WO2012146801A1 - Intelligent location-finding method using wireless sensor networks - Google Patents

Abstract

The invention relates to an intelligent localisation method using wireless sensor networks, comprising a first distributed process that is executed in each node and carries out an estimation of the possible location of the tag; and a second centralised process that combines all of said partial information to form a single solution.

Description

 INTELLIGENT LOCATION METHOD ON SENSOR NETWORKS

 WIRELESS

The present invention relates to a method and a system for estimating the position of mobile equipment, whose position is unknown from a series of known position equipment.

The main object of the present invention is an intelligent location system on wireless sensor networks, characterized by its low consumption and for being designed to be used in wireless sensor networks, and which determines the position of the mobile nodes by using computational intelligence techniques

BACKGROUND OF THE INVENTION

A network of wireless sensors is made up of a series of small, low-power devices, with the ability to communicate wirelessly with each other and collaborate to achieve a common goal. Each of the elements of this network, called nodes or tags, are formed by a wireless communications transceiver that allows communication with other nodes, a microcontroller that processes information, a periphery that allows the interconnection of sensors and a system of power that allows it to operate autonomously.

In a location system based on wireless sensor networks, two classes of devices are distinguished:

(i) Anchored nodes: those that are in fixed positions and known in advance.

 (ii) Nodes anchored: their positions are unknown.

They are the elements to locate. The location consists in determining the positions of the non-anchored nodes from the interfered data of the anchored nodes.

The localization methods existing in the literature can be classified into two large families:

1.- Methods based on measurements (Range-based)

 They are based on the estimation of the point-to-point distance between all nodes. From this information and using some graphical theory such as triangulation, the information of the non-anchored nodes can be inferred.

There are several distance measurement methods, such as:

 a) Those based on received field strength (RSSI) [A Awad, T Frunzke, F Dressler, Adaptive distanee estimation and localization in WSN using RSSI measures, Proc- Euromicro Conf. Digit. Syst Des. Archit. , Methods Tools, DSD. (2007) 471-478].

 b) Arrival time (TOA) [S-Wu, N- Zhang. Two-step TOA estimation method for UWB based wireless sensor networks, Ruan Jian Xue Bao. 18 (2007) 1164-1172].

 c) Arrival time difference (TDOA) [S Xiaoyan, L Jiandong, H Pengyu, P Jiyong, Total least-squares solution of active target localization using TDOA and FDOA measurements in WSN, Proc. Int. Conf. Adv. Netw Inf. Appl. AINA (2008) 995-999].

 d) Angle of incidence [D Niculescu, B Nath, Ad hoc positioning system (APS) using AOA, Proc IEEE INFOCOM. 3

(2003) 1734-1743] [K Kucuk, A Kavak, H Yigit, C Ozdemir, A novel localization technique for wireless sensor networks using adaptive antenna arrays, IEEE Radio Wirel. Sywp., RWS. (2008) 483-486] [D Niculescu, B Nath. ocalized positioning in ad hoc networks, Ad Hoc Networks. 1 (2003) 247-259] [K Kucuk, A Kavak. Scalable location estimation using smart antennas in wireless sensor networks, Ad Hoc Net-works 8 (2010) 889-903] [US 2007/0076638 Al] or those described in patents [US 2008/0080441 Al and US 2007/0005292 Al].

 e) Other recently proposed methods [F Su, W Ren, H Jin, Loca.l za.tion algorithm based on difference estimation for wireless sensor networks, Proc. Int. Commun. Conf. Softw. Netw. , ICCSN. (2009) 499-503], are based on the use of statistical calculations for the determination of distance.

Of all of them, the only one that does not require the use of additional hardware is the reading of the received field strength (RSSI, acronym in English of "Received Signal Strength Indication"), since this function is integrated by all radio transceivers current. However, these techniques present the problem of being very sensitive to electromagnetic noise. Therefore, to improve accuracy, these systems require a calibration that also depends on changing characteristics, such as the weather.

Therefore, the determination of better models and techniques to improve the behavior of location systems based on RSSI measurement is a very active field today [H Miura, K Hirano, N Matsuda, H Taki, N Abe, S Hori, Indoor localization for mobile node based on RSSI, Lect. Notes Comput. Sci. 4694 LNAI (2007) 1065-1072] [S Tian, X Z ang, P Liu, P Sun, X Wang, A RSSI-based DV-hop algorithm for wireless sensor networks, Int Conf Wirel Commun Networking Mob Comput. (2007) 2555-2558] [Z Shan, T-P Yum, Precise localization with smart antennas in ad-hoc networks, GLOBECOM IEEE Global Telecommun. Conf. (2007) 1053-1057].

2.- Methods based on connect.life (Range-free).

In these methods, the location is obtained from implicit characteristics in communications, such as: (i) Coverage conditions: The anchored nodes try to locate the non-anchored based on the information of whether or not they are in their range of communications coverage. Based on this data, the position can be estimated as the intersection of the areas of all the nodes with which you can establish communication. An example of this is US2005 / 0080924A1.

 (ü) Jump numbers: The location is established based on the number of jumps between the non-anchored nodes and some anchored nodes. With this data and considering an average distance between nodes, the location can be established using methods such as multirateration. An example of this is US2007 / 0159986A1.

The mechanisms used to obtain this information are generally based on the exchange of messages, generally referred to as beacons. The fundamental problem of these techniques lies in the fact of needing a large amount of message exchange to obtain the estimated position. This requires having to reach a compromise solution between the battery life and the number of messages sent, since the radio is the element that spends the most energy on these devices.

However, compared to measurement-based techniques, the consumption of message exchanges is generally lower than that consumed by the specific hardware required by measurement-based techniques. In addition, the use of additional hardware increases the cost and weight of the devices, an especially important feature in mobile nodes.

There are several localization methods in the literature, such as: a) The Centroid technique (CL) [N Bulusu, J Heidemann, D Estrln. GPS-less low-cost outdoor locallzatlon for very small devices, IEEE Pers Commun. 7 (2000) 28-34], DV-Hop [GQ Gao, L Lei, An Improved node locallzatlon algorithm based on DV-HOP in WSN, Proc. - IEEE Int. Conf. Adv. Comput Control, ICACC. 4 (2010) 321-324].

 b) Convex [L Doherty, KSJ Pister, L El Ghaoui, Convex position estimation in wireless sensor networks, Proc IEEE INFOCOM. 3 (2001) 1655-1663].

 c) APIT [Y Zhou, X Ao, S Xia, An Improved APIT node self-localization algorithm in WSN, Proc. World Congr. Intelligent Control Autom. WCICA (2008) 7576-7581].

The most representative being the centroid technique (CL) [Bulusu, op.cit]. This technique estimates the position of the non-anchored node by calculating the centroid of the positions of all anchored nodes that have received the beacon.

Some modifications of the centroid technique are proposed in the literature, either seeking to obtain a better performance under certain particular conditions, such as the use of weights in the estimate, obtained from the measured value of the RSSI [J Blumenthal, R Grossmann, F Golatowski , D Timmermann, Weighted centroid localization in Zigbee-based sensor networks, IEEE Int. Sy p. Intelligent Signal Process. , WISP. (2007)] or LQI [S Schuhmann, K Herrmann, K Rothermel, J Blvimenthal, D Timmermann, Improved weighted centroid localization in smart ubiquitous environments, Lect. Notes Comput. Sci. 5061 LNCS (2008) 20-34].

Other authors focus on reducing energy consumption [R Behnke, D Timmermann, AWC: Adaptive weighted centroid localization as an efficient and provement of coarse grained localization, Workshop Positioning, Navig. Commun. , WPNC. (2008) 243-250], modifying the method so that it does not require expensive functions for this type of processors, such as square root.

Other authors [R Behnke, J Salzmann, R GróbBmann, D Lleckfeldt, D Timmermann, K Thurow, Strategies to overeóme border area effeets of coarse gralned localization, Proc. - Workshop Positlonlng, Navlg. Commun. , WPNC. (2009) 95-102] focus on studying the areas where the system obtains the greatest error, as well as on defining its accuracy based on the characteristics of the installation.

Other connectivity-based location techniques have recently been proposed, some focusing on the study of the intersection of the coverage area between anchored nodes [H Lee, JL Welch, NH Vaidya. Locatlon tracking uslng quorums in mobile ad hoc networks, Ad Hoc Networks. 1 (2003) 371-381], and others focused on modifying characteristics of the beacon, such as the emission power [H- Chu, R- Jan. A GPS- less, outdoor, self-positlonlng method for wireless sensor networks, Ad Hoc Netw. 5 (2007) 547-557], in order to determine the minimum coverage area in which there is connectivity with the node to be located.

There are also some hybrid techniques that merge several different localization methods to improve accuracy, such as that described in the patent [US 2007/0111735 Al].

DESCRIPTION OF THE INVENTION

The location techniques for sensor networks are specific solutions that should try to meet the following objectives, on the one hand they should be easy to extend in external networks, on the other, use methods that do not consume much energy, and finally, try to use the least amount of additional hardware possible, in order to keep consumption, cost and dimensions at reduced values. In view of these objectives, in many of the applications of sensor networks the use of external devices, such as GPS is discarded due mainly to the fact that it consumes a lot of energy, so that a portable application of low weight and with a high autonomy (as needed for the Tag to be integrated in the location of animals). In addition to not being usable in some environments, as occurs indoors.

If you intend to have small, economical and high autonomy devices, the average flight time techniques of the communications signal are not a good option, since the computing power of these devices is very small and in order to have a close accuracy to the meter, clocks of at least 300 MHz would be required. Ultrasonic flight time measurements, although feasible with these systems, are ruled out by the need to require additional hardware, in addition to raising the need to require direct vision and without obstacles, having a typical range of about three meters.

Therefore, it is considered that the best solution involves the use of radio for localization, classical techniques based on measurements using RSSI not being considered valid due to their high sensitivity to disturbances.

Instead, a technique is proposed that uses the estimation of the power received from a node and its neighbors to make a local estimate of the position of the node to be located. This information is sent to the coordinating node that calculates the estimated position of the node from the reception of all the local processed sent by the anchored nodes, performing in the coordinating node a calculation that unites all the partial results in a single solution, called position Dear. To alleviate the problems described above, the intelligent location method and system on wireless sensor networks is presented, which is based on wireless sensor network techniques, seeking not to add hardware complexity to the device that increases its cost and weight. The main advantage of the present method with respect to the others that are proposed in the state of the art, is summarized, in the no need to have a pre-calibrated environment, being also a robust system against noise.

Unlike other proposed methods, in order to do this, no additional hardware is required beyond the RSSI meter, nowadays integrated into all commercial radio transceivers. This allows to design lightweight devices with limited autonomy, which results in a system applicable to all cases where autonomy and weight are important restrictions.

All this is achieved using computational intelligence techniques, based on a fuzzy inference engine and distributed processing that does not require large computing capabilities. In addition, the system allows monitoring of the environment and collecting information on possible sensors present in the mobile node, similar to what is usually done in wireless sensor networks.

The system and method of the invention allows the estimation of the position of mobile equipment, whose position is unknown from a series of equipment of known position.

It is part of a network of fixed position devices called anchored nodes. These positions are assumed to be known in advance and are static, and can be obtained at the time of installation by using some means of location, such as GPS technology, map plans, etc. The proposed location system It provides precision and robustness against noise, compared to the existing connectivity based location systems.

In order to estimate the position, it is necessary to measure the power of the electromagnetic field with which the message sent to the receiver arrives. This information can be acquired in two different ways: a) The remote node to be located sends a beacon: this situation will occur in those cases in which the mobile tag has additional sensors or information that it wishes to transmit to the coordinating node. In this case, the system will periodically send an empty message (without useful information), whose sole purpose is to be able to estimate the power with which this message arrives at the receiver. In order to save the battery, the tag only has the radio on long enough to send the beacon, not waiting for any type of acknowledgment.

The anchored nodes that receive the message, discard the data and measure the power with which the message arrives, since the information is irrelevant, not relaying it to the coordinating node.

The shipping period in this case is determined by the Tag, which can be a static or organized sampling rate using a calendar, which takes into account the needs of the application. For example, in wildlife monitoring, the system can reduce the sampling rate for as long as it is known in advance that the animal devotes to sleep, and therefore, is expected to remain for a prolonged period of time in a position. static b) The remote node to be located sends a message: this case will occur when the mobile node has additional sensors, whose measures are to be transmitted to the coordinating node. In this case, the operation of the system is identical to that described above, with the proviso that instead of sending an empty frame, it will contain useful information that the nodes near the Tag should reroute to the coordinating node. In order to save energy, the Tag will not wait for acknowledgment, turning off the radio immediately after sending the message. In this case, the sampling period of the messages will depend on the sensors of the system.

For the location it is necessary that the message sent to at least one anchored node of the system arrives, so the separation between them must be determined based on the accuracy that is desired in the location and the coverage of the mobile node. It is also necessary that the radio emission power be constant and known in advance.

It is not necessary for this system to perform a modeling of the environment, beyond making sure to separate the nodes at a distance that allows obtaining the precision required for the measurement. If a particular area is of special interest, to improve accuracy, you just have to increase the number of nodes anchored in that area.

More specifically, the method object of the present invention comprises the following steps:

- Stage 1: sending the beacon or the message by the mobile node. As mentioned above, this will be done periodically, depending on the needs of the system and the required autonomy. To save battery, the mobile node radio will only be on during this period of time and its resources will not be used to calculate the estimated position. - Stage 2: the anchored nodes receive the message and decide if the message corresponds to a beacon or if it contains useful information. In either case, the anchored nodes measure the field strength with which they received said message and relay it to their neighboring nodes.

- Stage 3: the nodes receive the power with which their neighbors have received the message. From these data and with the field strength measured by them, the nodes estimate a partial result of the position, using computational intelligence techniques, based on the use of a diffuse interference engine. Once the phases of the tensions in the substation have been estimated, and the phasors of the currents are known during the fault, already known and proven techniques are used to calculate the position of the fault.

- Stage 4: from the diffused interference motor output, the nodes that receive the Tag beacon estimate whether they have obtained a relevant result or not. If not, they discard the information and remain waiting for the reception of the next beacon. In the case of information, they send their partial estimate to the coordinating node, together with the information transmitted by the Tag sensors, in case it sends a message with data.

Stage 5: the coordinating node collects all partial estimates of the position and uses them to specify the estimated position, using the calculation of the center of gravity of all partial results. For a discrete set of solutions, as obtained in this case, the center of gravity coincides with the average of all the results received. In addition, in case of receiving data from the Tag sensors, it processes this information conveniently. Throughout the description and the claims the word "comprises" and its variants are not intended to exclude other technical characteristics, additives, components or steps. For those skilled in the art, other objects, advantages and features of the invention will be derived partly from the description and partly from the practice of the invention. The following examples and drawings are provided by way of illustration, and are not intended to be limiting of the present invention. In addition, the present invention covers all possible combinations of particular and preferred embodiments indicated herein.

BRIEF DESCRIPTION OF THE FIGURES

A series of drawings that help to better understand the invention and that expressly relate to an embodiment of said invention which is presented as a non-limiting example thereof is described very briefly below.

FIG. 1 schematically shows the different steps comprising the method object of the present invention in which FIG. 1A shows a beacon shipment by the non-anchored node; FIG. IB shows the retransmission of the power measured by the anchored nodes; FIG. 1C shows the estimation of the different partial results with the diffuse distributed method; and finally FIG. ID shows a determination of the estimated position in the coordinating node.

FIG. 2 shows a representation of the sectors formed by an anchored node and its neighbors.

FIG. 3 shows a representation of the distributed method.

FIG. 4 shows a representation of the fuzzy input sets. FIG. 5 shows a representation of the fuzzy output sets.

PREFERRED EMBODIMENT OF THE INVENTION

The intelligent location method on wireless sensor networks, object of the present invention comprises, in turn, two differentiated processes: (i) a first distributed process, executed in each node that estimates the possible location of the Tag; and (ii) a second centralized process, which unites all such partial information into a single solution.

In both cases, low computing power is required, and can be applied to large networks.

Therefore, the intelligent location method on wireless sensor networks, object of the present specification, is characterized in that it is sequenced in the following stages:

(i) A first stage of sending beacon by the Tag to be located, as shown in FIG. 1A; The unpinned Tag sends a message or a beacon, depending on the application you intend to develop. A message will be considered in the case of including useful information that has to be rerouted to the coordinating node. The emission power of the Tag must ensure that the communication reaches at least one fixed node. This message or beacon, will be sent cyclically by the unpinned node, this period being only during which it will have the radio activated, since this element will not participate in the computation process, in order to increase the battery life. The Tag can include systems to modify the sampling period according to the needs, either by means of a calendar or using external sensors, so that it is possible to increase even more battery life; subsequently, there is a reception of the beacon by the anchored nodes, the anchored nodes near the Tag that have received the beacon or the message, measure the level of RSSI with which it arrives, as well as the sensor information data, in case of existing; ii) A second stage of receiving the power received by the nodes anchored to their neighbors, as shown in FIG. 1B; Each anchored node that is close enough to be able to receive the message from the node to be located, sends a message to all its neighbors with the value of the power with which it has received the message. This causes the nodes around the node to be located to end a table with the strength of the signal they have received, as well as their neighbors. Neighboring nodes that have not received the "beacon" will not transmit any response; iii) A third stage of execution of a diffuse distributed location process, as shown in FIG. 1C; where from the power value received by the node and that which its neighbors have transmitted to it that have also received the message, the node executes a diffuse method, as shown in FIG. 3, to determine the most likely quadrants where the mobile node may be located.

The area enclosed by a particular node and two adjacent neighbors between them is considered quadrant. An example of this is shown in FIG. 2, where it can be seen that the quadrants or sectors considered always form triangular sections. The number of quadrants around a given node is variable and depends on the number of neighbors (that is, nodes with direct communication, without re-routing) that it has. It is important to keep in mind that each node has partial information, since it only has information on its estimate of the received power and that of its neighbors, but there may be other non-neighboring nodes that have also received the message, whose received power ignores this node

With said partial information, and from the execution of the fuzzy inference engine, the node determines the probable location quadrants, those that offer an output greater than 0.1 are considered probable. The diffuse method is always the same, and is executed once for each quadrant.

(iv) a fourth stage - of sending the proposed quadrants to the coordinating node, as shown in FIG. 1D, each of the anchored nodes that have received the message and that the diffuse method has given a significant output, they will send a message to the coordinating node indicating those quadrants for which they "vote", that is, they will inform those who have offered an inference motor output greater than 0.1. In addition, if the message contains useful information, it also sends it in the same frame to the coordinating node; where if none of the quadrants offer an output greater than 0.1, the anchored node, although it has received the message, will not "vote" for any quadrant, that is, it will not send any message; Y

(v) a fifth stage of calculating the estimated position with a centralized process, where from each of the partial results received from the different anchored nodes that have voted for a quadrant, the coordinating node prepares a map of points with all results, considering that the node is in the area delimited by all these quadrants. With this information, the system determines the estimated position from the centroid calculation of this entire area formed by the set of points for which the anchored nodes have voted. In this case, the centroid can be simplified by calculating the average of the partial results.

As you can see, the technique consists in the use of two differentiated processes, one of them being a diffuse distributed process, which is executed in each of the nodes that receive the "beacon", and subsequently, a centralized process carried out by the coordinating node that gathers all the partial results obtained by the different nodes.

A location method similar to the one proposed could have been performed using a single centralized method, but in this case, it would require the coordinating node to perform all calculations. In addition, it would present a much higher energy consumption due to the need to send all RSSI data received by all to the coordinating node.

This is why this division has been used in two methods, considering this system more efficient, both from the point of view of resources, since the system in this way performs a distributed computing, as well as from the energy point of view.

1. - Distributed diffuse process

FIG. 3 shows the fuzzy function for the location method, taking into account that this function should be executed in the speck once for each quadrant that is had, understood by quadrant, each of the areas formed by the node and two adjacent neighbors, that is, if there are four neighbors, In total, four quadrants are obtained around the anchored node, as shown in FIG. 2.

The diffuse output offers one result per quadrant, each node deciding from this information the "vote" that transmits to the coordinating node where it reports the representative quadrants found, those that offer an output greater than 0.1 being considered representative. This threshold value has been adjusted by simulation, being the best compromise found between precision and noise immunity.

To facilitate the explanation of the distributed method, it has been divided into several phases: a) explanation of the input assemblies to the diffuse inference engine; b) explanation of the output set of the diffuse inference motor; and c) explanation of the inference engine implemented.

1.1 Diffuse inference motor input assemblies

 The inputs of the fuzzy logic system are formed by an input that represents the power received by the node in question, plus a set of inputs that represent all its neighbors. Each of them has a blurred set consisting of three functions that represent the low, medium and high value of the defined power signal (as shown in FIG. 4).

It is understood as under a tropeizoidal set with membership function 1 for a power value equal to the sensitivity and falling linearly from this point, to the value defined as the center of the triangle of the average power. With this set, the received signal is represented if the anchored node is very far from the non-anchored one, therefore receiving a low or even null RSSI signal, as it would be received if it did not receive the beacon. It is defined as average power, a triangular set with the first vertex defined in the sensitivity, the central vertex (with membership value 1) to the power that would receive a node located in the center of the triangle formed by the three anchored nodes and the vertex on the right in the transmission power. This set represents the signal received at a strategic point, this point being the centroid formed by the three nodes that determine the quadrant. At this point, it is considered the midpoint, the Tag being defined if it is behind that point, and close otherwise.

Finally, a trapezoidal assembly is defined as high power that begins to literally grow from the point of the average power to the transmission power, having a membership value 1, for the transmission power. This set represents the signal that would be received if the non-anchored node is very close to the anchored node.

1.2 Diffuse inference motor output

 The output consists of three triangular assemblies, which correspond to the low, medium and high values, as shown in FIG. 5.

The triangle defined by points -0.5 and 0.5 is considered low, having a membership value of 1 at 0. The triangle defined between points 0 is considered medium; 0.5 and 1, having a membership value of 1 in 0.5. And finally, the triangle defined between points 0.5, 1 and 1.5 is considered high, having a membership value of 1 in 1.

The system will calculate an output per quadrant formed by the node in question, and their respective neighbors, making a transmission to the coordinating node if the output value is greater than 0.1. As can be seen in FIG. 5, if the value is only low, the centroid of the area corresponds to point 0, if it is the middle with point 0.5 and if it is high with point 1. 1.3 Inference engine

 The proposed inference engine corresponds to a knowledge base of Mandani rules, using an aggregation of the rules through the maximum function and an implication of the outputs through the minimum function. All this has been chosen according to computation efficiency criteria, since the processing capacity of the usual microcontrollers in the sensor networks is generally limited.

This knowledge base consists of evaluating each of the conditions of the rules, which are of the type: If A is ... and B is ... and ... THEN Exit is ...

The diffuse motor will calculate the antecedent of each rule through the intersection of the input sets, for which the minimum function for the operator "and" and the maximum for the function "o" have been used, which are easy to implement in devices With limited hardware.

With the calculated value of the antecedents with this method, the implication of the output is calculated using the minimum function. Subsequently, the union of all the rules is made and a specific value is calculated from said intersection. This process is called a constriction, having been used in this case a centroid concresor, which what it does is to offer as an output the center of gravity of the figure resulting from the union of the output of all the rules. Experimentally it has been proven that this concresor is the one that offers better results.

The inference engine consists of a total of eighteen rules, which must be executed once for each quadrant formed between the node and its neighbors, the rules can be summarized in the following table: RSSI of the neighbors RSSI node Exit

 Tall. All means. High.

 Low. All low. Low.

 Means, medium. All means. High.

 Means, medium. All low. Low.

 Tall. All tall. Half.

 Average in the current sector.

 Means, medium. High.

 Get off on the rest.

 High in some sector, except the current one.

 Means, medium. Low.

 Get off on the rest.

 High in a neighbor of the current sector.

 Tall. Means, medium.

 Get off on the rest.

 High in some node, except in the current sector.

 Tall. Low.

 Get off on the rest.

 Medium in some neighbor of the current sector.

 Means, medium. Means, medium.

 Get off on the rest.

 Medium in some node, except in the current sector.

 Means, medium. Low.

 Low on the rest.

2.- Centralized process.

The coordinating node is a special node of a sensor network, this element being a node that is responsible for receiving the data sent by all other nodes in the network and redirecting them through the appropriate network, acting as a gateway between the sensor networks , and an industrial PC or an embedded PC, which in turn will be responsible for distributing the information collected by other types of networks, such as Wi-Fi or Ethernet.

The node that fulfills the previous functions, and not the PC connected to it, will be considered as the coordinating node. Due to this, the proposed method will be implemented in a node with the same resources as the rest of the sensor nodes of the network, using the PC to transmit the obtained location results, as well as redirect the rest of the information of the present sensors In the net.

2.1 Stages of the centralized process

The coordinating node is a special node that has to perform the following actions to locate the non-anchored node. It is based on the fact that the coordinating node does not have access to a data table in which the distribution map of the anchored nodes is collected. It is also part of the fact that although its position is not known, the minimum period between messages sent by the Tag is known. The steps to be carried out are: a) Receiving data: from the message of the non-anchored node, and through the distributed method commenting above, some nodes of the network decide to vote for a location, for which they send to the coordinating node a data indicating by which of all the sectors formed between him, and his neighbors he decides to vote, in addition to the information collected from the non-anchored node, if he exists.

The coordinating node receives this message, and determines the coordinates of the nodes that make up the sector by which the node "votes" from reading the status of the bits that form the received word and knowing the identifier of the node that originated the message, since each of the sectors always forms a triangle. b) Determination of a representative point: once the previous step is completed, the coordinating node has the coordinates of the three vertices of each triangle. From them, it applies a method to determine a point that represents said triangle, which in this case has chosen to use the center of gravity.

Keep in mind that since the anchored nodes do not change their position, it is not necessary to continually recalculate these centers of gravity of all triangles, but it is enough to calculate them once (or have them previously calculated at the time of programming) and access its coordinates by accessing a table, saving resources in this way. c) Add the point to the voting table: the points obtained above are considered representative, so that in the event that the system only votes for a triangle, the center of gravity of that triangle will be considered the estimated position.

What the system does to be able to estimate the position is to add the points obtained to a table in which the coordinates of all the votes that decide to transmit the anchored nodes will be stored. d) Waiting for a specific time for the voting of the nodes: because the time between tag beacons is known in advance, the coordinating node establishes a time window from that period during which it will wait for the reception of the votes of the nodes.

Keep in mind that a node can vote for several regions, but can only vote once. In case several messages arrive from the same node in the same window, only the first one will be considered and all others will be rejected. e) Calculation of the estimated position: once the voting time window is closed, the coordinating node accesses the table with all the votes and specifies them in the estimated position.

For this purpose, the use of the method of calculation of the center of gravity has been chosen, in which the centers of gravity of each of the sectors that the coordinating node has received as partial network results are considered representative positions of sensors.

Once the estimated coordinates are obtained, they are transmitted from the coordinating node to the PC, and the node is waiting for the time to open a new window to be fulfilled.

Claims

1.- Intelligent location method on wireless sensor networks comprising a first distributed process, executed on each node that estimates the possible location of the Tag; and a second centralized process, which unites all these partial information in a single solution, where said method is characterized in that it comprises, in turn, the stages of:  (i) a first stage of sending a beacon by the Tag to be located, where the non-anchored Tag sends a message or a beacon; said message or beacon being cyclically sent to the node not anchored, this period being only during which it will have the radio activated;  (ii) a second stage of receiving the power received by the nodes anchored to their neighbors; where each anchored node that is close enough to be able to receive the message from the node to be located, sends a message to all its neighbors with the value of the power with which it has received the message; in such a way that the nodes around the node to be located complete a table with the strength of the signal they have received, as well as their neighbors; and where neighboring nodes that have not received the "beacon" will not transmit any response;  (iii) a third stage of execution of a diffuse distributed location process, where from the value of power received by the node and that transmitted by its neighbors who have also received the message, the node executes a diffuse process to determine the most probable quadrants where the mobile node may be located; (iv) a fourth stage of sending the proposed quadrants to the coordinating node, where each of the anchored nodes that have received the message and that the method diffused has given a significant output, they will send a message to the coordinating node indicating those quadrants for which they "vote", that is, they will report those that have offered an inference motor output greater than 0.1; Y  (v) a fifth stage of calculating the estimated position with a centralized process, where from each of the partial results received from the different anchored nodes that have voted for a quadrant, the coordinating node prepares a map of points with all results, considering that the node is in the area delimited by all these quadrants. 2. - Method according to claim 1 wherein in the first stage a message or beacon will be sent depending on the application that is intended to be developed, where a message will be considered in the case of including useful information to be rerouted to the coordinating node , and where the tag's emission power has to ensure that the communication reaches at least one fixed node. 3. - Method according to claims 1 and 2 wherein the Tag includes means to modify the sampling period according to the needs, either through a calendar or using external sensors. 4. - Method according to the preceding claims wherein in the reception of the beacon by the anchored nodes, said anchored nodes close to the Tag that have received the beacon or the message, measure the level of RSSI with which it arrives, as well as the information data of the sensors, if any. 5. - Method according to the preceding claims wherein in the fourth stage of sending the quadrants, in case the message contained information useful, it also sends it in the same frame to the coordinating node; and where if none of the quadrants offer an output greater than 0.1, the anchored node, even if it has received the message, will not "vote" for any quadrant, that is, it will not send any messages. 6. - Method according to the preceding claims wherein in the fifth stage of calculation, with the information received, the estimated position is determined from the centroid calculation of this entire area formed by the set of points for which the voters have voted. anchored nodes; and where the centroid can be simplified by calculating the average of the partial results. 7. - Method according to the preceding claims wherein the diffuse distributed process comprises the steps of a) entering the data set into the diffuse inference engine; b) output of the fuzzy inference engine data set; and c) blurred inference of the data received, where,  The inputs of the fuzzy logic system consist of an input that represents the power received by the node in question, plus a set of inputs representing all its neighbors; and where each of them has a blurred set consisting of three functions that represent the low, medium and high value of the defined power signal;  and where the output is formed by three triangular assemblies, which correspond to the low, medium and high values, calculating an output per quadrant formed by the node in question, and their respective neighbors, making a transmission to the coordinating node if the value of output is greater than 0.1; Y where the inference engine corresponds to a knowledge base of Mandani rules, which will calculate the antecedent of each rule through the intersection of the input sets, for which the minimum function has been used for the operator "y" and the maximum for the function "o", and where subsequently, the union of all the rules is made and a specific value is calculated from said intersection, offering as an output the center of gravity of the figure resulting from the union of the output of all the rules. 8. Method according to the preceding claims wherein the centralized process of the coordinating node comprises the steps of:  a) data reception, where from the message of the non-anchored node, and through the distributed process, some nodes of the network decide to vote for a location, for which they send to the coordinating node a data indicating which of all sectors formed between him, and his neighbors decides to vote, in addition to the information collected from the non-anchored node, if it exists; and where the coordinating node receives this message, and determines the coordinates of the nodes that form the sector by which the node "votes" from reading the state of the bits that form the received word and knowing the identifier of the node that originated the message, since each of the sectors always formed a triangle;  b) determining a representative point, where, once the previous step is completed, the coordinating node has the coordinates of the three vertices of each triangle; and where from them, a process is applied to determine a point that represents said triangle, which in this case has chosen to use the center of gravity; c) add the point to the voting table, and where the points obtained above are considered representative, so that in the case that the system only votes for a triangle, the center of gravity of said triangle will be considered the estimated position ; and where to be able to estimate the position is to add the points obtained to a table in which the coordinates of all the votes that decide to transmit the anchored nodes will be stored; d) wait for a specific time for the voting of the nodes, where due to the fact that the time between beacons of the Tag is known in advance, the coordinating node establishes from that period a time window during which it will wait for the reception of the votes of the nodes; and where a node can vote for several regions, but can only vote once; and where in case several messages arrive from the same node in the same window, only the first one will be considered and all others will be rejected; Y e) calculation of the estimated position, where, once the window of ^' voting time is closed, the coordinating node accesses the table with all the votes and specifies them in the estimated position;  and where once the estimated coordinates are obtained, they are transmitted from the coordinating node to the PC, and the node is waiting for the time to open a new window to be fulfilled. 9.- Intelligent location system on wireless sensor networks comprising a plurality of wireless sensors denoted as distributed nodes and at least one coordinating node, this element being a node that is responsible for receiving the data sent by all other nodes of the network and redirect them through the appropriate network, acting as a gateway between the sensor networks, and an external computing element, which in turn will be responsible for distributing the information collected by other types of networks, such as Wi-Fi or Ethernet, where said system is characterized in that each node comprises means for executing the method of claim 1.