CN105959042A - One-bit feedback collaborative beam forming method based on variable step size - Google Patents

Abstract

The present invention provides a one-bit feedback collaborative beam forming method based on variable step size, which makes full use of the variable step size to adjust the disturbance amplitude and the correction factor to accelerate the system's convergence speed. When the phase difference of a collaborative node is large, the disturbance amplitude keeps unchanged. When the phase difference is smaller than the disturbance amplitude, the disturbance amplitude is leveled down in time. When a random disturbance succeeds, the correction factor becomes more intense for higher convergence speed; when the random disturbance fails, the correction factor is updated to the inverse value of the random disturbance of the previous time slot with the hope of correcting the disturbance direction. According to the invention, through random controls over disturbance amplitude and the step size of the correction factor, the convergence speed of the system is accelerated with higher disturbance accuracy.

Description

An a kind of bit feedback cooperative beam manufacturing process based on variable step size

Technical field

The present invention relates to wireless communication technology, particularly to a bit feedback cooperative beam forming technique based on variable step size.

Technical background

Beam shaping is the combination of antenna technology and Digital Signal Processing, and purpose is for phasing signal transmission or receives.At letter Number reception or transmitting terminal, by each road signal weighting of multiple element antennas is synthesized, the ideal signal required for formation.From sky Line directional diagram (pattern) angle is seen, is equivalent to the wave beam defining on assigned direction.To bay feedback carry out amplitude with The adjustment of phase place, forms the directional diagram of required form.

Beam shaping is applicable to multiaerial system, the most in mimo systems.Although MIMO technology has been considered as new one One of key technology for radio communication, but in most practical wireless communication systems, only many with antenna in base station end configuration. Mobile portable terminal, self-organized network nodes etc. by own dimensions, power and other realize factor and limited, it is difficult to configure more Antenna element.Therefore, MIMO technology is extensively applied and is limited significantly.Thus distributed multi-antenna system, i.e. " association Communicate " become an effectively replacement technology.

Cooperative beam shapes also referred to as distributed beams and shapes (Distributed Beamforming) or network beam shaping (Network Beamforming).The basic thought that cooperative beam shapes is when the overall CSI of link is known, by adjusting Cooperative node sends the multiple weights of signal and forms a virtual beams being directed at destination node, thus obtains collaboration diversity.Cooperation Weight computing is emphasis and the difficult point of cooperative beam forming technique.

At present, the Weights-selected Algorithm during cooperative beam shapes mainly has two kinds: a kind of is that the method by Distributed Calculation realizes power The calculating of value, i.e. information between cooperative node is mutual；Another kind is the calculating utilizing feedback mechanism to realize weights, i.e. by all The node of participation beam shaping comes dynamically adjust the weights of this locality and make it restrain with destination node alternately.First method due to Quantity of information mutual between cooperative node is excessive, expense is too high and it is too high to require channel information and it is bigger to cause realizing difficulty.The Two kinds of methods are less demanding to cooperative node and channel information, and feasibility is higher, wherein cooperative beam based on a bit feedback Shaping is to study to obtain a kind of the most methods.

Following documents describes existing based on a bit feedback cooperative beam manufacturing process:

[1]Mudumbai R,Hespanha J,Madhow U,et al.Scalable feedback control for distributed beamforming in sensor networks[C]//Information Theory,2005.ISIT 2005.Proceedings. International Symposium on.IEEE,2005:137-141.

[2]Song S,Thompson J S,Chung P J,et al.Improving the one-bit feedback algorithm for distributed beamforming[C]//Wireless Communications and Networking Conference(WCNC), 2010IEEE.IEEE,2010:1-6.

[3]Song S,Thompson J S.One-bit feedback algorithm with decreasing step size for distributed beamforming[C]//Cognitive Wireless Systems(UKIWCWS),2010Second UK-India-IDRC International Workshop on.IEEE,2010:1-5.

[4] patent " method of a kind of 1bit based on direction disturbance feedback cooperative beam figuration ". Chinese patent, publication number: CN104243007A, the day for announcing: 2014-12-24.

Document [1] proposes for the first time by adding a random disturbance in the weights phase place of cooperative node, thus realizes phase alignment, makes The RSS (received signal strength) of destination node reaches maximum.Disturbance is according to feedback information, adds in former phase basis One correction value.Destination node feeds back to cooperative node 1bit information every time, but when only feeding back 1, weights are just disturbed by cooperative node Dynamic operation, thus cause convergence rate the slowest.A kind of upper method is improved by document [2] with document [4], and convergence rate obtains The biggest lifting, but the amplitude of random disturbance is required bigger, when phase contrast is less, the disturbance of amplitude can reduce receipts on the contrary Hold back speed and even stop convergence process.Disturbance step-length is studied by document [3], by real time monitoring feedback information adjust with The size of machine disturbance amplitude, especially phase difference variable hour, reduce disturbance amplitude and be conducive to correct disturbance, there is problems of receiving Hold back speed the slowest.

Summary of the invention

The bit based on variable step size that the technical problem to be solved is to provide a kind of convergence speed accelerating system is anti- Feedback beam-forming method.

The present invention solves that above-mentioned technical problem be employed technical scheme comprise that, an a kind of bit feedback cooperation based on variable step size Beam-forming method, comprises the following steps:

Step 1 time slot n cooperative node i uses self adaptation phase place weights φ_i[n] participates in beam shaping, and sends ripple to destination node Bundle；Self adaptation phase place weights are φ_i[n]=θ_i[n]+δ_i[n]+τ_i[n]；θ_i[n] be cooperative node i time slot n phase place weights, δ_i[n] is cooperative node i at the disturbance amplitude of time slot n, τ_i[n] is the cooperative node i correction factor at time slot n；Afterwards, enter Enter step 2；

Step 2 time slot n cooperative node i updates the number of times of failure disturbance continuously according to the control information receiving destination node feedback C_fIf the control information of feedback represents disturbance success, the number of times C of failure disturbance continuously is set_f=0, adjust next time slot The phase place weights θ of n+1_i[n+1]=φ_i[n] and correction factor τ_i[n+1]=(1/R_D)·τ_i[n], R_DRepresent zoom factor, 0≤R_D≤1；Afterwards, step 3 is entered；

If the control information of feedback represents disturbance failure, C_f=C_f+1；The number of times C of failure disturbance continuously_fAfter renewal, will be even Continue the number of times of unsuccessfully disturbance and upper limit C of failure disturbance continuously_tCompare, work as C_f＜ C_t, then current disturbance amplitude is kept not Become, work as C_f≥C_tTime, update disturbance amplitude δ₀=R_D·δ₀After, then update the number of times C of failure disturbance continuously_fIt is 0；Adjust The phase place weights θ of next time slot n+1_i[n+1]=θ_i[n] and correction factor τ_i[n+1]=-δ_i[n]；Afterwards, step is entered 3；

Whether step 3 receives the confirmation that destination node sends, and in this way, cooperative node keeps current state and sends follow-up number It is believed that breath, otherwise update time slot n=n+1, return step 1.

The present invention makes full use of variable step size and is adjusted disturbance amplitude and correction factor size, accelerates system convergence speed.When When phase contrast is bigger between cooperative node, keep disturbance amplitude；When phase contrast is less than disturbance amplitude, reduce disturbance amplitude in time； When random disturbance success, correction factor strengthens correction dynamics, accelerates convergence rate；When random disturbance lost efficacy, correction factor It is updated to the inverse value of a time slot random disturbance, is used for correcting perturbation direction.

The invention has the beneficial effects as follows, by mobility control disturbance amplitude and the step-length of correction factor, accelerate the convergence speed of system Degree, the accuracy making disturbance is higher.When especially phase contrast is less, reduces disturbance amplitude and adjust correction factor to adding rapid convergence tool There is remarkable result.

Accompanying drawing explanation

Fig. 1 is the system model figure of the present invention；

Fig. 2 is that the embodiment of the present invention and comparative example are at δ₀=pi/2 5 convergence rate comparison diagram；

Fig. 3 is that the embodiment of the present invention and comparative example are at δ₀=pi/2 00 convergence rate comparison diagram；

Fig. 4 is the embodiment of the present invention and the minimum time slot comparison diagram of comparative example convergence；

Fig. 5 is that the embodiment of the present invention and comparative example are relative to document [1] gain contrast figure.

Detailed description of the invention

The method that a present invention bit feedback cooperative beam based on variable step size shapes comprises the steps of

Step 1: model is set up: setting total N+1 node, wherein cooperative node number is N, and destination node number is 1.

Step 2: systematic parameter initializes: sets n and represents time slot, initial slot n=n_init, and θ_i[n]=θ_i-init, δ_i[n]=± δ₀, τ_i[n]=τ_i-init, RSS [n]=RSS_i-init>=0, wherein n_init, θ_i-init, τ_i-init, RSS_i-initIt is initial value.θ_i[n] It is the cooperative node i phase place weights at time slot n；δ_i[n] is the cooperative node i disturbance amplitude at time slot n, and equiprobability value ±δ₀；τ_i[n] is the cooperative node i correction factor at time slot n；RSS [n] is the destination node received signal strength at time slot n (Received Signal Strength)。

Step 3: scale parameter sets: C_fRepresent the number of times of failure disturbance continuously, C_tRepresent the upper limit of failure disturbance continuously, R_D Represent zoom factor, for adjusting the size of disturbance amplitude and correction factor, wherein C_f≥0,C_t≥1,0≤R_D≤1。

Step 4: time slot n, cooperative node i uses the phase place weights after self adaptation to participate in beam shaping, and self adaptation phase place weights are φ_i[n]=θ_i[n]+δ_i[n]+τ_i[n]。

Step 5: time slot n, destination node receives the beam signal of cooperative node transmission and calculates received signal strength, being designated as RSS [n], and it is accepted signal strength values RSS with the maximum locally recorded_best[n] compares.If RSS [n] ＞ RSS_best[n], Showing this time slot success disturbance, destination node controls information 1 to all cooperative nodes feedback 1bit；If RSS [n]≤RSS_best[n], Destination node controls information 0 to all cooperative nodes feedback 1bit.

Step 6: cooperative node receives the control information of feedback, C_fThe number of times of record failure disturbance continuously.If the control letter of feedback Breath is 1, C_f=0；If the control information of feedback is 0, C_f=C_f+1.Work as C_f≥C_tTime, disturbance amplitude is updated to δ₀=R_D·δ₀, and C_fIt is set to 0.

Step 7: cooperative node receive time slot n feedback information, and according to feedback information adjust time slot n+1 phase place weights with Correction factor.If the control information of feedback is 1, then phase place right value update is the self adaptation phase place weights of a upper time slot, i.e. θ_i[n+1]=φ_i[n]=θ_i[n]+δ_i[n]+τ_i[n]；Correction factor is updated to the 1/R of a time slot_DTimes, i.e. τ_i[n+1]=(1/R_D)·τ_i[n].If the control information of feedback is 0, the phase place weights of cooperative node keep constant, have θ_i[n+1]=θ_i[n]；Correction factor is updated to the opposite number of a time slot disturbance amplitude, has τ_i[n+1]=-δ_i[n]。

Step 8: destination node updates the maximum of received signal strength, and mathematic(al) representation is RSS_best[n+1]=max (RSS [n], RSS_best[n]), detect the maximum RSS of received signal strength simultaneously_best[n+1] Whether reach system minimum work thresholding.If meeting the minimum threshold of system received signal strength, then to all cooperative nodes Broadcast acknowledgements information ACK, cooperative node keeps current state and sends follow-up data information.Otherwise, time slot n adds 1 automatically, And repeat step 4-7.

Above step can represent with mathematic(al) representation, as follows:

Transmitting terminal, i-th (i=1,2 ..., N) individual cooperative node:

${\mathrm{\θ}}_i\[n+1\]=\left\{\begin{array}{cc}{\mathrm{\φ}}_i\[n\]={\mathrm{\θ}}_i\[n\]+{\mathrm{\δ}}_i\[n\]+{\mathrm{\τ}}_i\[n\]& RSS\[n\]>{\mathrm{RSS}}_{best}\[n\]\\ {\mathrm{\θ}}_i\[n\]& RSS\[n\]\≤{\mathrm{RSS}}_{best}\[n\]\end{array}\right.---\left(1\right)$

${\mathrm{\τ}}_i\[n+1\]=\left\{\begin{array}{cc}\frac{1}{R_D}\·{\mathrm{\τ}}_i\[n\]& RSS\[n\]>{\mathrm{RSS}}_{best}\[n\]\\ -{\mathrm{\δ}}_i\[n\]& RSS\[n\]>{\mathrm{RSS}}_{best}\[n\]\end{array}\right.---\left(2\right)$

$C_f=\left\{\begin{array}{cc}0& \begin{array}{ccc}RSS\[n\]>{\mathrm{RSS}}_{best}\[n\]& or& C_f\≥C_t\end{array}\\ C_f+1& otherwise\end{array}\right.---\left(3\right)$

${\mathrm{\&delta;}}_0=\left\{\begin{array}{cc}{\mathrm{\&delta;}}_0& C_f<C_t\\ R_D\&CenterDot;{\mathrm{\&delta;}}_0& C_f\&GreaterEqual;C_t\end{array}\right.---\left(4\right)$

Receiving terminal, destination node:

RSS_best[n+1]=max (RSS [n], RSS_best[n]) (5)

In order to be better understood from the present invention, below in conjunction with accompanying drawing, on the basis of technical scheme, describe the step that scheme is implemented in detail Suddenly, but the practical range of the present invention be not limited solely to describe scene.

Present case has 101 nodes, system model as it is shown in figure 1, wherein cooperative node 100, destination node 1, 100 cooperative nodes are randomly dispersed in the border circular areas of radius R, and destination node distance border circular areas center is D=100； X [n] is that cooperative node shares data message；Dotted line represents the beam pattern that cooperative node is formed.

Cooperative node is randomly dispersed in the border circular areas of radius R=4, and distance D=100 of regional center and destination node； Channel, for become channel slowly, only considers the phase effect that channel brings during whole.Meanwhile, the transmission merit of all cooperative nodes Rate is identical and normalization.

In order to make description closer to reality, it is assumed that starting time slot is 0；The initial phase weights of cooperative node are 0, i.e. θ_i[n]=θ_i-init=0, i=1,2 ... 100；The correction factor initial value of cooperative node is 0, i.e. τ_i[n]=τ_i-init=0, i=1,2 ... 100；The random disturbance value of cooperative node has δ_i[n]=± δ₀, i=1,2 ... 100；Purpose It is 0 that node accepts signal intensity maximum, i.e. RSS_best=0.

Whether scale parameter value is appropriate, the convergence rate of relation whole system fast and slow.By emulation several times and actual behaviour Make, C in this case_tWith R_DValue be respectively as follows: C_t=11, R_D=0.8.

So far system building is complete, and the enforcement step of a bit feedback beam-forming method based on variable step size is provided below, As follows:

Step 1: cooperative node generates self adaptation phase place weights.Each cooperative node produces a random disturbance value and a correction The factor.Random disturbance value δ_i[n], i=1,2 ... 100, and equiprobability value ± δ₀.Control information according to a upper time slot feedback Generate correction factor τ_i[n], i=1,2 ... 100.Then, self adaptation phase place weights are φ_i[n]=θ_i[n]+δ_i[n]+τ_i[n], i=1,2 ... 100, each cooperative node is as the transmission phase place of this time slot.

Step 2: cooperative node shares data message x [n], and carries the transmission of respective weights, and final each cooperative node sends Information is

Step 3: destination node reception is through the beam signal r [n] of Gaussian channel:

In formula (6), for the purpose of r [n], node receives the signal with white Gaussian noise；Represent cooperative node I is to the gradual change channel of destination node, wherein a_iRepresent channel magnitude,Represent channel phase；γ_iRepresent the phase of non-ideal synchronization Position influence；ω [n] represents white Gaussian noise signal, and ω [n]～N (0, σ²).Therefore compared with single antenna, beam shaping Signal gain is represented by:

Actually we are by the intensity of channel gain equivalent received signals, have RSS [n]=R [n] during time slot n.Represent the cooperative node i phase offset to destination node.When phase offset is full Foot Φ₁[n]=Φ₂[n]=...=Φ₁₀₀Time [n], received signal strength value is maximum:

$R\&lsqb;n\&rsqb;=\left|\underset{i=1}{\overset{100}{\&Sigma;}}{a}_i{e}^{{\mathrm{j\&Phi;}}_i\&lsqb;n\&rsqb;}\right|\&le;\underset{i=1}{\overset{100}{\&Sigma;}}{a}_i---\left(8\right)$

Owing to system channel is gradual change channel, γ_iWithFixing unknown parameter can be regarded as, so we send phase place by adjusting And find out the θ of optimum_i[n], makes received signal strength the strongest.

Step 4: time slot n, destination node calculates and accepts signal intensity RSS [n], and and RSS_best[n] compares: if RSS [n] ＞ RSS_best[n], shows this time slot disturbance success, and destination node controls information 1 to cooperative node broadcast 1bit；If RSS[n]≤RSS_best[n], shows this time slot disturbance failure, and destination node controls information 0 to cooperative node broadcast 1bit.

Step 5: cooperative node receives the control information of feedback, C_fThe number of times of record failure disturbance continuously.If the control letter of feedback Breath is 1, C_f=0；If the control information of feedback is 0, C_f=C_f+1.Work as C_fWhen >=11, disturbance amplitude is updated to δ₀=0.8 δ₀, and C_fIt is set to 0.

Step 6: cooperative node receive time slot n feedback information, and according to feedback information adjust time slot n+1 phase place weights with Correction factor.If the control information of feedback is 1, then phase place right value update is the self adaptation phase place weights of a upper time slot, i.e. θ_i[n+1]=φ_i[n]=θ_i[n]+δ_i[n]+τ_i[n]；Correction factor becomes the 1/R of a upper time slot_DTimes, i.e. τ_i[n+1]=(1/R_D)·τ_i[n].If the control information of feedback is 0, the phase place weights of cooperative node keep constant, have θ_i[n+1]=θ_i[n]；Correction factor is updated to the opposite number of a time slot disturbance amplitude, has τ_i[n+1]=-δ_i[n].Step 7: destination node updates the maximum of received signal strength, and mathematic(al) representation is RSS_best[n+1]=max (RSS [n], RSS_best[n]), detect the maximum RSS of received signal strength simultaneously_best[n+1] Whether reach system minimum work thresholding.If meeting the minimum threshold of system received signal strength, then to all cooperative nodes Broadcast acknowledgements information ACK, cooperative node keeps current state and sends follow-up data information.Otherwise, time slot n adds 1 automatically, And repeat step 1-6.

Fig. 2 from Fig. 3 gives 5 kinds of different 1bit and feeds back cooperative beam manufacturing process convergence processs under different disturbance amplitudes: Work as δ₀During=pi/2 5, present invention convergence rate after 500 time slots is significantly better than other four kinds of methods, and reaches at first System door limit value；Work as δ₀During=π/100, the present invention just shows convergence rate faster from initial slot.So, whether The quick regulation that disturbance amplitude is bigger, or the accurate regulation that disturbance amplitude is less, the present invention is significantly faster than in convergence rate Other four kinds of methods.

Fig. 4 features the minimum timeslot number when given different convergence threshold value.It is clear that the present invention arrives different convergence door The timeslot number of limit value is minimum, and especially when tending to system door limit value, the minimum timeslot number of the present invention is at least little by 2000 than additive method Individual.

Fig. 5 illustrates document [2], document [3] and the present invention relative to the convergence rate gain of original method in document [1].Such as figure Shown in, document [2] compared to the convergence rate gain of original method about 30%；Document [3] is compared to the convergence speed of original method Degree gain is about 70%；The present invention about 90%, reaches as high as 93% compared to the convergence rate gain of original method.

Claims (5)

1. a bit feedback cooperative beam manufacturing process based on variable step size, it is characterised in that comprise the following steps:

Step 1 time slot n cooperative node i uses self adaptation phase place weights φ_i[n] participates in beam shaping, and sends ripple to destination node Bundle；Self adaptation phase place weights are φ_i[n]=θ_i[n]+δ_i[n]+τ_i[n]；θ_i[n] be cooperative node i time slot n phase place weights, δ_i[n] is cooperative node i at the disturbance amplitude of time slot n, τ_i[n] is the cooperative node i correction factor at time slot n；Afterwards, enter Enter step 2；

Step 2 time slot n cooperative node i updates the number of times of failure disturbance continuously according to the control information receiving destination node feedback C_fIf the control information of feedback represents disturbance success, the number of times C of failure disturbance continuously is set_f=0, adjust next time slot The phase place weights θ of n+1_i[n+1]=φ_i[n] and correction factor τ_i[n+1]=(1/R_D)·τ_i[n], R_DRepresent zoom factor, 0≤R_D≤1；Afterwards, step 3 is entered；

If the control information of feedback represents disturbance failure, C_f=C_f+1；The number of times C of failure disturbance continuously_fAfter renewal, will be even Continue the number of times of unsuccessfully disturbance and upper limit C of failure disturbance continuously_tCompare, work as C_f<C_t, then current disturbance amplitude is kept not Become, work as C_f≥C_tTime, update disturbance amplitude δ₀=R_D·δ₀After, then update the number of times C of failure disturbance continuously_fIt is 0；Adjust The phase place weights θ of next time slot n+1_i[n+1]=θ_i[n] and correction factor τ_i[n+1]=-δ_i[n]；Afterwards, step is entered 3；

Whether step 3 receives the confirmation that destination node sends, and in this way, cooperative node keeps current state and sends follow-up number It is believed that breath, otherwise update time slot n=n+1, return step 1.

A kind of bit feedback cooperative beam manufacturing process based on variable step size, it is characterised in that Cooperative node i is at the phase place weights θ of time slot n_iThe initial value of [n] is 0, δ_i[n] is that cooperative node i is in the disturbance amplitude of time slot n ±δ₀, cooperative node i is at the correction factor τ of time slot n_i[n] is 0.

A kind of bit feedback cooperative beam manufacturing process based on variable step size, it is characterised in that Upper limit C of failure disturbance continuously_t=11, zoom factor R_D=0.8.

A kind of bit feedback cooperative beam manufacturing process based on variable step size, it is characterised in that The method of the control information of destination node feedback is:

Time slot n destination node receives the beam signal of cooperative node transmission and calculates received signal strength RSS [n], reception is believed Maximum received signal strength value RSS of number intensity RSS [n] and local record_best[n] compares, if RSS [n] > RSS_best[n], Destination node represents disturbance success to all cooperative node feedback control informations 1；If RSS [n]≤RSS_best[n], destination node Disturbance failure is represented to all cooperative node feedback control informations 0.

A kind of bit feedback cooperative beam manufacturing process based on variable step size, it is characterised in that After time slot n destination node receives transmission wave beam, the maximum updating next time slot n+1 received signal strength is RSS_best[n+1]=max (RSS [n], RSS_best[n]), max represents and takes maximum, and detection received signal strength is the most simultaneously It is worth greatly RSS_bestWhether [n+1] reaches system minimum work thresholding, in this way, then to all cooperative node broadcast acknowledgements information, no Then, time slot n=n+1 is updated.