CN101865993B - Target tracking method based on binary sensor Ad Hoc network - Google Patents

Abstract

The invention discloses a target tracking method based on a binary sensor Ad Hoc network. The target tracking method adopted by the invention comprises the following steps that: firstly, a sensor network system is deployed in site for carrying out network initialization and information updating; then, a sensor network node judges whether to trigger the position calculation or not; if the triggering conditions are satisfied, the node calculates the target position in the local place, and sends back the calculation results to a gateway gathering node of the sensor network system through a shortest route; and finally, a computer in the monitoring management center is responsible for collecting target tracking results of each node for forming the moving track of the tracked target. The method is characterized in that a vibration sensor uses an Ad Hoc network mode for forming a network system, realizes the optimization of two indexes: the target tracking precision and the calculation cost, the problem of large calculation amount is avoided, and in addition, the positioning precision is high. The method is also applicable to the network system formed by other binary detection sensors deployed outdoors, and realizes the tracking on moving vehicles or personnel.

Description

A kind of method for tracking target based on binary sensor Ad Hoc network

Technical field

The invention belongs to sensor network target and follow the tracks of the field; Relate to a kind of method that in the wireless sensor network system of self-organization, realizes motion target tracking; Relate in particular at the sensor network of arranging and have under the condition of two-value detection feature; For example adopt the earthquake motion miniature acceleration sensor, calculate the method for finding the solution the moving target position coordinates.

Background technology

Three big bases of modern information technologies are sensor technology, the communication technology and computer technology; They accomplish collection, transmission and the processing of information respectively; " sense organ " of configuration information system, " nerve " and " brain ", the combination of these three kinds of technology has promoted the process of IT application.The wireless sensor network technology typical case of this combination just embodies, and belongs to the dual-use strategic hi-tech of multidisciplinary intersection.

Self-organizing radio sensor network is the very popular field of research and development both at home and abroad in recent years, obtains attention and application highly in a plurality of industries and field at present.Target following is one of attractive and representational application of sensor network, in the military and civilian field wide application prospect is arranged.The optimal design of tracker is intended to make full use of the sensor network Limited resources, the Practical Performance of optimization system.

Sensor network has the function of microminiaturized end-probing and the characteristics of MANET, thereby more and more obvious in the target following application advantage, mainly comprise following some:

(1) follow the tracks of meticulous: the microminiaturized sensor node of intensive deployment can be surveyed moving target accurately, Position Tracking and control, can demonstrate the motion conditions of moving target in detail.

(2) follow the tracks of reliable: because the self-organization of wireless sensor network, when node failure or new node add fashionable, can configuration automatically in rugged environment, make sensor network when tracking target, have higher reliability, fault-tolerance and robustness.

(3) follow the tracks of in time: the collaboration type monitoring of multiple sensors makes that the discovery of moving target is more timely, realizes collaborative work between distributed data processing and the multiple heterogeneous sensor node easily, makes that the process of target following is more comprehensive.

(4) tracking is hidden: because the sensor node volume is little, wireless transmitted power is low, is difficult for being found by adverse party, thereby can the more hidden tracking of realization of goal be also made things convenient for to dispose simultaneously and use, and is applicable to the purpose of battlefield surroundings and military combat.

(5) low cost: the microsensor cost is low, can reduce the cost and the cost of whole tracing process.

(6) low-power consumption: the design of wireless sensor network is main target with the low-power consumption, and working in the open air etc. does not have fixed power source or change under the situation of battery inconvenience, make follow the tracks of use more convenient.

Based on the maneuvering target tracking position fixing process of sensor network, come down to a kind of computation optimization process that receives the constraint of target motion track, no matter be at military field or civil area, all be with a wide range of applications.

The self-organizing sensor network network has many advantages on the one hand when being applied to Target Tracking System, also brought some challenging technical matterss on the other hand.For example, wireless sensor network does not have center control machine system usually, does not have backbone network can supply scheduling and commander, that is to say the suitable distributed track algorithm that adopts.In addition, the power consumption of sensor network nodes requires low, the unsuitable computing method that adopt high complexity.

In current sensor network, a kind of interesting type is exactly the sensor network with two-value detection feature.This sensor network has application advantage preferably aspect target following, received the favor of academia and engineering department.

Two-value (Binary) sensor network is meant that the sensor network terminal node can be converted into the reading of target detection information to find target whether discrete value (0 or 1).For example, microseismic activity acceleration transducer, acoustic sensor, magnetic sensor etc. all have this two-value detection feature.Single-sensor with two-value detection feature has finite information for dbjective state, need realize the target following of collaboration type by networking technology.

In the prior art, mainly contain following several kinds of method for tracking target based on binary sensor Ad Hoc network:

(1) document " T.Jing, S.Hichem, and R.Cedric.Binary variational filtering for targettracking in sensor networks.Proceedings of 14th IEEE/SP Workshop on Statistical SignalProcessing; Madison, USA, August 2007; pp.685-689 (Chinese translation: T. respects, S. Hai Qimeng, R. Cedric. the two-value variation that sensor network target is followed the tracks of is filtered; the 14th IEEE/SP statistical signal handled Conference Papers collection, U.S.'s Madison, in August, 2007; page number: 685-689) " and document " P.M.Djuric, M.Vemula, and M.F.Bugallo.Signal processing by particle filtering for binary sensor networks.Proceedings of the 11th DigitalSignal Processing Workshop; New Mexico, USA, August 2004; Pp.263-267 (Chinese translation: P.M. enlightening turtledove Er Ruike, M.F. cloth loud, high-pitched sound Shandong. particulate filter is used for the signal Processing of binary sensor network. the 11st digital signal processing symposial, U.S. New Mexico; In August, 2004; The page number: 263-267) " in, propose the two-value detection information is transferred to gateway node, calculate the movement locus of target by gateway node operation population Filtering Model.Population filter tracking algorithm proposes on the maximal possibility estimation basis, but it is too expensive from the communication cost that each node is transmitted to gateway node with data.

(2) document " K.Mechitov, S.Sundresh, Y.Kwon; and G.Agha.Cooperative tracking withbinary-detection sensor networks.Tech.Rep.UIUCDCSR-2003-2379, University of Illinois atUrbana-Champaign, September 2003 (Chinese translation: K. Mo Qituwei; S. Sang Deruishi; Y. section is prosperous, G. A Ge. two-value acquisition sensor network is used for collaboration type to be followed the tracks of. the E Benna of University of Illinois-champagne branch school technical report is numbered UIUCDCSR-2003-2379; in September, 2003) " in; If the motor message that proposes target causes the sense events of sensor, then weighted mean is carried out in the position of these sensors, estimate each point coordinate on the target trajectory.

Document " W.Kim, K.Mechitov, J.-Y.Choi; and S.Ham.On target tracking with binaryproximity sensors.Proceedings of the 4th international conference on Embedded networkedsensor systems, Boulder, Colorado; USA, 2006, pp.301-308 (Chinese translation: W. Kern; K. Mo Qituwei, Cui J.Y., Korea Spro S.. approximate binary sensor is used for target following. the 4th embedded type networking sensing system international conference collection of thesis; the Colorado Bohr get,, the page number: 301-308) in 2006 " in; Proposition improves the weight computation method of this induction pick-up position through the movement velocity of estimating target.But the method in these two pieces of documents needs hypothetical target on rectilinear orbit, at the uniform velocity to go; And the position of only having adopted the sensor node that can produce the echo signal induction, in fact the target arrival event also can be improved the precision of target trajectory through the cooperation mode in the network range.

Because the binary sensor network objectives track algorithm that these two pieces of documents propose mainly is the foundation position of detecting sensor node, carries out the position weighted mean and calculates coordinates of targets, these class methods is referred to as method of weighted mean here.Because the calculation cost of these class methods and method provided by the invention is all less; Thereby they have the comparability on the performance; Performance index mainly are the errors that calculate the target location, and follow-up checking embodiment has shown that method provided by the invention has advantage aspect error criterion.

(3) document " N.Shrivastava, R.Mudumbai, U.Madhow; and S.Suri.Target tracking withbinary proximity sensors:Fundamental limits, minimal descriptions, and algorithms.Proceedings of the 4th international conference on Embedded networked sensor systems; Boulder, Colorado, USA; 2006, pp.251-264 (Chinese translation: N. Si Rui Wasita, R. Mu Dubai; U. Mai Dehaou; S. Rui that revives. approximate binary sensor is used for target following: basic restrictive condition, minimize and describe and algorithm. the 4th embedded type networking sensing system international conference collection of thesis, the Colorado Bohr get, 2006 years; The page number: 251-264) " in; propose the information of target through the sensor investigative range is used for constructing regional area, if target is about to walk these zones, regional extent constitutes round intersection arc by the detection border of interdependent node and confirms.According to the sequence of target passing, the minimum straight line clip path that target is walked All Ranges is defined as the movement locus of target.Yet this method belongs to centralized solution, and computation complexity is high, is not real time execution in addition, exist acquisition of information and fused data than long time delay.

Therefore, the method for tracking target that the designing and calculating amount is little, bearing accuracy is high, significant for the application of self-organizing sensor network network.

Summary of the invention

The object of the present invention is to provide a kind of method that under the binary sensor Ad Hoc network system, realizes target following.Theoretical according to computational geometry, based on the position that three circle Intersection set method are calculated tracked target, make computation amount, bearing accuracy also has raising by a relatively large margin simultaneously.

So-called two-value (Binary) surveys to be meant sensor reading to be converted into discovery target whether discrete value (0 or 1), for example common miniature ground microseismic activity acceleration transducer, acoustic sensor, magnetic sensor etc. all have the physical attribute that this two-value is surveyed.Constitute simply and the sensor network system of practicability through this two-value acquisition sensor, realize the function of target following.Single microsensor with two-value detection feature has finite information for dbjective state, thereby need realize the target following of collaboration type by networking technology.

A lot of sensors all have the attribute that this two-value is surveyed, and when moving target during proximity transducer, can produce tangible signal and change in the certain distance scope, show the existence and the appearance of target, threshold value is set eliminate the signal noise that possibly exist.Utilize the binary sensor network to realize the tracking to moving target, have lot of advantages, Practical Performance is good.When target signal to noise ratio is big, the good reliability that two-value is surveyed, designed system is stable, reliable, can reduce the false dismissed rate and the false alarm rate of Target Tracking System.

If any point in the sensor network monitoring zone can both be monitored by k sensor node, claim that network is that k-covers, the coverage of perhaps claiming network is k.Target Tracking System is system one to one normally, and promptly a detector only aims at and follow the tracks of a target continuously.If realization is carried out accurate tracking and location to target, must realize that node is covered by 3 degree, promptly exists three sensor probe nodes to survey simultaneously target at least, otherwise only can play the function of discovery and early warning for monitoring objective.

The invention is characterized in: to sensor network with two-value detection feature; Design the distributed method for tracking target of a kind of cooperation based on three circle Intersection set method; Survey the intermediate points of the public intersection segmental arc of circle with three of compact way calculating sensor network nodes, with this as coordinates of targets.The characteristics of this method have been to realize the balance optimizing of precision and two kinds of indexs of calculation cost of target following, and not only calculated amount is little, is applicable to the sensor network of low-power consumption, and bearing accuracy is high.

Method content of the present invention realizes on the node of the microseismic activity acceleration transducer with ZigBee wireless self-networking function.The architecture of sensor network system comprises distributed sensor node, aggregation node and task management background program.Sensor node has the dual-use function of the network terminal and route: realize the collection and the processing of data on the one hand; Realize the fusion and the route of data on the other hand, data of itself gathering and the data that other node of receiving sends are carried out comprehensively being transmitted to gateway node.Sensor network nodes is made up of processor unit, wireless transmission unit, sensor unit and PMU four parts.

Condition and corresponding measure that method provided by the invention is required are following:

(1) sensor node that is used for the pursuit movement target have and neighbor node between carry out wireless communication ability, wireless communication ability is the radio function that common MANET sensor node all has.

Adopt the CC2420 wireless communication chips, operation has the ZigBee agreement of short-range communications functionality, accomplishes to organize themselves between the network node net.The node wireless transmission unit of design mainly provides physical layer data service and physical layer management service; Physical layer data service transceive data from the wireless channel; Management service is safeguarded the database that a Physical layer related data is formed, and comprises following service: activate at any time and the dormancy wireless transmit/receive units; Energy to wireless channel detects, and selects foundation for network layer provides channel; Transceive data.

(2) need sensor can judge that a moving target is whether within its maximum detectable range scope.If the detection perception radius of sensor is R, when the distance of target and sensor during less than R, sensor reading is that target is found in " 1 " expression; When the distance of target and sensor during greater than R, sensor reading is not found target for " 0 " expression.

Adopt ADXL202 microseismic activity sensor, the output digital signal, its pulse duty cycle is directly proportional with two acceleration that sensing shaft is experienced separately.Acceleration transducer ADXL202 is the bidimensional digital acceleration sensor of AD company, and working temperature is-40 ℃～85 ℃, adopts advanced MEMS technology, measures oscillatory acceleration and static acceleration.The accekeration that calculates according to the dutycycle and the resistor values of this sensor; And get two on axially acceleration and numerical value as final useful signal data; Can design and judge the size of the vibration signal value of target, distinguish target and whether be positioned within its maximum range.

(3) exist three sensor probe nodes to survey simultaneously target at least, promptly the coverage in sensor network monitoring zone is more than or equal to 3.

(4) sensor network nodes has been accomplished the location through manual type or GPS technology, i.e. self coordinate figure of known sensor network node.

Effect of the present invention is: this target following distributed method avoids directly finding the solution the big problem of Floating-point Computation amount of binary quadratic equation; Be applicable to the sensor network nodes of low-power consumption; It is reducing manyly than traditional method of weighted mean aspect the error in addition, and the precision of setting movement target can satisfy the demand of practicability.Therefore, it realizes balance optimizing aspect the precision of target localization and two kinds of indexs of calculation cost.This method is applicable to the sensor network system of outdoor deployment, is specially adapted to the target monitoring task in key road segment or the zone, can realize the tracking to moving vehicle or personnel.

Accompanying drawing 1 is to adopt three circle Intersection set method to realize the principle schematic of target following.Particularly, method for tracking target provided by the invention comprises following the implementation phase and step successively:

The deployment of step 1. sensor network system, initialization and information updating

Dispose whole network system through the mode that manual work is buried underground or machine dispenses, information is converged to the computing machine of Surveillance center by gateway node.The architecture of each network node comprises sensor assembly, processor module, radio receiving transmitting module and energy supply module four parts composition.Microprocessor adopts C8051F121, and wireless communication chips adopts CC2420, and sensor chip adopts ADXL202 microseismic activity sensor, adopts powered battery.The carrier frequency of CC2420 is 2.4GHz, and each network node utilizes carrier wave detection multiple access/collision to avoid (CSMA/CA) avoidance mechanism to obtain wireless channel.

After system deployment, carry out netinit; Each network node makes up its neighbor list and to the coordinate of neighbors circular self; Neighbor list information comprises that 1 bit status Information sign of the target detection incident that neighbors numbering (ID), neighbors coordinate, neighbors produce is (with " 1/0 " expression; Target is found in " 1 " expression, and target is not found in " 0 " expression).The coordinate of each network node is participated in follow-up calculating.Each network node periodically receives and the target detection information of upgrading the neighbors that it preserves, simultaneously also with the own information announcement neighbors of finding target, with the calculating of cooperation mode completion target location in this locality.

The terminal node of step 2. sensor network judges whether to trigger the behavior that calculate the target location

The sampling rate that sensor is got 90sps in Design of System Software realizes the monitoring to target.According to the signal of gathering, each network node judges whether to trigger the position calculation of tracked target.If satisfy following a kind of condition, then produce and trigger calculating operation:

1) target of existing two neighbors arrival Status Flag is " 1 " in the neighbor list of each network node, and target is just getting into the investigative range of this network node this moment;

2) when target gets into the investigative range of each network node; In the neighbor list of this network node, existed target arrival Status Flag to surpass 2 for the neighbors quantity of " 1 ", this moment is according to the ID numbering of neighbors; By ascending order; Select two neighbors, as the basis that three circles occur simultaneously and calculate, promptly the coordinate of these two neighbors is participated in follow-up calculating.

Calculate step 3. target location

If each network node satisfies one of above-mentioned trigger condition, then calculate the position of target according to three circle Intersection set method.The process of three circle Intersection set method comprises: when target gets into the investigative range of three sensor nodes successively, calculate the 3rd sensor node shown in the accompanying drawing 1 earlier and survey circle S ₃Survey circle S with first sensor respectively ₁, second sensor survey circle S ₂4 intersecting point coordinates; Whether the distance of judging these 4 intersection points and three centers of circle then respectively is equal to or less than the detection radius of circle; It is to be equal to or less than radius of circle that 2 intersection points are wherein arranged; These 2 intersection points are selected as two the end points a and the b of the common intersecting arcs of three circles, calculate the intermediate point of the line segment that these two end points form at last, with this as target in this residing estimated position constantly.

Gathering of step 4. target following information

Trigger the result of calculation of the sensor node of target location calculating with step 3; Through shortest path by the gateway aggregation node that sends it back sensor network; And be transferred to the computing machine of Surveillance center, and gather the position calculation result of target following, form the movement locus of tracked target.

Native system adopts the graphical interfaces handling procedure on the Visual Basic design PC.Computing machine adopts the baud rate of serial ports and 11520Baud/s and network node to transmit data, with the trace of the mode display machines moving-target of figure.In addition, the host computer interface software is preserved raw data with the form of database file on the backstage, supplies to analyze to use.

Motion target tracking method provided by the invention is characterized in that, in the network system that microsensor is formed; Realize the precision of target following and the balance optimizing of two kinds of indexs of calculation cost; Not only calculated amount is little, and bearing accuracy is high, has reduced error significantly than classic method.

Below in conjunction with accompanying drawing further explain specific embodiment of the present invention.

Description of drawings

Fig. 1 is that binary sensor Ad Hoc network adopts three circle Intersection set method to realize the principle schematic of target following;

Fig. 2 is the deployment design diagram of sensor network target tracker;

Fig. 3 is the statistics synoptic diagram that the binary sensor network system realizes target following;

Fig. 4 is based on the target following pilot system of binary sensor Ad Hoc network and disposes embodiment.

Embodiment

Be described in greater detail with reference to the attached drawings some embodiment of the present invention.

Accompanying drawing 1 is the exemplary plot of carrying out target following location according to sensor two-value apperceive characteristic, and localized network is by with S ₁, S ₂, S ₃Form for three sensor nodes in the center of circle, their search coverage is respectively X, Y and Z, supposes that they have unified radius of investigation.When by the target of being followed the tracks of along the arrow curve when the monitored area is walked, A ₁, A ₂, A ₃, A ₄, A ₅, A ₆Represent 6 ad-hoc locations of target in motion process respectively, promptly target gets into/withdraws from the investigative range of a certain sensor node.

Here at t ₁, t ₂, t ₃Constantly, target gets into sensor node S successively ₁, S ₂, S ₃Investigative range, trigger the signal that the sensor export target arrives, promptly be masked as " 1 "; At t ₄, t ₅, t ₆Constantly, when target separating sensor S successively ₁, S ₂, S ₃Investigative range the time, trigger the signal that the sensor export target is lost, promptly be masked as " 0 ".

In case when target got into the Z zone, it just met the condition that 3-covers, and can realize the estimation of target location.Because being sensor network k-, accompanying drawing 1 covers the example fragment in the monitored area, in fact in the target continuous tracing process, if target gets into X zone, the A in the trigger point ₁The place also can come the current location of target is calculated according to the neighbors information of front.

Here still with target get into the Z zone, the A in the trigger point ₃The due in t at place ₃Be example, set forth three circle Intersection set method and realize the target location calculating principle.

Only need one to jump and can carry out radio communication with neighbor node if the communication radius of sensor network nodes, guarantees node much larger than radius of investigation R, this normally satisfies in practical application.In native system; The wireless communication distance of CC2420 chip is about 100 meters; Although the maximum range of microseismic activity sensors A DXL202 can reach 300 meters, effective detection range under normal conditions is in tens meters, thereby can satisfy the condition of target following.

Find out round S by accompanying drawing 1 ₁, S ₂, S ₃Common factor be circular arc ab, the some a with the some b be its two end points, this is the public boundary of three sensor search coverages.If the central point of arc ab is a c, promptly the pentagram mark (★) in the accompanying drawing 1 with the coordinate of this point as the target that is positioned, is remembered and is made (x _T, y _T).Key is to obtain through compact way the coordinate of this point, need not the time synchronized of whole network, makes paying under the very little calculation cost condition, obtains high orientation precision, is key point of the present invention.

The coordinate of solution point c is through calculating the end points a of arc ab, the position of b, known three sensor node S earlier ₁, S ₂, S ₃Coordinate be respectively (x ₁, y ₁), (x ₂, y ₂), (x ₃, y ₃), the coordinate of two-end-point is remembered work (x respectively _a, y _a), (x _b, y _b), then draw following simultaneous equations according to geometric relationship:

$\left\{\begin{array}{c}{(x_a-x_2)}^2+{(y_a-y_2)}^2=R^2\\ {(x_a-x_3)}^2+{(y_a-y_3)}^2=R^2\\ {(x_a-x_1)}^2+{(y_a-y_1)}^2\≤R^2\end{array}\right.---\left(1\right)$

$\left\{\begin{array}{c}{(x_b-x_1)}^2+{(y_b-y_1)}^2=R^2\\ {(x_b-x_3)}^2+{(y_b-y_3)}^2=R^2\\ {(x_b-x_2)}^2+{(y_b-y_2)}^2\≤R^2\end{array}\right.---\left(2\right)$

Directly accurately find the solution above-mentioned binary quadratic equation group (1) and (2), their Floating-point Computation amount can be very big, and calculation cost is high, on the embedded system of low-power consumption, implements very difficulty.Here adopt a kind of three circle common factor diagnostic methods to calculate the coordinate of end points a and b, only need calculated amount seldom.

Usually the coordinate that calculates the central point c of circular arc ab relatively takes twists and turns, considers that the error that causes is little, chooses the intermediate point with the line segment of end points a and b line here, and approximate getting made the c point coordinate, and with this estimated position as tracked target.Therefore, coordinates of targets (x _T, y _T) draw by computes:

$\left\{\begin{array}{c}{x}_T=\frac{(x_a+x_b)}{2}\\ y_T=\frac{(y_a+y_b)}{2}\end{array}\right.---\left(3\right)$

If network coverage degree greater than 3, promptly can be surveyed by the sensor node more than 3 in the regional area internal object motion of network and find, then set the 3rd node according to finding that preceding two sensor nodes in the sequence come calculated target positions.Although consider that the sensor node number in the target detection range is many more, then the target location estimation for accuracy can be high more, can bring surge and the complicacy of processing on the calculated amount simultaneously.Therefore, based target finds that three circles in the formation come calculated target positions, makes tracking have better terseness, and bearing accuracy also can satisfy practical requirement simultaneously.

Accompanying drawing 2 is deployment design diagrams of sensor network target tracker.Sensor node constitutes wireless network architecture through the self-organization mode, realizes the collection and the processing of network's coverage area internal object trace information with cooperation mode.In the monitored area of sensor network, sensor node carries out position calculation to the target of being followed the tracks of, and the result is reported to the gateway aggregation node, finally is responsible for calculating the entire motion track that provides target by the computing machine of supervision and management center.

The step embodiment that objectives are followed the tracks of is following:

The deployment of step 1. sensor network system, initialization and information updating.

Dispose whole network system through the mode that manual work is buried underground or machine dispenses, information is converged to the computing machine of Surveillance center by gateway node.The architecture of node comprises sensor assembly, processor module, radio receiving transmitting module and energy supply module four parts composition.The micro OS software design of network terminal node comprises functional requirement analysis, task division and priority, adopts the suitable communication and the method for synchronization between the task.Microprocessor adopts C8051F121, and wireless communication chips adopts CC2420, and node utilizes carrier wave detection multiple access/collision to avoid (CSMA/CA) avoidance mechanism to obtain channel.

In initialization procedure, each network node makes up its neighbor list and to the coordinate of neighbors circular self, list information comprises 1 bit status Information sign (1/0) of the target detection incident that neighbors numbering ID, neighbors coordinate, neighbors produce.This node coordinate is as (x ₃, y ₃), participate in follow-up calculating.Node periodically receives and the target detection status information of upgrading the neighbors that it preserves, simultaneously also with the own status information announcement neighbors of finding target, with the calculating of cooperation mode completion target location in this locality.

Step 2. node judges whether to trigger the position calculation of tracked target, and trigger condition is following:

1) target of existing two neighbors arrival Status Flag is " 1 " in the neighbor list information of node, and target is just getting into the investigative range of this node this moment;

2) when target gets into the investigative range of this node; Existed in the neighbor list information target arrive Status Flag for the neighbors quantity of " 1 " above 2; Number according to neighbors ID this moment; Therefrom select two neighbors by ascending order, as the basis of three circle common factor calculating, i.e. their coordinate (x ₁, y ₁), (x ₂, y ₂) the follow-up calculating of participation.

In system design; The slewing rate of ADC is very crucial, possibly catch the vibration signal less than maneuvering targets such as vehicles too slowly, can strengthen the load of system too soon again; Waste sensor node originally with regard to limited energy, thereby will be through calculating the balanced sample rate of a centre.Get a median speed 100km/h, i.e. the vehicle 27.8m that can advance in 1s surveys three points and calculates by the vehicle 1m that advances, and the sampling rate that draws system is 83.4.Native system is got 90sps in software design sampling rate realizes.

If step 3. satisfies trigger condition, node comes calculated target positions with distributed way according to three circle Intersection set method in this locality, promptly calculates the round S shown in the accompanying drawing 1 earlier ₃Respectively with circle S ₁, the circle S ₂4 intersecting point coordinates, then according to formula (1) and formula (2), judge respectively whether the distance in intersection point and three centers of circle is equal to or less than radius of circle, therefrom select at last two end points of the common intersecting arcs of three circles, as the coordinate of an a with some b.

The concrete computation process of this step is following:

1) defines following variable symbol: L, K ₁, K ₂, X ^*, Y ^*, D, x ', y ', x ", y ",

Order $L=\sqrt{{(x_3-x_1)}^2+{(y_3-y_1)}^2},$

${\mathrm{<figure-callout\; id="1"\; label="K"\; filenames="H200910185077XE00011.png,H200910185077XE00021.png"\; state="\{\{state\}\}">K</figure-callout>}}_1=\frac{(y_3-y_1)}{(x_3-x_1)},$ ${\mathrm{<figure-callout\; id="2"\; label="K"\; filenames="H200910185077XE00011.png"\; state="\{\{state\}\}">K</figure-callout>}}_2=-\frac{1}{{\mathrm{<figure-callout\; id="1"\; label="K"\; filenames="H200910185077XE00011.png,H200910185077XE00021.png"\; state="\{\{state\}\}">K</figure-callout>}}_1},$

$\left\{\begin{array}{c}{X}^{*}=x_1+\frac{(x_3-x_1)}{2}\\ Y^{*}={\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="H200910185077XE00011.png,H200910185077XE00021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1+K_1(X^{*}-x_1)\end{array}\right.,$

D=R then ²-(X ^*-x ₁) ²-(Y ^*-y ₁) ²

2) the intersecting point coordinates note that intersects of two circles do (x ', y '), (x ", y "), calculate as follows:

$\left\{\begin{array}{c}{x}^{\&prime;}=X^{*}+\sqrt{\frac{D}{{\left({1+K}_2\right)}^2}}\\ y^{\&prime;}=Y^{*}+K_2(x^{\&prime;}-X^{*})\end{array}\right.$

$\left\{\begin{array}{c}{x}^{\&prime;\&prime;}=X^{*}+\sqrt{\frac{D}{{\left({1+K}_2\right)}^2}}\\ y^{\&prime;\&prime;}=Y^{*}+K_2(x^{\&prime;\&prime;}-X^{*})\end{array}\right.$

3) according to the decision condition of formula (1) and formula (2), judge respectively intersection point (x ', y '), (x ", whether y ") is equal to or less than radius of circle with the distance in three centers of circle, selects one of wherein eligible person, as end of arc of three circle common factors, like a point (x _a, y _a).

4) with x ₂, y ₂Replacement process 1) x that begins to locate ₁, y ₁, repeat the aforementioned calculation flow process, obtain another end of arc that three circles occur simultaneously, like b point (x _b, y _b).

5) average a point and the b coordinate of ordering is as the approximate estimation of coordinates of targets, shown in (3).

The node that calculate step 4. triggering target location is the result of calculation of step 3, through shortest path by the gateway aggregation node that sends it back sensor network.The main frame of supervision and management center is responsible for collecting the target following result of each node, forms the movement locus of tracked target.

For the network system host computer, native system adopts the graphical interfaces handling procedure on the Visual Basic design PC.Computing machine adopts the baud rate and network node transmission data of serial ports and 11520Baud/s, and is with the trace of the mode display machines moving-target of figure, vivid and intuitively.In addition, upper computer software is preserved raw data with the form of database file on the backstage, supplies to analyze to use.

Performance for practice examining method provided by the present invention; We have designed the shock sensor node with ZigBee network function; Mainly, guarantee meeting under the network condition of operation the effect that the assessment objective track and localization is calculated according to the condition of sensor node deployment.

Fig. 3 is the statistics synoptic diagram that the binary sensor network system realizes target following.Sensor network is made up of these 7 sensor nodes of S1～S7; Tracked target passes through from the network detection area along the arrow curve; Asterism is 6 ad-hoc locations of measured target in motion process; Be the investigative range that target got into/withdrawed from a certain sensor node, the flag-shaped punctuate is represented the valuation position of method to positioning through target that utilize this patent to provide.

Accompanying drawing 4 is based on the embodiment that the target following test platform of binary sensor Ad Hoc network is disposed, and sets the target of being followed the tracks of and passes the regional area of sensor network disposition along straight line, is convenient to the performance of appraisal procedure like this.In fact target can be moved along free routing, and can be located by sensor node tracking.

The above-mentioned target following method of weighted mean of mentioning compare all less aspect the calculation cost with method provided by the invention, thereby they has the comparability on the performance.Performance index mainly are the errors that calculate the target location.

Test statistics shows; Method provided by the invention and the traditional average relative error ratio of method of weighted mean on the target following precision are 15.8%; The former is 83.1% than the latter at the average ratio that has reduced aspect the error, thereby has remarkable advantages aspect the bearing accuracy of target following.Because practical Target Tracking System is being had relatively high expectations aspect calculated amount and two indexs of bearing accuracy, the method here meets the requirement of these two indexs just, thereby it has potential practical function.

The above is the preferable main implementation process of the present invention, is not to be used for limiting practical range of the present invention; Every according to equivalence variation and modification that the present invention did, also contained by claim of the present invention.

Claims (1)

1. method for tracking target based on binary sensor Ad Hoc network; It is characterized in that: if in the overlay area of two-value microsensor MANET system deployment; The all the sensors node has identical radius of investigation, and moving target can be surveyed by 3 or above sensor node and find, detection information in the net that then provides through MANET; The 3rd sensor node calculates the position of target according to preceding two information that sensor node provides in the target detection sequence;

The step that realizes the aforesaid operations process is following:

The deployment of step 1. network system, initialization and information updating

Through the network system that the mode that manual work is buried underground or machine dispenses is disposed microminiaturized sensor, dispose the initialization that network system is carried out in the back, target approach monitoring state at hardware; Each sensor node is through communication, makes up its neighbor list and to the coordinate of neighbors circular self, neighbor list information comprises the status information sign of the target detection incident of neighbors numbering, neighbors coordinate, neighbors generation; Each sensor node periodically receives and the target detection information of upgrading the neighbors that it preserves, also gives neighbors with the own information announcement of target of finding simultaneously, with the calculation task of cooperation mode completion target location in this locality;

The sensor node of step 2. sensor network judges whether to trigger the behavior that calculate the target location

If satisfy following a kind of condition, then produce and trigger calculating operation:

(1) in the neighbor list of a sensor node, existing two neighbors have target and arrive Status Flag, and target is just getting into the investigative range of this sensor node this moment;

(2) when target gets into the investigative range of a sensor node; If the neighbors quantity that has had target arrival Status Flag in the neighbor list of this sensor node is above 2; Then,, therefrom select two neighbors by ascending order according to the numbering of neighbors; As the basis that three circles occur simultaneously and calculate, the coordinate of these two neighbors is participated in follow-up calculating;

Calculate step 3. target location

If a sensor node satisfies one of above-mentioned trigger condition, then calculate the current location of target according to three circle Intersection set method; The process of three circle Intersection set method comprises: when target gets into the investigative range of three sensor nodes successively; Calculate the 3rd sensor node earlier and survey circle S3 surveys circle S2 respectively with first sensor node detection circle S1, second sensor node 4 intersecting point coordinates; Whether the distance of judging these 4 intersection points and three centers of circle then respectively is equal to or less than the detection radius of circle; It is to be equal to or less than radius of circle that 2 intersection points are wherein arranged; These 2 intersection points are selected as two the end points a and the b of the common intersecting arcs of three circles, calculate the intermediate point of the line segment that these two end points form at last, with this as target in this residing estimated position constantly;

Gathering of step 4. target following information

Trigger the result of calculation of the sensor node of target location calculating with step 3; Through shortest path by the gateway aggregation node that sends it back sensor network; And be transferred to the computing machine of Surveillance center, and gather the position calculation result of target following, form the movement locus of tracked target.

Images