CN104217231A

CN104217231A - RFID positioning system and positioning method based on non-accurate anchor nodes

Info

Publication number: CN104217231A
Application number: CN201410439297.1A
Authority: CN
Inventors: 谢磊; 鲁翔; 殷亚凤; 王炜; 陆桑璐
Original assignee: Nanjing University
Current assignee: Nanjing University
Priority date: 2014-08-29
Filing date: 2014-08-29
Publication date: 2014-12-17

Abstract

The invention discloses an RFID positioning system and positioning method based on non-accurate anchor nodes. The system mainly includes shelves, goods, RFID tags, readers, antennas and the like. In the environment to be positioned, shelves with known coordinate positions are deployed, and goods with RFID tags are placed on the shelves. The precise location of the goods is unknown, and only the start and end positions of the goods on the shelf can be obtained. The present invention uses these goods labels without precise location information as anchor nodes, builds an RFID positioning system based on the distribution of goods types, and proposes an effective positioning method. The system does not need to obtain the precise location information of the goods, but only needs to complete the positioning operation through the scanned goods classification information.

Description

An RFID positioning system and positioning method based on non-accurate anchor nodes

技术领域technical field

本发明涉及一种RFID定位方法，具体是一种基于非精确锚节点的RFID定位系统及定位方法。The invention relates to an RFID positioning method, in particular to an RFID positioning system and positioning method based on non-accurate anchor nodes.

背景技术Background technique

随着RFID技术的广泛应用，越来越多的RFID标签被使用，使得在一些环境中(比如超市、图书馆、货品仓库等)会有着大量RFID标签的存在，虽然不能获取这些锚节点的精确位置，但其分类所在位置范围已知。上述环境中也有着对人员进行定位的需求，以提供更为人性化的服务，如：导航、附近货品查询等。然而目前仅仅依靠现有环境中RFID标签，无法准确定位人员，必须借助于参考标签。With the widespread application of RFID technology, more and more RFID tags are used, so that there will be a large number of RFID tags in some environments (such as supermarkets, libraries, goods warehouses, etc.), although the precise location of these anchor nodes cannot be obtained. location, but the range of locations where it is classified is known. In the above-mentioned environment, there is also a need to locate personnel to provide more humanized services, such as: navigation, query of nearby goods, etc. However, only relying on the RFID tags in the existing environment at present cannot accurately locate personnel, and must rely on reference tags.

而货品上的标签，本来是用于标识货品的信息便于获知其相关信息，但是由于RFID技术中包含的基础无线特性(信号衰减、信号传播的方向性等)，实际可以扩展其他功能，然而现有技术并没有充分挖掘其中的潜力。本发明通过利用这些看似别无他用的货品标签，即可对目标物体进行定位工作。The label on the goods is originally used to identify the information of the goods so as to obtain its relevant information, but due to the basic wireless characteristics (signal attenuation, signal propagation directionality, etc.) contained in RFID technology, other functions can actually be expanded, but now There are technologies that are not exploiting their full potential. The present invention can locate the target object by using these seemingly useless goods labels.

发明内容Contents of the invention

本发明所要解决的技术问题是提供一种基于非精确锚节点的RFID定位系统及其方法，该系统无须另外布置参考标签节点，只利用已有货品上的RFID标签进行定位操作，即可准确定位用户位置。The technical problem to be solved by the present invention is to provide an RFID positioning system and method based on non-accurate anchor nodes. The system does not need to arrange additional reference tag nodes, and only uses the RFID tags on the existing goods for positioning operations to accurately locate user location.

本发明所述的一种基于非精确锚节点的RFID定位系统，其包括以下组件：A kind of RFID positioning system based on the non-accurate anchor node of the present invention, it comprises the following components:

货架：用于摆放贴有RFID标签的货品，货架的位置固定，并知道每个货架上货品分类的起止位置；Shelves: used to place goods with RFID tags, the positions of the shelves are fixed, and the start and end positions of the classification of goods on each shelf are known;

RFID标签：贴在货品上的包含货品信息的普通RFID标签；RFID tag: an ordinary RFID tag that contains product information attached to the product;

移动装置：其上搭载RFID阅读器和天线，由用户推动，通过对移动装置的定位，即可得到用户的位置信息；Mobile device: It is equipped with RFID reader and antenna, which is driven by the user, and the location information of the user can be obtained by locating the mobile device;

RFID阅读器和天线：用于收集货品的标签信息，并通过网络传输给运算服务器中的信息处理模块；RFID reader and antenna: used to collect the label information of the goods, and transmit it to the information processing module in the computing server through the network;

信息处理模块：由收集到的货品的标签信息，计算出移动装置的位置。Information processing module: Calculate the position of the mobile device from the collected label information of the goods.

本发明还提供了基于非精确锚节点的RFID定位系统的定位方法，其包括以下步骤：The present invention also provides a positioning method based on the RFID positioning system of the non-accurate anchor node, which includes the following steps:

1)用户推动装载RFID阅读器及天线的移动装置，进入在货架的物品上部署RFID标签的可定位环境中，该环境中货架的位置固定，即各排货架的y坐标位置已知；1) The user pushes the mobile device loaded with RFID readers and antennas to enter the positionable environment where RFID tags are deployed on the items on the shelf. The position of the shelf in this environment is fixed, that is, the y-coordinate position of each row of shelves is known;

2)RFID阅读器及天线收集周边物品上的RFID标签信息，并通过网络传入信息处理模块进行位置运算操作；2) The RFID reader and antenna collect the RFID tag information on the surrounding items, and transmit the information processing module through the network for position calculation operation;

3)信息处理模块获得各RFID标签所属分类的信息，由于待定位目标所在货架y坐标位置已知，只需计算待定位目标所在的x坐标位置，即可定位目标。3) The information processing module obtains the classification information of each RFID tag. Since the y-coordinate position of the shelf where the target is located is known, it only needs to calculate the x-coordinate position of the target to be located to locate the target.

对于步骤3)的具体定位过程有以下几种并列方法：For the specific positioning process of step 3), there are the following parallel methods:

其一：信息处理模块获得各RFID标签所属分类的信息，因为每个种类的起始位置x_i,s及结束位置x_i,e已知,使用各种类中被识别的标签数n_i为加权权值,通过加权求平均的方法即可以计算得出目标的x坐标位置，该计算过程用以下公式进行描述:First: the information processing module obtains the information of the classification of each RFID tag, because the starting position x _{i, s} and the end position x _{i, e} of each category are known, and the number of identified tags in each category n _i is Weighted weight, the x-coordinate position of the target can be calculated by weighted average method, the calculation process is described by the following formula:

$\overset{^^}{x x} = = {Σ Σ}_{i i = = 11}^{k k} {w w}_{i i} \cdot &Center Dot; \frac{{x x}_{i i,, s the s} + + {x x}_{i i,, e e}}{22},,$

${w w}_{i i} = = \frac{{n no}_{i i}}{{Σ Σ}_{i i = = 11}^{k k} {n no}_{i i}}$

这里,k是被识别的种类数,w_i是相对应分类的权值,其可通过使用该种类中被识别的标签数除以所有被识别的标签数得到；在这个方法中,使用货品种类的中间位置作为种类的位置。Here, k is the number of identified categories, and w _i is the weight of the corresponding category, which can be obtained by dividing the number of identified tags in this category by the number of all identified tags; in this method, use the category of goods middle position position as a species.

其二：Second:

信息处理模块获得各RFID标签所属分类的信息，因为每个种类的起始位置x_i,s及结束位置x_i,e已知,而各RFID标签距离天线在识别区域上的投影距离越近,其RSSI值就越高，这里使用每个货品种类的总RSSI值做为权值,通过加权求平均的方法即可以计算得出目标的x坐标位置，计算方法描述如下:The information processing module obtains the classification information of each RFID tag, because the starting position x _{i, s} and the end position x _{i, e} of each type are known, and the closer the projection distance of each RFID tag to the antenna on the identification area, The higher the RSSI value is, the total RSSI value of each product type is used as the weight here, and the x-coordinate position of the target can be calculated by weighted average method. The calculation method is described as follows:

${w w}_{i i} = = \frac{{s the s}_{i i}}{{Σ Σ}_{i i = = 11}^{k k} {s the s}_{i i}}$

其中，k表示选取的k个具有最强RSSI平均值的货品分类，s_i为获取的货品分类中所有标签的RSSI值的平均值，w_i为根据s_i算出的每个分类权值，由于每个分类的起止位置已知，即可根据该算法算出移动装置的位置信息。Among them, k represents the selected k product categories with the strongest average RSSI value, s _i is the average value of RSSI values of all tags in the obtained product category, and w _i is the weight value of each category calculated according to _si , because The start and end positions of each classification are known, and the position information of the mobile device can be calculated according to the algorithm.

其三：Third:

信息处理模块获得各RFID标签所属分类的信息，因为每个种类的起始位置x_i,s及结束位置x_i,e已知；与此同时，在扫描过程中,可以获知每个分类里面的标签数,把这些获得的各分类标签数作为一个向量，所有被识别到的标签分类所覆盖的范围作为可能位置范围,从可能位置范围的左侧开始,每隔△d距离,就以可能的位置为中心,以主要识别区间的半径向外画一个圈,可以获得一系列的标签数向量表，通过匹配识别的标签数向量和计算得出的标签向量,通过一个k近邻算法,找到k个最类似的向量,再根据他们的可能位置,估算出目标的x坐标位置。The information processing module obtains the classification information of each RFID tag, because the starting position x _{i, s} and the end position x _{i, e} of each type are known; at the same time, during the scanning process, the information in each classification can be known The number of labels, the number of labels of each category obtained as a vector, the range covered by all identified label categories as the possible position range, starting from the left side of the possible position range, every △d distance, with the possible position range Position as the center, draw a circle outward with the radius of the main recognition interval, you can get a series of label number vector table, by matching the recognized label number vector and the calculated label vector, through a k-nearest neighbor algorithm, find k The most similar vectors, and then according to their possible positions, estimate the x-coordinate position of the target.

其四：Fourth:

通过实验得到了不同RFID标签密度情况下，RSSI强度与标签到天线投影距离间的关系训练集T_d，通过计算每个标签距离天线投影距离，给出每个标签对于投影位置的预判位置；通过计算这些预判位置的中心位置，即为需要计算的目标位置。The training set T _d of the relationship between RSSI strength and the projection distance from the tag to the antenna is obtained through experiments under different RFID tag densities. By calculating the distance between each tag and the antenna projection, the predicted position of each tag for the projection position is given; By calculating the center position of these predicted positions, it is the target position that needs to be calculated.

本发明直接使用附着在货品上的标签,对用户进行定位,以提供更为个性化的服务,如:导航、附近货品查询、实时广告推送等基于位置的服务。当用户推着购物车行进购物时,购物车上装有的阅读器通过天线扫描并获取周边环境中的标签信息,使用本文后续提出的相应算法即可计算得出顾客的大概位置。本发明通过使用这些没有精确位置信息的货品标签为锚节点，基于货品种类的分布情况，搭建了一种RFID定位系统，并提出四种行之有效的定位方法。该系统不需要获得货品的精确位置信息，无需建立参考标签网络，只需要通过扫描到的货品分类信息即可完成定位操作，简化了操作手续，降低了使用开销。The present invention directly uses the tags attached to the goods to locate the users to provide more personalized services, such as location-based services such as navigation, query of nearby goods, and real-time advertisement push. When the user pushes the shopping cart to go shopping, the reader installed on the shopping cart scans through the antenna and obtains the tag information in the surrounding environment, and the approximate location of the customer can be calculated by using the corresponding algorithm proposed in this paper. The present invention builds an RFID positioning system based on the distribution of goods types by using these goods tags without precise location information as anchor nodes, and proposes four effective positioning methods. The system does not need to obtain the precise location information of the goods, and does not need to establish a reference label network. It only needs to complete the positioning operation through the scanned goods classification information, which simplifies the operation procedures and reduces the cost of use.

附图说明Description of drawings

图1是本发明的系统架构图，Fig. 1 is a system architecture diagram of the present invention,

图2是本发明中货架及货品部署的原型图，Fig. 2 is a prototype diagram of shelves and goods deployment in the present invention,

图3是本发明的整体部署图，Fig. 3 is the overall deployment diagram of the present invention,

图4是基于RSSI方法的工作流程图，Fig. 4 is a workflow diagram based on the RSSI method,

图5是本发明中分类匹配算法原理图，Fig. 5 is a principle diagram of classification matching algorithm in the present invention,

图6是本发明中距离投票算法原理图，Fig. 6 is a schematic diagram of the middle-distance voting algorithm of the present invention,

图7是本发明中距离投票算法流程图。Fig. 7 is a flow chart of the middle-distance voting algorithm of the present invention.

具体实施方法Specific implementation method

下面结合附图对本发明作进一步描述。The present invention will be further described below in conjunction with the accompanying drawings.

如图1，本发明的RFID定位系统，其组成包括货架、无源RFID标签、移动装置、RFID阅读器+天线、具有信息处理模块的运算服务器及无线网络。图2是本发明中货架及货品部署的原型图。其中RFID标签附着在货品上，货品按其分类摆放在货架上，货架的纵横位置已知，上面摆放货品分类的起止位置已知。移动装置上安放有RFID阅读器和天线，用于获取周边的标签信息。信息处理模块为ARM9开发板，通过处理获得的标签信息的统计数据及RSSI信息，返回移动装置的位置信息，即用户的位置信息。As shown in Figure 1, the RFID positioning system of the present invention consists of shelves, passive RFID tags, mobile devices, RFID readers + antennas, computing servers with information processing modules, and wireless networks. Fig. 2 is a prototype diagram of shelves and goods deployment in the present invention. Among them, RFID tags are attached to the goods, and the goods are placed on the shelves according to their categories. The vertical and horizontal positions of the shelves are known, and the start and end positions of the categories of goods placed on them are known. An RFID reader and an antenna are placed on the mobile device to obtain surrounding tag information. The information processing module is an ARM9 development board, which returns the location information of the mobile device, that is, the location information of the user, by processing the obtained statistical data of the tag information and RSSI information.

本发明的定位过程为：The positioning process of the present invention is:

上述步骤3)使用的几种定位算法描述如下：The above steps 3) several positioning algorithms used are described as follows:

如图3中系统整体部署图所示，由于待定位目标物体所在货架y坐标位置已知，在该情况下，只需计算x坐标位置，即可获得待定位目标物体的所在位置。因此该问题转化为求一维坐标问题。As shown in the overall deployment diagram of the system in Figure 3, since the y-coordinate position of the shelf where the target object to be located is known, in this case, only the x-coordinate position needs to be calculated to obtain the location of the target object to be located. Therefore, the problem is transformed into a one-dimensional coordinate problem.

算法一：基于分类方法Algorithm 1: Based on the classification method

这是一个较为基础的解决方案。通过扫描,阅读器可以获得一系列的标签信息,并能获得标签所属分类的信息。因为每个种类的起始位置x_i,s及结束位置x_i,e已知,使用各种类中被识别的标签数n_i为加权权值,通过加权求平均的方法即可以计算得出目标物体的位置。该计算过程可以用以下公式进行描述:This is a more basic solution. By scanning, the reader can obtain a series of label information, and can obtain the information of the category to which the label belongs. Because the starting position x _{i, s} and end position x _{i, e} of each category are known, the number of identified labels in each category n _i is used as the weighted weight value, which can be calculated by weighted average method The location of the target object. The calculation process can be described by the following formula:

${w w}_{i i} = = \frac{{n no}_{i i}}{{Σ Σ}_{i i = = 11}^{k k} {n no}_{i i}}$

这里,k是被识别的种类数,w_i是相对应分类的权值,其可通过使用该种类中被识别的标签数除以所有被识别的标签数得到。在这个方法中,使用货品种类的中间位置作为种类的位置。而即为所求。Here, k is the number of identified categories, and _wi is the weight of the corresponding category, which can be obtained by dividing the number of identified labels in this category by the number of all identified labels. In this method, use the middle position of the item type position as a species. And that is what you want.

算法二：基于RSSI方法Algorithm 2: Based on RSSI method

在上一个算法中,仅仅只用到了各个识别到的货品种类的标签个数,然而RSSI对于衡量标签分布来首,是一个非常重要的因素。在前文中提到,距离天线在识别区域上的投影距离越近,其RSSI值就越高。这里使用每个货品种类的总RSSI值做为权值,计算方法可以描述如下:In the previous algorithm, only the number of tags of each identified product category is used, but RSSI is a very important factor for measuring the distribution of tags. As mentioned above, the closer the projection distance of the antenna on the identification area is, the higher the RSSI value will be. Here, the total RSSI value of each commodity type is used as the weight, and the calculation method can be described as follows:

$\overset{^^}{x x} = = {Σ Σ}_{i i = = 11}^{k k} {w w}_{i i} \cdot \cdot \frac{{x x}_{i i,, s the s} + + {x x}_{i i,, e e}}{22},,$

其中，k表示选取的k个具有最强RSSI平均值的货品分类，s_i为获取的货品分类中所有标签的RSSI值的平均值，w_i为根据s_i算出的每个分类权值，由于每个分类的起止位置已知，即可根据该算法算出移动装置的位置信息(图4)。Among them, k represents the selected k product categories with the strongest average RSSI value, s _i is the average value of RSSI values of all tags in the obtained product category, and w _i is the weight value of each category calculated according to _si , because The start and end positions of each category are known, and the location information of the mobile device can be calculated according to this algorithm (Figure 4).

算法三分类匹配算法Algorithm Three Classification Matching Algorithm

在扫描过程中,可以获知每个分类里面的标签数,把这些获得的各分类标签数作为一个向量。把所有被识别到的标签分类所覆盖的范围作为可能位置范围,从可能位置范围的左侧开始,每隔△d距离,就以可能的位置为中心,以主要识别区间的半径向外画一个圈,我们可以获得一系列的标签数向量表。通过匹配识别的标签数向量和计算得出的标签向量,同过一个k近邻算法,找到k个最类似的向量,再根据他们的可能位置,估算出最终位置。During the scanning process, the number of labels in each category can be known, and the obtained number of labels in each category can be used as a vector. Take the range covered by all identified label categories as the possible position range, start from the left side of the possible position range, and draw a circle outward with the radius of the main recognition interval centered on the possible position at every △d distance. Circle, we can get a series of vector tables of label numbers. By matching the recognized label number vector and the calculated label vector, with a k-nearest neighbor algorithm, find the k most similar vectors, and then estimate the final position according to their possible positions.

如图5所示,浅色圆圈所示区域即为主要识别区间,其半径为r。根据扫描结果,我们可以得到一个分类标签数向量V₀＝{n_0,1,n_0,2,...,n_0,s}。基于识别半径与标签密度及阅读器能量的关系训练集T_a，我们可以得到当前情况下的主要识别区间的半径r,图中虚线圆圈所示区域即为可能的主要识别区间。基于识别标签数与标签密度及阅读器能量的关系训练集T_b,我们可以算出标签密度ρ。以此虚线圈框住的的部分,各个分类也可以生成一个分类标签数向量V_i＝(n_i,1,n_i,2,...,n_i,s)。通过比较向量V₀和向量V_i间的相似度,我们可以得到k个最类似的向量,并获得它们的可能位置x₁,...,x_k,再通过一个向量距离的加权算法,我们即可计算出最终的定位位置。As shown in Figure 5, the area indicated by the light-colored circle is the main identification interval, and its radius is r. According to the scanning result, we can obtain a classification label number vector V ₀ ={n _0,1 ,n _0,2 ,...,n _0,s }. Based on the training set T _a of the relationship between the recognition radius, tag density and reader energy, we can obtain the radius r of the main recognition interval in the current situation, and the area indicated by the dotted circle in the figure is the possible main recognition interval. Based on the training set T _b of the relationship between the number of identified tags and the tag density and reader energy, we can calculate the tag density ρ. Each category can also generate a category label number vector V _i =(n _i,1 ,n _i,2 ,...,n _i,s ) for the part enclosed by the dotted circle. By comparing the similarity between vector V ₀ and vector V _i , we can get the k most similar vectors, and obtain their possible positions x ₁ ,...,x _k , and then through a vector distance weighting algorithm, we The final positioning position can be calculated.

算法四距离投票算法Algorithm Four Distance Voting Algorithm

为解决基于RSSI加权平均的算法中的限制,我们做了大量实验，得到了不同标签密度情况下，RSSI强度与标签到天线投影距离间的关系训练集T_d。通过计算每个标签距离天线投影距离，给出每个标签对于投影位置的预判位置。通过计算这些预判位置的中心位置，即为需要计算的目标位置。In order to solve the limitations of the algorithm based on RSSI weighted average, we have done a lot of experiments and obtained the training set T _d of the relationship between the RSSI intensity and the projection distance from the tag to the antenna under different tag densities. By calculating the projection distance of each tag from the antenna, the predicted position of each tag for the projected position is given. By calculating the center position of these predicted positions, it is the target position that needs to be calculated.

基于训练集T_d，我们可以通过插值法算出每个tag距离天线投影的距离。我们以标签分类的中间位置作为标签的位置d_i,j。因此每个分类都可以算出一个预判位置:其中d_i,j越小，其预判位置越准确，我们对算出的d_i,j进行正序排序，获得前k个预判位置，算其平均值，即可计算出最终的定位位置。图6为该定位方法示意图,图7为该定位方法流程图。Based on the training set T _d , we can calculate the distance of each tag from the antenna projection by interpolation. The middle position we classify by label as position d _i,j of the label. Therefore, each classification can calculate a predicted position: The smaller d _{i, j} is, the more accurate the predicted position is. We sort the calculated d _{i, j} in positive order to obtain the first k predicted positions, and calculate the average value to calculate the final positioning position. FIG. 6 is a schematic diagram of the positioning method, and FIG. 7 is a flowchart of the positioning method.

Claims

1. An RFID positioning system based on inaccurate anchor nodes, characterized in that it comprises the following components:

Shelves: used to place goods with RFID tags, the positions of the shelves are fixed, and the start and end positions of the classification of goods on each shelf are known;

RFID tag: an ordinary RFID tag that contains product information attached to the product;

Mobile device: It is equipped with RFID reader and antenna, which is driven by the user, and the location information of the user can be obtained by locating the mobile device;

RFID reader and antenna: used to collect the label information of the goods, and transmit it to the information processing module in the computing server through the network;

Information processing module: Calculate the location of the mobile device from the collected label information of the goods.

2. the positioning method of the RFID positioning system based on the non-accurate anchor node of claim 1, is characterized in that comprising the following steps:

1) The user pushes the mobile device loaded with RFID readers and antennas to enter the positionable environment where RFID tags are deployed on the items on the shelf. The position of the shelf in this environment is fixed, that is, the y-coordinate position of each row of shelves is known;

2) The RFID reader and antenna collect the RFID tag information on the surrounding items, and transmit the information processing module through the network to perform position calculation operations;

3) The information processing module obtains the classification information of each RFID tag. Since the y-coordinate position of the shelf where the target is located is known, it only needs to calculate the x-coordinate position of the target to be located to locate the target.

3. the positioning method based on the RFID positioning system of non-accurate anchor node according to claim 2, it is characterized in that the concrete process of step 3) is:

The information processing module obtains the classification information of each RFID tag, because the starting position x _{i, s} and the end position x _{i, e} of each category are known, and the number of identified tags in each category n _i is used as the weighted value , the x-coordinate position of the target can be calculated by weighted average method, and the calculation process is described by the following formula:

\overset{^^}{x x} = = {Σ Σ}_{i i = = 11}^{k k} {w w}_{i i} \cdot \cdot \frac{{x x}_{i i,, s the s} + + {x x}_{i i,, e e}}{22},,

{w w}_{i i} = = \frac{{s the s}_{i i}}{{Σ Σ}_{i i = = 11}^{k k} {s the s}_{i i}}

Here, k is the number of identified categories, and w _i is the weight of the corresponding category, which can be obtained by dividing the number of identified tags in this category by the number of all identified tags; in this method, use the category of goods middle position position as a species.

4. the positioning method based on the RFID positioning system of non-accurate anchor node according to claim 2, is characterized in that the concrete process of step 3) is:

The information processing module obtains the classification information of each RFID tag, because the starting position x _{i, s} and the end position x _{i, e} of each type are known, and the closer the projection distance of each RFID tag to the antenna on the identification area, The higher the RSSI value is, the total RSSI value of each product type is used as the weight here, and the x-coordinate position of the target can be calculated by weighted average method. The calculation method is described as follows:

\overset{^^}{x x} = = {Σ Σ}_{i i = = 11}^{k k} {w w}_{i i} \cdot &Center Dot; \frac{{x x}_{i i,, s the s} + + {x x}_{i i,, e e}}{22},,

{w w}_{i i} = = \frac{{s the s}_{i i}}{{Σ Σ}_{i i = = 11}^{k k} {s the s}_{i i}}

Among them, k represents the selected k product categories with the strongest average RSSI value, s _i is the average value of RSSI values of all tags in the obtained product category, and w _i is the weight value of each category calculated according to _si , because The start and end positions of each classification are known, and the position information of the mobile device can be calculated according to the algorithm.

5. the positioning method based on the RFID positioning system of non-accurate anchor node according to claim 2, it is characterized in that the concrete process of step 3) is:

The information processing module obtains the classification information of each RFID tag, because the starting position x _{i, s} and the end position x _{i, e} of each type are known; at the same time, during the scanning process, the information in each classification can be known The number of labels, the number of labels of each category obtained as a vector, the range covered by all identified label categories as the possible position range, starting from the left side of the possible position range, every △d distance, with the possible position range Position as the center, draw a circle outward with the radius of the main recognition interval, you can get a series of label number vector table, by matching the recognized label number vector and the calculated label vector, through a k-nearest neighbor algorithm, find k The most similar vectors, and then according to their possible positions, estimate the x-coordinate position of the target.

6. the positioning method based on the RFID positioning system of non-accurate anchor node according to claim 2, it is characterized in that the concrete process of step 3) is:

The training set T _d of the relationship between RSSI strength and the projection distance from the tag to the antenna is obtained through experiments under different RFID tag densities. By calculating the distance between each tag and the antenna projection, the predicted position of each tag for the projection position is given; By calculating the center position of these predicted positions, it is the target position that needs to be calculated.