CN103716898B - Underwater sensor network media access control method based on motion prediction - Google Patents

Abstract

The invention belongs to underwater sensor network technical field, a kind of media access control method based on motion prediction is proposed, use RTS, CTS, DATA, ACK tetra-class message communicate: all node random broadcast signals, the distance that each node calculates between self and neighbor node according to the propagation time and known spread speed that receive signal, and the distance obtained between self and neighbor node is broadcasted again, thus each node have recorded all internodal mutual distance values, Distance matrix D is utilized to characterize；By node present speed and past movement velocity, the movement velocity that node is following is carried out linear prediction, then update the distance value in Distance matrix D, ask for maximum in Distance matrix D, and calculate reserving time slots length；Application time-division slot reservation transmission mechanism communicates.The present invention, on the basis of realizing inter-node communication conflict avoidance, takes full advantage of time interval resource, improves the utilization rate of channel.

Description

Underwater sensor network media access control method based on motion prediction

Technical field

The present invention relates to underwater sensor network (Underwater Sensor Networks, UWSN) communication technical field, specifically The media access control method of conflict avoidance and efficient communication is realized between sensor node based on motion prediction.

Background technology

Underwater sensor network (UWSN) has become as existing a kind of preferable exploration and observes the instrument of ocean, has in the following areas The biggest application: oceanographic data collection, pollution monitoring, water-bed explore, disaster is avoided, succour navigation, military strategy observe apply, Mineral resources exploration, seismic monitoring, pipeline under the ocean leakage monitoring, the auxiliary of submerged structure, and ecological monitoring etc..But, The characteristic of marine environment brings a lot of difficulty to the structure of sensor network, such as: power problems, length and variable propagation delay time, have The available bandwidth of limit, high bit error rate etc., simultaneously as acted on by ocean current and wave, whole UWSN network is in motion shape all the time State, therefore, the existing more perfect media access control method being applied to land wireless network is not all suitable for network under water, it is necessary to pin The communication requirement of UWSN is designed new media access control method.

The main purpose of network medium access method is contemplated to solve data packet collisions problem under water, simultaneously need to consider energy efficiency, low Channel access postpones and fairness between node.Currently, the new media access control method of many of UWSN it is suitable for it has been proposed that on the whole Two classes can be divided into: uncompetitive method and method based on competition.Uncompetitive as the methods such as FDMA, CDMA be limited to respectively bandwidth and Encoded question, fails to become the main flow of access control method for network medium under water.Based on competition method be divided into random division multiple accessing method and based on Channel booking class method, former approach simple in construction, it is achieved convenient, but data packet collisions probability along with node traffic amount increase exponentially Increase, be not suitable for bursty data scene；The latter is based on channel booking class method, and such as R-MAC, collision avoidance mechanism is complicated, network delay Greatly, the node location uncertainty that the dynamic characteristic of network causes under water affects the degree of accuracy of its reservation so that R-MAC is in long propagation delay time Network in throughput performance performance not good enough.

Existing access control method for network medium under water is both for what static network proposed, does not dynamically change in view of node location Situation, and in actual environment, owing to being acted on by ocean current and wave, whole UWSN network is kept in motion all the time.In order to dynamic State underwater sensor network realizes node data packet conflict avoid, improve communication efficiency, improve network throughput, it is necessary to for moving under water The impact that network service is caused by state environment proposes a kind of reasonably solution.

Summary of the invention

In order to improve the data packet collisions caused in UWSN media access control method and the problem such as communication efficiency is low due to joint movements, the present invention carries Go out a kind of connected applications motion prediction and channel booking mechanism, the side of reserving time slots length can be adaptively adjusted according to the change of euclidean distance between node pair Method, improves communication efficiency, improves network throughput.Technical scheme is as follows:

A kind of media access control method based on motion prediction, use RTS, CTS, DATA, ACK tetra-class message communicate, take into account Joint movements prediction and network communication efficiency, it is adaptable to underwater sensor network, the method comprises the following steps:

The first step, all node random broadcast signals, each node calculates self with adjacent according to the propagation time and known spread speed that receive signal Occupy internodal distance, and the distance obtained between self and neighbor node be broadcasted again, thus each node have recorded all internodal mutually away from Distance values, utilizes Distance matrix D to characterize；

Second step, carries out linear prediction by node present speed and past movement velocity to the movement velocity that node is following, then updates Distance matrix D In distance value, ask for maximum in Distance matrix D, and calculate reserving time slots length:

$T_{\mathrm{slot}}=\frac{\max \left(D\right)}{v}+t_{\mathrm{cts}}+t_d$

Wherein, t_ctsFor the transmission time of CTS bag, t_dFor guard time, v is the velocity of sound in water；

3rd step, the reserving time slots length tried to achieve in conjunction with second step, application time-division slot reservation sends mechanism, communicates, and the transmission of all packets all must Must be at the initial time of time slot:

1) when certain node has data to need to send, it will wait until first time slot start time, and send RTS bag and preengage；

2) destination node receives RTS bag, then can send CTS bag at the initial time of next time slot, inform all neighbor nodes, and self has been prepared for Good acceptance is from the packet of sending node, and other nodes are kept out of the way voluntarily；

3) source node receives the CTS bag of destination node, represents the success of self reserve channel, and sends DATA at next time slot initial time；

4), after destination node has received packet, can send at next time slot initial time and be properly received confirmation message ACK.

Preferably, in second step, use autoregression (Auto Regressive, AR) model prediction, by node present speed With past movement velocity, the movement velocity that node is following is predicted, is divided into multiple predicting unit according to the time, is located in adjacent predicting unit, The motion mode of object is held essentially constant, and the speed of application m predicting unit in the past is predicted obtaining:

$\stackrel{\→}{V\left(n\right)}=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{a}_i\stackrel{\→}{V(n-i)}$

Wherein V (n) is the velocity of the n-th predicting unit, a_iIt is predictive coefficient, can be tried to achieve by Levison-Durbin recursive algorithm, so It is integrated speed afterwards i.e. can get distance value new in Distance matrix D.

It is adaptively adjusted channel booking slot length according to the change of mutual distance between node, causes network area to shrink, in region when node due to motion When the spacing of portion's node diminishes, reserving time slots shortens, otherwise, reserving time slots increases accordingly.

The beneficial effects of the present invention is, compared with existing media access control method, the program can better adapt to dynamic environment under water, improves Network communication efficiency, improves network throughput.The program has taken into full account the motion conditions of node under water, by grinding immersed body kinetic characteristic Study carefully design node motion prediction algorithm, be adaptively adjusted channel booking slot length according to the change of mutual distance between node, when node is due to motion Causing network area to shrink, when the spacing of intra-zone node diminishes, reserving time slots shortens, otherwise, time slot increases accordingly.With it, The conflict avoidance under UWSN network dynamic environment can not only be realized, and substantially increase the utilization rate of channel resource, improve the handling capacity of network Performance.

Accompanying drawing explanation

Fig. 1 is media access control method flow chart based on motion prediction

Fig. 2 is internodal distance map

Fig. 3 is that reservation sends schematic diagram of mechanism

Fig. 4 is RTS bag form, and wherein Packet type refers to that type of message, Packet length refer to message length, sender ID and receiver ID Refer to that sending node and the address of receiving node, Data packet length represent the required message length sending data of node respectively.

Fig. 5 is CTS bag form, and wherein Packet type refers to that type of message, Packet length refer to message length, sender ID and receiver ID Refer to that sending node and the address of receiving node, Data slot represent the number of time slots taken needed for packet transmission respectively.

Detailed description of the invention

Before the detailed description of the invention of the present invention is illustrated, first illustrate floating motion and the basic theories of linear prediction:

The motion of node under water is mainly affected by wave, and wave can regard gravitational wave as, and the motion of water particle has to come and go shakes the characteristics of motion. Using linear wave theory to be analyzed gravitational wave, ignore the non-linear factor under boundary condition, ocean wave motion is reduced to rule traveling wave, it is special Levy parameter:

$\mathrm{\η}=\frac{h}{2}\cos (\mathrm{kx}-\mathrm{\ωt})---\left(1\right)$

Wherein ripples are high: h, wave number: k=2 π/λ, wavelength: λ=gT/2 π, cycle: T, wave frequency: ω=2 π/T

For simplifying the analysis, set the direct of travel of wave as x-axis direction, and true environment situation be closer to the superposition of N number of velocity, So solve mode according to the movement velocity vector of water particle, simplify through a large amount of, obtain the velocity expression formula of ocean wave motion:

$\stackrel{\→}{V}=\left(\begin{array}{c}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{a}_i\cos (k_{i}x-\mathrm{\ωt})\\ 0\\ \underset{i=1}{\overset{M}{\mathrm{\Σ}}}{b}_i\sin (k_{i}z-\mathrm{\ωt})\end{array}\right)---\left(2\right)$

It can be seen that the complicated wave motion under true environment can regard N(M as) synthesis of individual simple rule ripple.

For the calculating of wave force suffered by the object in water of sea, it is respectively adopted two kinds of different methods by dimension of object size and is analyzed, large sized object Use diffraction theory；And small-size object (article diameters and the ratio of wave wavelength :), use semiempirical semi-theoretical Morison formula.Literary composition Middle studied node under water, with reference to UWM1000 underwater sound MODEM, belongs to small-size object.Assuming that wave incidence wave is transmitted by x-axis direction, Wave force suffered by object in water is represented by:

$f=\frac{\mathrm{dF}}{\mathrm{dz}}=C_d\mathrm{A\ρ}\frac{v}{2}\left|v\right|+C_m\frac{\mathrm{\π}}{4}{D}^2\mathrm{\ρv}---\left(3\right)$

Wherein C_m、C_dRepresenting quality coefficient and resistance coefficient respectively, v is the speed of water particle.

The motion of object in water owing to being caused by wave force, by formula (2) wave equation it can be seen that the motion of water particle be not with Machine moves, but rhythmic, and in like manner, the motion of object in water is also periodic, can draw the motion speed of immersed body according to formula (2) Degree vector expression is:

$\stackrel{\→}{V_M}=\left(\begin{array}{c}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{c}_i\cos (k_{i}x-\mathrm{\ωt})\\ 0\\ \underset{i=1}{\overset{M}{\mathrm{\Σ}}}{d}_i\sin (k_{i}z-\mathrm{\ωt})\end{array}\right)---\left(4\right)$

Individual sine-wave superimposed equally, it is also possible to find out that the equation of motion of immersed body is by N(M) forms.In real scene, underwater sensing Device node is to lie under water surface float, and the impact of its shift in position in vertical direction (z-axis) is negligible, and therefore, enters formula (4) One step simplifies, and can obtain the equation of motion expression formula of underwater sensor node:

$\stackrel{\→}{V_M}=\left(\begin{array}{c}a\·\cos \left({\mathrm{\ω}}_{a}t\right)+b\·\cos \left({\mathrm{\ω}}_{b}t\right)+c\·\cos \left({\mathrm{\ω}}_{c}t\right)+d\·\cos \left({\mathrm{\ω}}_{d}t\right)\\ 0\\ 0\end{array}\right)---\left(5\right)$

Wherein a, b, c, d, ω_a、ω_b、ω_c、ω_dAll relevant to concrete scene etc. ambient parameter value.

By the equation of motion described above it can be seen that the motion of underwater sensor node has half period property feature, say, that the position of node It is foreseeable.By the motion model of analysis node, it is considered to use AR model prediction, by node present speed and past movement velocity pair The following movement velocity of node is predicted.First dividing time into multiple predicting unit, length can be set to T, and the motion of immersed body has Having half period property rule, it is assumed here that in adjacent predicting unit, the motion mode of object is held essentially constant, i.e. the coefficient of AR model is constant, Can obtain:

$\stackrel{\→}{V\left(n\right)}=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{a}_i\stackrel{\→}{V(n-i)}---\left(6\right)$

Wherein V (n) is the velocity of the n-th predicting unit, a_iIt is predictive coefficient, can be tried to achieve by Levison-Durbin recursive algorithm.By Formula (6) can be seen that node present speed is the linear combination of its front m speed, and this is also the reason of referred to as linear prediction algorithm.Thus, On the premise of speed is known, its integration can be tried to achieve internodal distance relation:

$d_t=\underset{t_0}{\overset{t_0+T}{\∫}}\stackrel{\→}{V\left(t\right)}\mathrm{dt}+d_0---\left(7\right)$

Accompanying drawing 1 be the program implement flow process.The media access control method based on motion prediction that the present invention proposes, uses request Transmission message (RTS), clear to send message (CTS), data (DATA), confirmation message (ACK) four class message communicate, and hold concurrently Turn round and look at joint movements prediction and network communication efficiency.

The first step that scheme performs is exactly to obtain the positional information of region interior nodes.Internodal mutual distance can use TOA(in time synchronized On the basis of, propagation time and known spread speed according to signal calculate internodal distance) method records.As shown in Figure 2, node 1 The distance between other nodes can be calculated, obtain distance vector D1=[0, d12, d13, d14, d15], and be broadcasted, the most each node Have recorded the distance value of all nodes, obtain distance matrix:

$D=\left[\begin{array}{ccccc}0& d_{12}& d_{13}& d_{14}& d_{15}\\ d_{21}& 0& d_{23}& d_{24}& d_{25}\\ d_{31}& d_{32}& 0& d_{34}& d_{35}\\ d_{41}& d_{42}& d_{43}& 0& d_{45}\\ d_{51}& d_{52}& d_{53}& d_{54}& 0\end{array}\right]---\left(8\right)$

Wherein d_ijBeing that node i records and distance value between node j, Distance matrix D will be used for determining reserving time slots length.

Second step, cycle movement is predicted.Make full use of motion prediction, it is possible to reduce the communication consumption caused because obtaining positional information, obtain simultaneously Obtain high range accuracy.Node is on the basis of learning Distance matrix D, it is possible to be calculated slot length by formula (3), slot length by Maximum transmission distance, the transmission time of CTS bag and protection interval determine:

T_slot=f (predict (D), max (d_ij))+t_cts+t_d(9)

Wherein t_ctsFor the transmission time of CTS bag, t_dFor guard time, d_ijBeing the distance value in matrix D, function f (.) comprises three mistakes Journey, first applies predictive algorithms periodically predict the variation of euclidean distance between node pair, then ask for the maximum max (d in matrix D_ij), in conjunction with The signal velocity known tries to achieve maximum transmitted time delay in region.

Calculate gained T_slotValue is then the slot length of agreement reservation transmission phase.Additionally, in order to avoid the accumulation of forecast error, node is one To again position, to obtain internodal accurate distance after section time (much larger than predetermined period).

3rd step, the program uses time-division slot to send data, carries out conflict avoidance, solves hidden terminal problem.Scheme use RTS, CTS, DATA, ACK tetra-class message communicates, and the transmission of all packets all must be at the initial time of time slot, and slot length is the T that forecast period obtains_slotValue, Slot length ensure that RTS/CTS bag can be received by all nodes in region, and periodically updates it, as shown in Figure 3.

Wherein R, C, A represent RTS, CTS, ACK message, T respectively₁Represent slot length, it can be seen that through after a while After, the mobility of node causes node N₃With N₁Between distance become big, slot length is updated to T accordingly₂。

RTS: request sends message

When a node have data need send time, it is necessary first to by send RTS bag preengage, bag form as shown in Figure 4, wherein Packet Type refer to type of message, Packet length refer to message length, sender ID and receiver ID refer to respectively sending node and the address of receiving node, Data packet length represents the required message length sending data of node.Node is sending before RTS bag, elder generation's monitor channel, if channel busy Or when judging to clash, then actively keep out of the way, back off time is random slot length.

CTS: clear to send message

When node receives the request message RTS of other nodes, then can send CTS bag at the initial time of next time slot, inform all neighbours Occupying node, self be already prepared to accept the packet from sending node, other nodes keep out of the way the time of corresponding slot length voluntarily.Bag form As shown in Figure 5, wherein Packet type refers to that type of message, Packet length refer to that message length, sender ID and receiver ID refer to send out respectively Node and the address of receiving node, Data slot is sent to represent the number of time slots taken needed for packet transmission.

DATA: packet

Source node receives the CTS bag of destination node, represents the success of self reserve channel, and sends DATA at next time slot initial time.

ACK: confirmation message

After destination node has received packet, can send at next time slot initial time and be properly received confirmation message ACK.

Claims (3)

1. a media access control method based on motion prediction, use RTS, CTS, DATA, ACK tetra-class message communicate, hold concurrently Turning round and look at joint movements prediction and network communication efficiency, it is adaptable to underwater sensor network, the method comprises the following steps:

The first step, all node random broadcast signals, each node calculates self according to the propagation time and known spread speed that receive signal And the distance between neighbor node, and the distance obtained between self and neighbor node is broadcasted again, thus each node have recorded between all nodes Mutual distance value, utilize Distance matrix D to characterize；

Second step, carries out linear prediction by node present speed and past movement velocity to the movement velocity that node is following, then updates distance matrix Distance value in D, asks for maximum in Distance matrix D, and calculates reserving time slots length:

$T_{slot}=\frac{max\left(D\right)}{v}+t_{cts}+t_d$

Wherein t_ctsFor the transmission time of CTS bag, t_dFor guard time, v is the velocity of sound in water；

3rd step, the reserving time slots length tried to achieve in conjunction with second step, application time-division slot reservation sends mechanism, communicates, the transmission of all packets All must be at the initial time of time slot:

1) when certain node has data to need to send, it will wait until first time slot start time, and send RTS bag and preengage；

2) destination node receives RTS bag, then can send CTS bag at the initial time of next time slot, inform all neighbor nodes, and self is Being ready to accept the packet from sending node, other nodes are kept out of the way voluntarily；

3) source node receives the CTS bag of destination node, represents the success of self reserve channel, and sends DATA at next time slot initial time；

4), after destination node has received packet, can send at next time slot initial time and be properly received confirmation message ACK.

2. media access control method based on motion prediction as claimed in claim 1, it is characterised in that: in second step, use autoregression (Auto Regressive, AR) model prediction, by node present speed and in the past movement velocity the movement velocity that node is following is predicted, according to time Between be divided into multiple predicting unit, be located in adjacent predicting unit, the motion mode of object is held essentially constant, application in the past m predicting unit Speed be predicted obtaining:

Wherein V (n) is the velocity of the n-th predicting unit, a_iIt is predictive coefficient, Levison-Durbin recursive algorithm tries to achieve, so It is integrated speed afterwards i.e. can get distance value new in Distance matrix D.

3. media access control method based on motion prediction as claimed in claim 1, it is characterised in that: according to the changing of mutual distance between node Become and be adaptively adjusted channel booking slot length, cause network area to shrink when node due to motion, when the spacing of intra-zone node diminishes, Reserving time slots shortens, otherwise, reserving time slots increases accordingly.