast divergence algorithm further includes:
通过比较各个隐藏层输入与输出之间的误差，并利用误差反向传播算法不断调整不同隐藏层之间的网络模型参数，最终得到最优网络模型参数。By comparing the error between the input and output of each hidden layer, and using the error back propagation algorithm to continuously adjust the network model parameters between different hidden layers, the optimal network model parameters are finally obtained.
进一步的，对多组子载波相位数据进行校准的步骤包括：Further, the step of calibrating the multiple groups of subcarrier phase data includes:
通过对多组子载波相位数据进行线性变换处理，得到校准后与多个位置一一对应的子载波相位数据。By performing linear transformation processing on multiple groups of sub-carrier phase data, the calibrated sub-carrier phase data corresponding to multiple positions one-to-one is obtained.
本发明另一方面提供一种基于信道状态信息的定位系统，系统包括：Another aspect of the present invention provides a positioning system based on channel state information, the system comprising:
训练集构建模块，用于将目标位置预先划分为多个位置，获取与多个位置一一对应的多组子载波相位数据，对多组子载波相位数据进行校准，根据校准后的子载波相位数据构建训练集；The training set building module is used to pre-divide the target position into multiple positions, obtain multiple sets of subcarrier phase data corresponding to the multiple positions one-to-one, and calibrate the multiple sets of subcarrier phase data. data to build a training set;
模型训练模块，用于构建深度神经网络模型，将训练集输入深度神经网络模型中进行训练，得到最优深度神经网络模型及对应目标位置的指纹数据；The model training module is used to construct a deep neural network model, input the training set into the deep neural network model for training, and obtain the optimal deep neural network model and fingerprint data corresponding to the target position;
位置计算模块，用于根据指纹数据建立指纹库，获取移动目标的子载波相位数据，将移动目标的子载波相位数据与指纹库进行匹配，根据匹配结果确定移动目标的位置。The position calculation module is used for establishing a fingerprint database according to the fingerprint data, acquiring the sub-carrier phase data of the moving target, matching the sub-carrier phase data of the moving target with the fingerprint database, and determining the position of the moving target according to the matching result.
本发明另一方面提供一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时实现如上述任一项的基于信道状态信息的定位方法。Another aspect of the present invention provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, implements the channel state information-based positioning method as described above.
本发明另一方面还提供一种数据处理设备，包括存储器、处理器以及存储在存储器上并可在处理器上运行的计算机程序，处理器执行程序时实现如上述任一项的基于信道状态信息的定位方法。Another aspect of the present invention also provides a data processing device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, the processor implements the channel state information based on any of the above when the processor executes the program positioning method.
附图说明Description of drawings
图1为本发明第一实施例中基于信道状态信息的定位方法流程图；FIG. 1 is a flowchart of a positioning method based on channel state information in a first embodiment of the present invention;
图2为本发明第二实施例中基于信道状态信息的定位方法流程图；2 is a flowchart of a positioning method based on channel state information in a second embodiment of the present invention;
图3为本发明第四实施例中基于信道状态信息的定位系统的系统框图；3 is a system block diagram of a positioning system based on channel state information in a fourth embodiment of the present invention;
图4为本发明实施例中通过深度神经网络训练后的建立指纹库的流程示意图；4 is a schematic flow chart of establishing a fingerprint database after training by a deep neural network according to an embodiment of the present invention;
图5为本发明实施例中采用对比散度算法对模型训练时层与层之间关系的示意图；5 is a schematic diagram of the relationship between layers when a model is trained using a contrastive divergence algorithm in an embodiment of the present invention;
图6为本发明实施例中采用误差反向传播算法调整模型的流程示意图；FIG. 6 is a schematic flowchart of adjusting a model by using an error back-propagation algorithm in an embodiment of the present invention;
图7为本发明实施例中25m×20m的目标环境下检测移动目标的流程示意图；7 is a schematic flowchart of detecting a moving target in a target environment of 25m×20m according to an embodiment of the present invention;
图8为本发明实施例中基于信道状态信息的定位方法的应用场景示意图；8 is a schematic diagram of an application scenario of a positioning method based on channel state information in an embodiment of the present invention;
如下具体实施方式将结合上述附图进一步说明本发明。The following specific embodiments will further illustrate the present invention in conjunction with the above drawings.
具体实施方式Detailed ways
为了便于理解本发明，下面将参照相关附图对本发明进行更全面的描述。附图中给出了本发明的若干实施例。但是，本发明可以以许多不同的形式来实现，并不限于本文所描述的实施例。相反地，提供这些实施例的目的是使对本发明的公开内容更加透彻全面。In order to facilitate understanding of the present invention, the present invention will be described more fully hereinafter with reference to the related drawings. Several embodiments of the invention are presented in the accompanying drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
除非另有定义，本文所使用的所有的技术和科学术语与属于本发明的技术领域的技术人员通常理解的含义相同。本文中在本发明的说明书中所使用的术语只是为了描述具体的实施例的目的，不是旨在于限制本发明。本文所使用的术语“及／或”包括一个或多个相关的所列项目的任意的和所有的组合。Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terms used herein in the description of the present invention are for the purpose of describing specific embodiments only, and are not intended to limit the present invention. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.
本发明主要应用于基于无线信号进行环境检测的场合，如图8所示为本发明的应用场景示意图，包括目标环境、网络设备和定位装置。在目标环境中预先划分出多个不同位置，使网络设备发送的网络信号覆盖整个目标环境，网络设备将接收目标环境中不同位置的子载波信号，并将子载波信号发送至定位装置，定位装置将根据子载波信号进行匹配定位。The present invention is mainly applied to the occasion of environment detection based on wireless signals. FIG. 8 is a schematic diagram of the application scenario of the present invention, including target environment, network equipment and positioning device. A number of different locations are pre-divided in the target environment, so that the network signal sent by the network device covers the entire target environment. The network device will receive sub-carrier signals at different locations in the target environment, and send the sub-carrier signals to the positioning device. The positioning device Matching and positioning will be performed according to the subcarrier signal.
本发明首先将智能化环境检测问题，分解为训练和在线实时计算两个阶段，在离线训练阶段中利用带有三层隐藏层的深度神经网络对校准后的子载波相位数据本发明首先将智能化环境检测问题，分解为训练和在线实时计算两个阶段，在离线训练阶段中利用带有三层隐藏层的深度神经网络对校准后的子载波相位数据
实施例一Example 1
请参阅图1，所示为本发明第一实施例中的基于信道状态信息的定位方法，包括步骤S11-S13。Please refer to FIG. 1, which shows a channel state information-based positioning method in the first embodiment of the present invention, including steps S11-S13.
S11、将目标位置预先划分为多个位置，获取与多个位置一一对应的多组子载波相位数据，对多组子载波相位数据进行校准，根据校准后的子载波相位数据构建训练集。S11. Divide the target position into multiple positions in advance, acquire multiple sets of subcarrier phase data corresponding to the multiple positions one-to-one, calibrate the multiple sets of subcarrier phase data, and construct a training set according to the calibrated subcarrier phase data.
构建训练集：训练集即收集得到并通过线性变换校准后的n组i个子载波相位数据，根据校准后的n组i个子载波相位数据建立输入矩阵h 0 ，输入h 0 是n×i维的矩阵，n表示建立指纹信息的设定位置总数，i表示在某一位置上所输入的i个校准后子载波相位数据，输出结果同样是n×i维的指纹输出矩阵，表示在n个不同位置下输出的的i个经过模型训练后获得的指纹信息数据，输入矩阵h 0 的每一行与指纹输出矩阵的每一行对应。Construction of training set: The training set is the n groups of i sub-carrier phase data that are collected and calibrated by linear transformation, and the input matrix h 0 is established according to the calibrated n groups of i sub-carrier phase data. The input h 0 is n×i -dimensional Matrix, n represents the total number of set positions for establishing fingerprint information, i represents the i calibrated sub-carrier phase data input at a certain position, and the output result is also an n×i -dimensional fingerprint output matrix, which means that in n different The i pieces of fingerprint information data obtained after model training are output under the position, and each row of the input matrix h 0 corresponds to each row of the fingerprint output matrix.
S12、构建深度神经网络模型，将训练集输入深度神经网络模型中进行训练，得到最优深度神经网络模型及对应目标位置的指纹数据。S12 , constructing a deep neural network model, inputting the training set into the deep neural network model for training, and obtaining an optimal deep neural network model and fingerprint data corresponding to the target position.
构建深度神经网络：该深度神经网络含有一层输入层，三层隐藏层，以及一层输出层。h 0 表示输入矩阵，即n组i个校准后子载波相位数据，W 1 、W 2 、W 3 表示输入层与第一隐藏层，第一隐藏层与第二隐藏层，第二隐藏层与第三隐藏层之间的网络模型参数。h 0 通过输入层送入深度神经网络模型中，经过层与层之间的训练，最后由输出层输出结果。Building a deep neural network: The deep neural network consists of one input layer, three hidden layers, and one output layer. h 0 represents the input matrix, that is, n groups of i calibrated sub-carrier phase data, W 1 , W 2 , W 3 represent the input layer and the first hidden layer, the first hidden layer and the second hidden layer, and the second hidden layer and Network model parameters between the third hidden layer. h 0 is sent into the deep neural network model through the input layer, after training between layers, and finally the result is output by the output layer.
其中，收集的子载波数据输入深度神经网络中进行训练时，n组i个子载波相位数据作为输入矩阵输入神经网络中的输入层中，进而在三个隐藏层中依次进行训练，在输出层中得到由训练后的子载波相位数据建立的输出矩阵。Among them, when the collected sub-carrier data is input into the deep neural network for training, n groups of i sub-carrier phase data are input into the input layer of the neural network as an input matrix, and then trained in the three hidden layers in turn, and in the output layer An output matrix built from the trained subcarrier phase data is obtained.
训练过程时，为了更好的获取到最优的网络模型参数，使模型训练至更优，构建模型损失函数，模型损失函数如下：During the training process, in order to better obtain the optimal network model parameters and make the model training more optimal, the model loss function is constructed. The model loss function is as follows:
其中，h k-1 、h k 分别表示在第K-1层和第K层的输出，h 0 表示输入数据，h 1 、h 2 、h 3 分别表示输入数据h 0 在经过模型隐藏层第一层K 1 ，第二层K 2 ，第三层K 3 的输出结果，b k-1 、b k 分别表示在模型第K-1层和第K层的偏差，b 0 、b 1 、b 2 、b 3 分别表示在模型输入层，隐藏层第一层K 1 ，第二层K 2 ，第三层K 3 的偏差。使得上式损失函数输出值达到预设值的网络模型参数W 1 、W 2 、W 3 ，即为模型训练所得的最优模型参数。Among them, h k-1 and h k represent the output at the K-1th layer and the Kth layer, respectively, h 0 represents the input data, h 1 , h 2 , h 3 represent the input data h 0 after passing through the hidden layer of the model. The output results of the first layer K 1 , the second layer K 2 , and the third layer K 3 , b k-1 and b k represent the deviations at the K-1th layer and the Kth layer of the model, respectively, b 0 , b 1 , b 2 and b 3 respectively represent the deviation of the input layer of the model, the first layer K 1 of the hidden layer, the second layer K 2 , and the third layer K 3 . The network model parameters W 1 , W 2 , and W 3 that make the output value of the loss function of the above formula reach the preset value are the optimal model parameters obtained by the model training.
输入矩阵h 0 在经过具备最优模型参数的深度神经网络训练后，会输出n×i维的输出矩阵，即对应n个不同位置下的指纹数据
S13、根据指纹数据建立指纹库，获取移动目标的子载波相位数据，将移动目标的子载波相位数据与指纹库进行匹配，根据匹配结果确定移动目标的位置。S13 , establishing a fingerprint database according to the fingerprint data, acquiring sub-carrier phase data of the moving target, matching the sub-carrier phase data of the moving target with the fingerprint database, and determining the position of the moving target according to the matching result.
如图4所示，根据经模型训练后得到的指纹数据
在指纹匹配过程中，因为是在同一收发端和共同信道环境下，所以CSI所携带的子载波相位信息中的个数X和其子载波间频率间隔
故在实际的指纹匹配操作中，只需比较校准后的子载波相位数据
可知，在某一位置下接收移动目标的CSI所携带的子载波信息经过校准后，有校准后的子载波相位数据
与已建立指纹库中各位置上的指纹数据：
两者进行比较匹配：Compare the two to match:
观察上式，根据所得结果可以发现，当
获取移动目标的子载波相位数据，并进行线性变换校准，将接收并处理得到的移动目标校准后的子载波相位数据
上式中：P表示进行P次指纹匹配，N表示白噪声对CSI接收和指纹匹配的干扰。In the above formula: P represents the fingerprint matching for P times, and N represents the interference of white noise on CSI reception and fingerprint matching.
在输出的指纹匹配结果中，我们可以得到移动目标CSI校准后的子载波相位数据
综上，本发明上述实施例当中的基于信道状态信息的定位方法，通过构建深度神经网络训练模型和预先划分的多个位置的子载波相位数据，通过训练深度神经网络，得到最优深度神经网络模型，通过深度神经网络训练，能够减少子载波相位数据在低信噪比环境中因环境等因素所受到的的影响，再根据训练后校准的子载波相位数据建立对应目标位置的指纹库，将移动目标的子载波数据与指纹库进行比对，得到移动目标的位置信息，通过深度神经网络训练，提高了接收的子载波相位数据的准确性和可靠性，进而提高了对移动目标位置检测的可靠性和准确性，解决了背景技术中在复杂环境下，移动目标位置检测的可靠性和准确性会变差的问题。To sum up, in the channel state information-based positioning method in the above-mentioned embodiments of the present invention, the optimal deep neural network is obtained by constructing a deep neural network training model and pre-divided sub-carrier phase data of multiple locations, and by training the deep neural network. The model, through deep neural network training, can reduce the influence of the sub-carrier phase data in the low signal-to-noise ratio environment due to factors such as the environment, and then establish a fingerprint database corresponding to the target position according to the sub-carrier phase data calibrated after training, and the The sub-carrier data of the moving target is compared with the fingerprint database to obtain the position information of the moving target. Through deep neural network training, the accuracy and reliability of the received sub-carrier phase data are improved, thereby improving the detection accuracy of the moving target position. Reliability and accuracy solve the problem that the reliability and accuracy of moving target position detection will deteriorate under complex environments in the background art.
实施例二Embodiment 2
请参阅图2，所示为本发明第二实施例中的基于信道状态信息的定位方法，包括步骤S21-S24。Please refer to FIG. 2, which shows a positioning method based on channel state information in a second embodiment of the present invention, including steps S21-S24.
S21、将目标位置预先划分为多个位置，获取与多个位置一一对应的多组子载波相位数据，对多组子载波相位数据进行校准，根据校准后的子载波相位数据构建训练集。S21. Divide the target position into multiple positions in advance, acquire multiple sets of subcarrier phase data corresponding to the multiple positions one-to-one, calibrate the multiple sets of subcarrier phase data, and construct a training set according to the calibrated subcarrier phase data.
构建训练集：训练集即收集得到并通过线性变换校准后的n组i个子载波相位数据，根据校准后的n组i个子载波相位数据建立输入矩阵h 0 ，输入h 0 是n×i维的矩阵，n表示建立指纹信息的设定位置总数，i表示在某一位置上所输入的i个校准后子载波相位数据，输出结果同样是n×i维的指纹输出矩阵，表示在n个不同位置下输出的的i个经过模型训练后获得的指纹信息数据，输入矩阵h 0 的每一行与指纹输出矩阵的每一行对应。Construction of training set: The training set is the n groups of i sub-carrier phase data that are collected and calibrated by linear transformation, and the input matrix h 0 is established according to the calibrated n groups of i sub-carrier phase data. The input h 0 is n×i -dimensional Matrix, n represents the total number of set positions for establishing fingerprint information, i represents the i calibrated sub-carrier phase data input at a certain position, and the output result is also an n×i -dimensional fingerprint output matrix, which means that in n different The i pieces of fingerprint information data obtained after model training are output under the position, and each row of the input matrix h 0 corresponds to each row of the fingerprint output matrix.
子载波相位数据校准包括对在环境中不同n个位置下采集到的
根据校准后的n组i个子载波相位数据建立输入矩阵h 0 ，输入h 0 是n×i维的矩阵，n表示建立指纹信息的设定位置总数，i表示在某一位置上所输入的i个校准后子载波相位数据，输出结果同样是n×i维的指纹输出矩阵，表示在n个不同位置下输出的的i个经过模型训练后获得的指纹信息数据，输入矩阵h 0 的每一行与指纹输出矩阵的每一行对应。An input matrix h 0 is established according to the calibrated n groups of i sub-carrier phase data, the input h 0 is an n×i -dimensional matrix, n represents the total number of set positions for establishing fingerprint information, and i represents the i input at a certain position The output result is also an n×i -dimensional fingerprint output matrix, which represents the i fingerprint information data obtained after model training output at n different positions, and each row of the input matrix h 0 Corresponds to each row of the fingerprint output matrix.
S22、构建深度神经网络模型，将训练集输入深度神经网络模型中，采取对比散度算法对模型进行训练。S22 , constructing a deep neural network model, inputting the training set into the deep neural network model, and using a contrastive divergence algorithm to train the model.
构建深度神经网络：该深度神经网络含有一层输入层，三层隐藏层，以及一层输出层。h 0 表示输入矩阵，即n组i个校准后子载波相位数据，W 1 、W 2 、W 3 表示输入层与第一隐藏层，第一隐藏层与第二隐藏层，第二隐藏层与第三隐藏层之间的网络模型参数。h 0 通过输入层送入深度神经网络模型中，经过层与层之间的训练，最后由输出层输出结果。Building a deep neural network: The deep neural network consists of one input layer, three hidden layers, and one output layer. h 0 represents the input matrix, that is, n groups of i calibrated sub-carrier phase data, W 1 , W 2 , W 3 represent the input layer and the first hidden layer, the first hidden layer and the second hidden layer, and the second hidden layer and Network model parameters between the third hidden layer. h 0 is sent into the deep neural network model through the input layer, after training between layers, and finally the result is output by the output layer.
其中，收集的子载波数据输入深度神经网络中进行训练时，n组i个子载波相位数据作为输入矩阵输入神经网络中的输入层中，进而在三个隐藏层中依次进行训练，在输出层中得到由训练后的子载波相位数据建立的输出矩阵。Among them, when the collected sub-carrier data is input into the deep neural network for training, n groups of i sub-carrier phase data are input into the input layer of the neural network as an input matrix, and then trained in the three hidden layers in turn, and in the output layer An output matrix built from the trained subcarrier phase data is obtained.
其中，h k-1 、h k 分别表示在第K-1层和第K层的输出，h 0 表示输入数据，h 1 、h 2 、h 3 分别表示输入数据h 0 在经过模型隐藏层第一层K 1 ，第二层K 2 ，第三层K 3 的输出结果，b k-1 、b k 分别表示在模型第K-1层和第K层的偏差，b 0 、b 1 、b 2 、b 3 分别表示在模型输入层，隐藏层第一层K 1 ，第二层K 2 ，第三层K 3 的偏差。使得上式损失函数输出值达到预设值的网络模型参数W 1 、W 2 、W 3 ，即为模型训练所得的最优模型参数。Among them, h k-1 and h k represent the output at the K-1th layer and the Kth layer, respectively, h 0 represents the input data, h 1 , h 2 , h 3 represent the input data h 0 after passing through the hidden layer of the model. The output results of the first layer K 1 , the second layer K 2 , and the third layer K 3 , b k-1 and b k represent the deviations at the K-1th layer and the Kth layer of the model, respectively, b 0 , b 1 , b 2 and b 3 respectively represent the deviation of the input layer of the model, the first layer K 1 of the hidden layer, the second layer K 2 , and the third layer K 3 . The network model parameters W 1 , W 2 , and W 3 that make the output value of the loss function of the above formula reach the preset value are the optimal model parameters obtained by the model training.
现有技术中，通常采取随机梯度下降法来更新网络模型参数对模型进行训练，但随机梯度下降法会造成模型训练复杂度过高，模型训练过慢，训练效率过低等问题，因此，本实施中通过对比散度算法在模型第K-1层和第K层之间进行训练，如图5所示，当第K-1层的输入经过网络训练传递至第K层时，各个神经元之间状态相互独立，训练结果只是在层与层之间的神经元传递，则可利用K-1层的输入h k-1 经过模型网络模型参数训练后与第K层各神经元
其中，
S23、根据误差反向传播算法调整经训练后估计的网络模型参数，得到最优深度神经网络模型及对应目标位置的指纹数据。S23: Adjust the parameters of the network model estimated after training according to the error back-propagation algorithm to obtain an optimal deep neural network model and fingerprint data corresponding to the target position.
如图6所示，输入h 0 在经过模型输入层和第一层K 1 时，根据由拟合得到的初始网络模型参数b 0 、b 1 、W 1 的值来计算输入层K 0 和第一层K 1 中的
根据每一次的训练得到的输出
其中，∂是在模型训练前预先设置好的学习率。Among them, ∂ is the learning rate preset before model training.
输入矩阵h 0 在经过具备最优模型参数的深度神经网络训练后，会输出n×i维的输出矩阵，即对应n个不同位置下的指纹数据
S24、根据指纹数据建立指纹库，获取移动目标的子载波相位数据，将移动目标的子载波相位数据与指纹库进行匹配，根据匹配结果确定移动目标的位置。S24. Establish a fingerprint database according to the fingerprint data, obtain subcarrier phase data of the moving target, match the subcarrier phase data of the moving target with the fingerprint database, and determine the position of the moving target according to the matching result.
如图4所示，根据经模型训练后得到的指纹数据
在指纹匹配过程中，因为是在同一收发端和共同信道环境下，所以CSI所携带的子载波相位信息中的个数X和其子载波间频率间隔
故在实际的指纹匹配操作中，只需比较校准后的子载波相位数据
可知，在某一位置下接收移动目标的CSI所携带的子载波信息经过校准后，有校准后的子载波相位数据
与已建立指纹库中各位置上的指纹数据
两者进行比较匹配：Compare the two to match:
观察上式，根据所得结果可以发现，当
获取移动目标的子载波相位数据，并进行线性变换校准，将接收并处理得到的移动目标校准后的子载波相位数据
上式中：P表示进行P次指纹匹配，N表示白噪声对CSI接收和指纹匹配的干扰。In the above formula: P represents the fingerprint matching for P times, and N represents the interference of white noise on CSI reception and fingerprint matching.
在输出的指纹匹配结果中，我们可以得到移动目标CSI校准后的子载波相位数据
综上，本发明上述实施例当中的基于信道状态信息的定位方法，通过构建深度神经网络训练模型和预先划分的多个位置的子载波相位数据，通过训练深度神经网络，得到最优深度神经网络模型，通过深度神经网络训练，能够减少子载波相位数据在低信噪比环境中因环境等因素所受到的的影响，再根据训练后校准的子载波相位数据建立对应目标位置的指纹库，将移动目标的子载波相位数据与指纹库进行比对，得到移动目标的位置信息，通过深度神经网络训练，提高了接收的子载波相位数据的准确性和可靠性，进而提高了对移动目标位置检测的可靠性和准确性，解决了背景技术中在复杂环境下，移动目标位置检测的可靠性和准确性会变差的问题。To sum up, in the channel state information-based positioning method in the above-mentioned embodiments of the present invention, the optimal deep neural network is obtained by constructing a deep neural network training model and pre-divided sub-carrier phase data of multiple locations, and by training the deep neural network. The model, through deep neural network training, can reduce the influence of the sub-carrier phase data in the low signal-to-noise ratio environment due to factors such as the environment, and then establish a fingerprint database corresponding to the target position according to the sub-carrier phase data calibrated after training, and the The sub-carrier phase data of the moving target is compared with the fingerprint database to obtain the position information of the moving target. Through deep neural network training, the accuracy and reliability of the received sub-carrier phase data are improved, and the detection of the moving target position is improved. It solves the problem that the reliability and accuracy of moving target position detection will deteriorate under complex environments in the background technology.
实施例三Embodiment 3
如图7所示，本实施例中还提供一种根据上述步骤S21-S24的方法对25m×20m的目标环境中移动目标检测方法。As shown in FIG. 7 , this embodiment also provides a method for detecting a moving target in a target environment of 25m×20m according to the method in the above steps S21-S24.
S31：在25m×20m的目标环境下，确定n=2000个不同训练位置0.5m×0.5m且利用设备采集每个位置上的i=90个子载波相位数据
S32：可知每个位置可以用一组i=90个校准后子载波相位数据
S33：采用在输入层k 0带有90个输入节点，在第一隐藏层K 1 带有60个节点，在第二隐藏层K 2 带有30个节点，在第三隐藏层K 3 带有15个节点，在输出层带有6个输出节点的深度神经网络，将输入数据的每一行的数据对应输入到每一个输入节点，输入数据在经过模型不同层时会对应输出不同数据，每进行一次完整的训练，会根据输入h 0 和输出
S34：当移动目标在已建立好的指纹库中进行移动时，接收设备可以获知其CSI所携带的子载波信息
为了消除环境白噪声N对CSI接收和匹配结果的影响，在同一段时间内，由接收设备得到的一组移动目标CSI所携带的子载波信息
最终输出的2000个不同位置下20次累加后的指纹匹配结果The final output fingerprint matching results after 20 accumulations in 2000 different positions
S35：每隔△T=1s，接收设备可以获知移动目标的一组CSI所携带的子载波信息
在一段时间T(T＞1)内，经过多组匹配结果比较后，若指纹匹配结果并无明显变化，则说明在该时间内，移动目标静止，即在目标环境中并未移动；若相邻两组匹配之间，指纹匹配结果发生变化，则对应指纹库中位置变化可确定移动目标在该时间内的运动方向和轨迹，进而计算其运动速度，实现在该环境中移动目标的运动检测。Within a period of time T (T>1) , after comparing multiple sets of matching results, if there is no significant change in the fingerprint matching results, it means that the moving target is stationary during this period of time, that is, it does not move in the target environment; If the fingerprint matching result changes between the adjacent two groups of matches, the position change in the corresponding fingerprint database can determine the moving direction and trajectory of the moving target within this time, and then calculate its moving speed to realize the motion detection of the moving target in this environment.
综上，本发明上述实施例当中的基于信道状态信息的定位方法，通过构建深度神经网络训练模型和预先划分的多个位置的子载波相位数据，通过训练深度神经网络，得到最优深度神经网络模型，通过深度神经网络训练，能够减少子载波相位数据在低信噪比环境中因环境等因素所受到的的影响，再根据训练后校准的子载波相位数据建立对应目标位置的指纹库，将移动目标的子载波数据与指纹库进行比对，得到移动目标的位置信息，通过深度神经网络训练，提高了接收的子载波相位数据的准确性和可靠性，进而提高了对移动目标位置检测的可靠性和准确性，解决了背景技术中在复杂环境下，移动目标位置检测的可靠性和准确性会变差的问题。To sum up, in the channel state information-based positioning method in the above-mentioned embodiments of the present invention, the optimal deep neural network is obtained by constructing a deep neural network training model and pre-divided sub-carrier phase data of multiple locations, and by training the deep neural network. The model, through deep neural network training, can reduce the influence of the sub-carrier phase data in the low signal-to-noise ratio environment due to factors such as the environment, and then establish a fingerprint database corresponding to the target position according to the sub-carrier phase data calibrated after training, and the The sub-carrier data of the moving target is compared with the fingerprint database to obtain the position information of the moving target. Through deep neural network training, the accuracy and reliability of the received sub-carrier phase data are improved, thereby improving the detection accuracy of the moving target position. Reliability and accuracy solve the problem that the reliability and accuracy of moving target position detection will deteriorate under complex environments in the background art.
实施例四Embodiment 4
本发明另一方面还提供一种基于信道状态信息的定位系统，请参阅图3，所示本发明申请第三实施例中的基于信道状态信息的定位系统，所述系统包括：Another aspect of the present invention also provides a positioning system based on channel state information. Please refer to FIG. 3 , which shows the channel state information-based positioning system in the third embodiment of the present application. The system includes:
训练集构建模块，用于将目标位置预先划分为多个位置，获取与多个位置一一对应的多组子载波相位数据，对所述多组子载波相位数据进行校准，根据校准后的子载波相位数据构建训练集；The training set building module is used to pre-divide the target position into multiple positions, obtain multiple groups of subcarrier phase data corresponding to the multiple positions one-to-one, and calibrate the multiple groups of subcarrier phase data. The carrier phase data constructs a training set;
模型训练模块，用于构建深度神经网络模型，将所述训练集输入所述深度神经网络模型中进行训练，得到最优深度神经网络模型及对应目标位置的指纹数据；A model training module, used for constructing a deep neural network model, and inputting the training set into the deep neural network model for training to obtain an optimal deep neural network model and fingerprint data corresponding to the target position;
位置计算模块，用于根据所述指纹数据建立指纹库，获取移动目标的子载波相位数据，将所述移动目标的子载波相位数据与所述指纹库进行匹配，根据匹配结果确定所述移动目标的位置。The position calculation module is used to establish a fingerprint database according to the fingerprint data, obtain the subcarrier phase data of the moving target, match the subcarrier phase data of the moving target with the fingerprint database, and determine the moving target according to the matching result s position.
进一步的，在一些其他可选实施例中，所述模型构建模块包括：Further, in some other optional embodiments, the model building module includes:
深度神经网络单元，包括输入层、多个隐藏层和输出层，所述输入层用于输入所述多组子载波相位数据建立的输入矩阵，所述输出层用于输出所述多组子载波相位数据经所述多个隐藏层训练后得到的输出矩阵，所述输出矩阵包括最优深度神经网络模型及对应目标位置的指纹数据。A deep neural network unit, including an input layer, a plurality of hidden layers and an output layer, the input layer is used for inputting the input matrix established by the multiple groups of subcarrier phase data, and the output layer is used for outputting the multiple groups of subcarriers The output matrix obtained after the phase data is trained by the multiple hidden layers, the output matrix includes the optimal deep neural network model and the fingerprint data corresponding to the target position.
进一步的，在一些其他可选实施例中，所述模型训练模块包括：Further, in some other optional embodiments, the model training module includes:
损失函数构建单元，用于基于所述模型损失函数的深度神经网络模型对所述训练集进行训练，当所述模型损失函数输出值达到预设值的网络模型参数，即是所述深度神经网络模型最优模型参数，输出所述深度神经网络模型的最优模型参数及对应目标位置的输出指纹数据。The loss function construction unit is used to train the training set based on the deep neural network model of the model loss function. When the output value of the model loss function reaches the preset network model parameters, the deep neural network is the network model parameter. Optimal model parameters of the model, and outputting the optimal model parameters of the deep neural network model and the output fingerprint data corresponding to the target position.
进一步的，在一些其他可选实施例中，所述损失函数构建单元包括：Further, in some other optional embodiments, the loss function construction unit includes:
对比散度算法子单元，用于将所述多组子载波相位数据经所述输入层后输入所述隐藏层中，根据对比散度算法对所述子载波相位数据的每个隐藏层的训练结果进行估计，根据每个隐藏层的训练结果估计不同隐藏层之间的网络模型参数。A contrastive divergence algorithm subunit, configured to input the multiple groups of subcarrier phase data into the hidden layer after passing through the input layer, and train each hidden layer of the subcarrier phase data according to the contrastive divergence algorithm The results are estimated, and the network model parameters between different hidden layers are estimated according to the training results of each hidden layer.
进一步的，在一些其他可选实施例中，所述隐藏层包括第一隐藏层、第二隐藏层和第三隐藏层；所述对比散度算法子单元包括：Further, in some other optional embodiments, the hidden layer includes a first hidden layer, a second hidden layer and a third hidden layer; the contrastive divergence algorithm subunit includes:
联合概率分布相乘子单元，用于将所述输入层的输入与所述第一隐藏层的各个神经元的输出状态进行联合概率分布相乘，以估计所述第一隐藏层的输出；a joint probability distribution multiplier subunit, configured to perform a joint probability distribution multiplication between the input of the input layer and the output state of each neuron of the first hidden layer to estimate the output of the first hidden layer;
再将所述第一隐藏层的输出与所述第二隐藏层的各个神经元的输出状态进行联合概率分布相乘，以估计所述第二隐藏层的输出；Then, the output of the first hidden layer is multiplied by a joint probability distribution with the output state of each neuron of the second hidden layer to estimate the output of the second hidden layer;
将所述第二隐藏层的输出与所述第三隐藏层的各个神经元的输出状态进行联合概率分布相乘，以估计所述第三隐藏层的输出；Multiplying the output of the second hidden layer with the output state of each neuron of the third hidden layer by a joint probability distribution to estimate the output of the third hidden layer;
根据每个隐藏层的输出反向推算出各个隐藏层的输入，进而根据各个隐藏层的输入得到不同隐藏层之间的网络模型参数。According to the output of each hidden layer, the input of each hidden layer is reversely calculated, and then the network model parameters between different hidden layers are obtained according to the input of each hidden layer.
进一步的，在一些其他可选实施例中，所述损失函数构建单元还包括：Further, in some other optional embodiments, the loss function construction unit further includes:
误差反向传播算法子单元，用于通过比较各个隐藏层输入与输出之间的误差，并利用误差反向传播算法不断调整不同隐藏层之间的网络模型参数，最终得到最优网络模型参数。The error back propagation algorithm sub-unit is used to compare the errors between the input and output of each hidden layer, and use the error back propagation algorithm to continuously adjust the network model parameters between different hidden layers, and finally obtain the optimal network model parameters.
进一步的，在一些其他可选实施例中，所述训练集构建模块包括：Further, in some other optional embodiments, the training set building module includes:
线性变换单元，用于通过对所述多组子载波相位数据进行线性变换处理，得到校准后与所述多个位置一一对应的子载波相位数据。The linear transformation unit is configured to perform linear transformation processing on the multiple groups of sub-carrier phase data to obtain the calibrated sub-carrier phase data corresponding to the multiple positions one-to-one.
上述各模块、单元被执行时所实现的功能或操作步骤与上述方法实施例大体相同，在此不再赘述。The functions or operation steps implemented by the foregoing modules and units when executed are substantially the same as those in the foregoing method embodiments, and will not be repeated here.
综上，本发明上述实施例当中的基于信道状态信息的定位系统，通过构建深度神经网络训练模型和预先划分的多个位置的子载波相位数据，通过训练深度神经网络，得到最优深度神经网络模型，通过深度神经网络训练，能够减少子载波相位数据在低信噪比环境中因环境等因素所受到的的影响，再根据训练后校准的子载波相位数据建立对应目标位置的指纹库，将移动目标的子载波数据与指纹库进行比对，得到移动目标的位置信息，通过深度神经网络训练，提高了接收的子载波相位数据的准确性和可靠性，进而提高了对移动目标位置检测的可靠性和准确性，解决了背景技术中在复杂环境下，移动目标位置检测的可靠性和准确性会变差的问题。To sum up, the positioning system based on the channel state information in the above-mentioned embodiments of the present invention obtains the optimal deep neural network by constructing a deep neural network training model and pre-divided sub-carrier phase data of multiple positions, and by training the deep neural network. The model, through deep neural network training, can reduce the influence of the sub-carrier phase data in the low signal-to-noise ratio environment due to factors such as the environment, and then establish a fingerprint database corresponding to the target position according to the sub-carrier phase data calibrated after training, and the The sub-carrier data of the moving target is compared with the fingerprint database to obtain the position information of the moving target. Through deep neural network training, the accuracy and reliability of the received sub-carrier phase data are improved, thereby improving the detection accuracy of the moving target position. Reliability and accuracy solve the problem that the reliability and accuracy of moving target position detection will deteriorate under complex environments in the background art.
本发明实施例还提出一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时实现如上述实施例中的基于信道状态信息的定位方法的步骤。An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, implements the steps of the channel state information-based positioning method in the foregoing embodiment.
实施例五Embodiment 5
本发明另一方面还提出一种设备，所述系统包括存储器、处理器以及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述程序时实现上述实施例中基于信道状态信息的定位方法。其中，处理器在一些实施例中可以是电子控制单元(ElectronicControlUnit，简称ECU，又称行车电脑)、中央处理器(CentralProcessingUnit，CPU)、控制器、微控制器、微处理器或其他数据处理芯片，用于运行存储器中存储的程序代码或处理数据，例如执行访问限制程序等。Another aspect of the present invention also provides a device. The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the processor implements the based on A method for locating channel state information. Wherein, the processor may be an electronic control unit (Electronic Control Unit, ECU for short, also known as a trip computer), a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor or other data processing chips in some embodiments , used to run program code or process data stored in memory, such as executing access-restricted programs, etc.
其中，存储器至少包括一种类型的可读存储介质，所述可读存储介质包括闪存、硬盘、多媒体卡、卡型存储器(例如，SD或DX存储器等)、磁性存储器、磁盘、光盘等。存储器在一些实施例中可以是设备的内部存储单元，例如该设备的硬盘。存储器在另一些实施例中也可以是设备的外部存储装置，例如设备上配备的插接式硬盘，智能存储卡(SmartMediaCard，SMC)，安全数字(SecureDigital，SD)卡，闪存卡(FlashCard)等。进一步地，存储器还可以既包括设备的内部存储单元也包括外部存储装置。存储器不仅可以用于存储安装于设备的应用软件及各类数据，还可以用于暂时地存储已经输出或者将要输出的数据。Wherein, the memory includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (eg, SD or DX memory, etc.), magnetic memory, magnetic disk, optical disk, and the like. The memory may in some embodiments be an internal storage unit of the device, such as the device's hard disk. In other embodiments, the memory may also be an external storage device of the device, such as a plug-in hard disk equipped on the device, a smart memory card (SmartMediaCard, SMC), a secure digital (SD) card, a flash memory card (FlashCard), etc. . Further, the memory may also include both an internal storage unit of the device and an external storage device. The memory can be used not only to store application software installed in the device and various types of data, but also to temporarily store data that has been output or will be output.
综上，本发明上述实施例当中的设备，通过构建深度神经网络训练模型和预先划分的多个位置的子载波相位数据，通过训练深度神经网络，得到最优深度神经网络模型，通过深度神经网络训练，能够减少子载波相位数据在低信噪比环境中因环境等因素所受到的的影响，再根据训练后校准的子载波相位数据建立对应目标位置的指纹库，将移动目标的子载波数据与指纹库进行比对，得到移动目标的位置信息，通过深度神经网络训练，提高了接收的子载波相位数据的准确性和可靠性，进而提高了对移动目标位置检测的可靠性和准确性，解决了背景技术中在复杂环境下，移动目标位置检测的可靠性和准确性会变差的问题。To sum up, the device in the above-mentioned embodiments of the present invention obtains the optimal deep neural network model by constructing a deep neural network training model and pre-divided sub-carrier phase data of multiple positions, and training the deep neural network. Training can reduce the influence of sub-carrier phase data in low signal-to-noise ratio environment due to environmental and other factors, and then establish a fingerprint database corresponding to the target position according to the sub-carrier phase data calibrated after training, and transfer the sub-carrier data of the moving target. Compared with the fingerprint database, the position information of the moving target is obtained. Through the training of the deep neural network, the accuracy and reliability of the received subcarrier phase data are improved, and the reliability and accuracy of the position detection of the moving target are further improved. It solves the problem that the reliability and accuracy of moving target position detection will deteriorate under complex environments in the background art.
在流程图中表示或在此以其他方式描述的逻辑和/或步骤，例如，可以被认为是用于实现逻辑功能的可执行指令的定序列表，可以具体实现在任何计算机可读介质中，以供指令执行系统、装置或设备（如基于计算机的系统、包括处理器的系统或其他可以从指令执行系统、装置或设备取指令并执行指令的系统）使用，或结合这些指令执行系统、装置或设备而使用。就本说明书而言，“计算机可读介质”可以是任何可以包含、存储、通信、传播或传输程序以供指令执行系统、装置或设备或结合这些指令执行系统、装置或设备而使用的装置。The logic and/or steps represented in flowcharts or otherwise described herein, for example, may be considered an ordered listing of executable instructions for implementing the logical functions, may be embodied in any computer-readable medium, For use with, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other system that can fetch instructions from and execute instructions from an instruction execution system, apparatus, or apparatus) or equipment. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport the program for use by or in connection with an instruction execution system, apparatus, or apparatus.
计算机可读介质的更具体的示例（非穷尽性列表）包括以下：具有一个或多个布线的电连接部（电子装置），便携式计算机盘盒（磁装置），随机存取存储器（RAM），只读存储器（ROM），可擦除可编辑只读存储器（EPROM或闪速存储器），光纤装置，以及便携式光盘只读存储器（CDROM）。另外，计算机可读介质甚至可以是可在其上打印程序的纸或其他合适的介质，因为可以例如通过对纸或其他介质进行光学扫描，接着进行编辑、解译或必要时以其他合适方式进行处理来以电子方式获得程序，然后将其存储在计算机存储器中。More specific examples (non-exhaustive list) of computer readable media include the following: electrical connections with one or more wiring (electronic devices), portable computer disk cartridges (magnetic devices), random access memory (RAM), Read Only Memory (ROM), Erasable Editable Read Only Memory (EPROM or Flash Memory), Fiber Optic Devices, and Portable Compact Disc Read Only Memory (CDROM). In addition, the computer readable medium may even be paper or other suitable medium on which the program may be printed, as may be done, for example, by optically scanning the paper or other medium, followed by editing, interpretation, or other suitable means as necessary process to obtain the program electronically and then store it in computer memory.
应当理解，本发明的各部分可以用硬件、软件、固件或它们的组合来实现。在上述实施方式中，多个步骤或方法可以用存储在存储器中且由合适的指令执行系统执行的软件或固件来实现。例如，如果用硬件来实现，和在另一实施方式中一样，可用本领域公知的下列技术中的任一项或他们的组合来实现：具有用于对数据信号实现逻辑功能的逻辑门电路的离散逻辑电路，具有合适的组合逻辑门电路的专用集成电路，可编程门阵列（PGA），现场可编程门阵列（FPGA）等。It should be understood that various parts of the present invention may be implemented in hardware, software, firmware or a combination thereof. In the above-described embodiments, various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented by any one or a combination of the following techniques known in the art: Discrete logic circuits, ASICs with suitable combinational logic gates, Programmable Gate Arrays (PGA), Field Programmable Gate Arrays (FPGA), etc.
在本说明书的描述中，参考术语“一个实施例”、“一些实施例”、“示例”、“具体示例”、或“一些示例”等的描述意指结合该实施例或示例描述的具体特征、结构、材料或者特点包含于本发明的至少一个实施例或示例中。在本说明书中，对上述术语的示意性表述不一定指的是相同的实施例或示例。而且，描述的具体特征、结构、材料或者特点可以在任何的一个或多个实施例或示例中以合适的方式结合。In the description of this specification, description with reference to the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples", etc., mean specific features described in connection with the embodiment or example , structure, material or feature is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
以上所述实施例仅表达了本发明的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对本发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本发明构思的前提下，还可以做出若干变形和改进，这些都属于本发明的保护范围。因此，本发明专利的保护范围应以所附权利要求为准。The above-mentioned embodiments only represent several embodiments of the present invention, and the descriptions thereof are specific and detailed, but should not be construed as a limitation on the scope of the patent of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can also be made, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention should be subject to the appended claims.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210432885.7A CN114531729B (en) | 2022-04-24 | 2022-04-24 | Positioning method, system, storage medium and device based on channel state information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210432885.7A CN114531729B (en) | 2022-04-24 | 2022-04-24 | Positioning method, system, storage medium and device based on channel state information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114531729A true CN114531729A (en) | 2022-05-24 |
| CN114531729B CN114531729B (en) | 2022-08-09 |
Family
ID=81628121
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210432885.7A Active CN114531729B (en) | 2022-04-24 | 2022-04-24 | Positioning method, system, storage medium and device based on channel state information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114531729B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116343966A (en) * | 2023-03-27 | 2023-06-27 | 山东大学 | Probabilistic multiplication and accumulation structural damage imaging positioning method and system based on delay factor |
| CN116527462A (en) * | 2023-05-25 | 2023-08-01 | 兰州交通大学 | A wireless positioning method based on channel state change CSI value |
| CN117062001A (en) * | 2023-08-01 | 2023-11-14 | 哈尔滨工程大学 | 5G NR indoor positioning method and system based on interpretable deep learning |
| WO2024138739A1 (en) * | 2022-12-30 | 2024-07-04 | Oppo广东移动通信有限公司 | Method and apparatus for acquiring label information, and device and storage medium |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105929419A (en) * | 2016-04-13 | 2016-09-07 | 南京理工大学 | BP artificial neural network-based GPS carrier tracking method |
| CN106772219A (en) * | 2017-03-08 | 2017-05-31 | 南京大学 | Indoor orientation method based on CSI signals |
| WO2017100706A1 (en) * | 2015-12-09 | 2017-06-15 | Origin Wireless, Inc. | Method, apparatus, and systems for wireless event detection and monitoring |
| CN109151731A (en) * | 2018-06-14 | 2019-01-04 | 西北师范大学 | A kind of indoor orientation method based on CSI |
| US20190049551A1 (en) * | 2017-08-09 | 2019-02-14 | SWFL, Inc., d/b/a "Filament" | Systems and methods for coherence based positioning |
| CN110222776A (en) * | 2019-06-10 | 2019-09-10 | 安徽磐众信息科技有限公司 | A kind of indoor Passive Location based on CSI finger print information |
| CN111556431A (en) * | 2020-04-21 | 2020-08-18 | 西安交通大学 | Fingerprint indoor positioning method based on signal subspace |
| CN111770436A (en) * | 2020-05-07 | 2020-10-13 | 普联国际有限公司 | A CSI-based indoor Wi-Fi positioning method, device and storage medium |
| CN111866713A (en) * | 2020-07-09 | 2020-10-30 | 北京邮电大学 | A positioning method, device and electronic device |
| CN112040400A (en) * | 2020-08-25 | 2020-12-04 | 西安交通大学 | Single-site indoor fingerprint positioning method based on MIMO-CSI, storage medium and equipment |
| CN112147573A (en) * | 2020-09-14 | 2020-12-29 | 山东科技大学 | A passive positioning method based on CSI amplitude and phase information |
| CN112954791A (en) * | 2021-01-26 | 2021-06-11 | 电子科技大学 | Channel State Information (CSI) positioning method based on subcarrier screening |
| CN113852908A (en) * | 2021-07-29 | 2021-12-28 | 上海大学 | Indoor positioning method based on antenna extension |
-
2022
- 2022-04-24 CN CN202210432885.7A patent/CN114531729B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017100706A1 (en) * | 2015-12-09 | 2017-06-15 | Origin Wireless, Inc. | Method, apparatus, and systems for wireless event detection and monitoring |
| CN105929419A (en) * | 2016-04-13 | 2016-09-07 | 南京理工大学 | BP artificial neural network-based GPS carrier tracking method |
| CN106772219A (en) * | 2017-03-08 | 2017-05-31 | 南京大学 | Indoor orientation method based on CSI signals |
| US20190049551A1 (en) * | 2017-08-09 | 2019-02-14 | SWFL, Inc., d/b/a "Filament" | Systems and methods for coherence based positioning |
| CN109151731A (en) * | 2018-06-14 | 2019-01-04 | 西北师范大学 | A kind of indoor orientation method based on CSI |
| CN110222776A (en) * | 2019-06-10 | 2019-09-10 | 安徽磐众信息科技有限公司 | A kind of indoor Passive Location based on CSI finger print information |
| CN111556431A (en) * | 2020-04-21 | 2020-08-18 | 西安交通大学 | Fingerprint indoor positioning method based on signal subspace |
| CN111770436A (en) * | 2020-05-07 | 2020-10-13 | 普联国际有限公司 | A CSI-based indoor Wi-Fi positioning method, device and storage medium |
| CN111866713A (en) * | 2020-07-09 | 2020-10-30 | 北京邮电大学 | A positioning method, device and electronic device |
| CN112040400A (en) * | 2020-08-25 | 2020-12-04 | 西安交通大学 | Single-site indoor fingerprint positioning method based on MIMO-CSI, storage medium and equipment |
| CN112147573A (en) * | 2020-09-14 | 2020-12-29 | 山东科技大学 | A passive positioning method based on CSI amplitude and phase information |
| CN112954791A (en) * | 2021-01-26 | 2021-06-11 | 电子科技大学 | Channel State Information (CSI) positioning method based on subcarrier screening |
| CN113852908A (en) * | 2021-07-29 | 2021-12-28 | 上海大学 | Indoor positioning method based on antenna extension |
Non-Patent Citations (5)
| Title |
|---|
| JINGJING WANG等: "first_page settings", 《SENSORS》 * |
| XIANDI LI等: "DeFe: indoor localization based on channel state information feature using deep learning", 《JOURNAL OF PHYSICS: CONFERENCE SERIES 1303》 * |
| 刘帅等: "一种基于卷积神经网络的CSI指纹室内定位方法", 《工程科学学报》 * |
| 江小平等: "基于信道状态信息幅值-相位的被动式室内指纹定位", 《电子与信息学报》 * |
| 韩卓宜等: "基于CSI指纹序列匹配的LTE室内定位方法", 《无线电工程》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024138739A1 (en) * | 2022-12-30 | 2024-07-04 | Oppo广东移动通信有限公司 | Method and apparatus for acquiring label information, and device and storage medium |
| CN116343966A (en) * | 2023-03-27 | 2023-06-27 | 山东大学 | Probabilistic multiplication and accumulation structural damage imaging positioning method and system based on delay factor |
| CN116343966B (en) * | 2023-03-27 | 2023-11-17 | 山东大学 | Probabilistic multiplication-accumulation structural damage imaging positioning method and system based on delay factor |
| US12062167B1 (en) | 2023-03-27 | 2024-08-13 | Shandong University | Probability multiply-sum structural damage imaging positioning method and system based on delay factor |
| CN116527462A (en) * | 2023-05-25 | 2023-08-01 | 兰州交通大学 | A wireless positioning method based on channel state change CSI value |
| CN116527462B (en) * | 2023-05-25 | 2024-02-02 | 兰州交通大学 | A wireless positioning method based on channel state changes CSI value |
| CN117062001A (en) * | 2023-08-01 | 2023-11-14 | 哈尔滨工程大学 | 5G NR indoor positioning method and system based on interpretable deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114531729B (en) | 2022-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114531729B (en) | Positioning method, system, storage medium and device based on channel state information | |
| CN110632572B (en) | Radar radiation source individual identification method and device based on unintentional phase modulation characteristics | |
| CN107396322B (en) | Indoor positioning method based on path matching and coding-decoding cyclic neural network | |
| CN108696331B (en) | A Signal Reconstruction Method Based on Generative Adversarial Networks | |
| CN105607039B (en) | Robust least squares localization method based on reaching time-difference under nlos environment | |
| CN108169708B (en) | Direct positioning method of modular neural network | |
| CN109581281B (en) | Moving target positioning method based on arrival time difference and arrival frequency difference | |
| CN112014791B (en) | Near-field source positioning method of array PCA-BP algorithm with array error | |
| CN110187335A (en) | Particle filter tracking before detection method for targets with discontinuous characteristics | |
| CN114449452A (en) | An Indoor Localization Algorithm for Heterogeneous Devices | |
| CN114201987A (en) | An Active Interference Identification Method Based on Adaptive Identification Network | |
| CN110261818B (en) | Indirect ultra-wideband signal identification and error elimination method and device, storage medium | |
| CN116166642A (en) | Spatio-temporal data filling method, system, equipment and medium based on guide information | |
| CN115422977A (en) | Radar Radiator Signal Recognition Method Based on CNN-BLS Network | |
| CN110568406B (en) | A localization method based on acoustic energy under the condition of unknown energy attenuation factor | |
| CN106612158A (en) | Signal blind detection method based on complex sinusoidal chaotic neural network | |
| CN111505575A (en) | Sensor selection method aiming at TDOA (time difference of arrival) location based on conversion TOA (time of arrival) model | |
| CN111505576B (en) | Sensor selection method aiming at TDOA (time difference of arrival) location | |
| CN116645973A (en) | Directional audio enhancement method and device, storage medium and electronic equipment | |
| CN108564171A (en) | A kind of neural network sound source angle method of estimation based on quick global K mean cluster | |
| CN110988800A (en) | Semi-positive relaxation positioning method based on acoustic energy | |
| CN111277950B (en) | Positioning method, device and equipment based on arrival time difference and arrival frequency difference | |
| KR20230001777A (en) | Method for classifying electromagnetic pulse based on neural network | |
| CN113255887A (en) | Radar error compensation method and system based on genetic algorithm optimization BP neural network | |
| CN117829274B (en) | Model fusion method, device, equipment, federal learning system and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |