CN105656551A - Precoding MIMO-OOFDM-VLC (Multiple-input Multiple-output-Optical Frequency Division Multiplexing-Visible Light Communication) imaging communication method based on PDS (Photodetector Selection) - Google Patents

Abstract

The invention discloses a precoding MIMO-OOFDM-VLC (Multiple-input Multiple-output-Optical Frequency Division Multiplexing-Visible Light Communication) imaging communication method based on PDS (Photodetector Selection). In the method, imaging reception and a receiving antenna selection technology are combined, and the VLC channel relevance among PDs (Photodetectors) of a same user terminal or different user terminals can be lowered remarkably, so that a stable high-performance gain is obtained for a multi-user system in a typical indoor environment; a more robust transmission link is formed for a plurality of users in the typical indoor environment; and the link bit error rate performance of a communication system is improved effectively.

Description

Precoding MIMO-OOFDM-VLC imaging type communication means based on PDS

Technical field

The present invention is towards visible light communication (VisibleLightCommunication, VLC) field, propose a kind of based on dynamic optical receiver selection (PhotodetectorSelection, PDS) multi-User Indoor precoding (Multi-userPrecoded, MUP) multiple-input and multiple-output (Multiple-inputMultiple-output, MIMO) the imaging type VLC communication means of light OFDM (OpticalFrequencyDivisionMultiplexing, OOFDM).

Background technology

Along with the development of technology, adopting the illuminator of light emitting diode (Light-emittingDiode, LED) to obtain increasingly extensive application, the VLC technology based on LED have also been obtained paying close attention in a large number of researcher. VLC technology is based primarily upon intensity modulated (IntensityModulation, IM) and directly detects (DirectDetection, DD) technology, and it utilizes the human eye fast-changing light intensity of insignificant power to carry out the transmission of information. In traditional wireless communication, wide variety of MIMO technology also obtains application by its advantage in System Expansion etc. in VLC technology, by utilizing multiple LED to realize high speed data transfer simultaneously, improves the transfer rate of VLC system. Therefore, MIMO and VLC technology combines and is increasingly becoming the study hotspot of current high speed VLC communication system.

Multiuser MIMO (MU-MIMO) technology, as the extension of MIMO technology and application, is widely used in Modern wireless communication field. Such as, Long Term Evolution (LongTermEvolution, LTE) system just have employed MU-MIMO technology. In recent years, MU-MIMO technology is used for VLC system also become the effective means of a kind of VLC of lifting systematic function and receive the concern of industry. The situation belonging to same user terminal with all photodetectors (Photodetector, PD) in common MIMO-VLC system is different, and multiuser MIMO-VLC system supports multiple user terminals, and each terminal comprises one or more PD. MU-MIMO system eliminates inter-user interference (Multi-UserInterference, MUI) by specific interference cancellation algorithm so that multiple user terminals can utilize identical LED light array to communicate without interfering with each other.

Although multi-user system being conducted extensive research in traditional wireless communication, but for the design of indoor MU-MIMO-VLC system, it is necessary to consider the special nature of indoor VLC channel, not yet studied fully at present.Current indoor MU-MIMO-VLC systematic difference also exists a major issue, namely in order to meet the requirement of user terminal miniaturization, with usual apart from close between the different PD of a user terminal in MU-MIMO system, and the channel gain between LED with the PD of close together is generally similar, this can introduce serious channel relevancy, making the performance dramatic drop-off of MU-MIMO-VLC system, this is also the difficult point that current indoor VLC application is encountered. Accordingly, it is considered to MU-MIMO-VLC combines effectively to reduce the dependency of MU-MIMO channel with other technology, thus obtaining systematic function more preferably, development and the application of VLC technology are had important realistic meaning by this.

Summary of the invention

Existing indoor MU-MIMO-VLC technology has been extended by the present invention, receives and Receiving antenna selection technology in conjunction with imaging type, it is proposed to a kind of precoding MIMO-OOFDM-VLC imaging type communication means based on PDS.

For solving above-mentioned technical problem, technical scheme is as follows:

A kind of precoding MIMO-OOFDM-VLC imaging type communication means based on PDS, comprises the following steps:

Step 1: the complete or collected works of all PD in definition multi-user system are ��_p, each user is furnished with M_r,jIndividual PD, j=1,2 ..., K, K represents the user terminal sum in room; M from each user_r,jIndividual PD selects N_r,jDuring individual PD, constitute one and compriseThe alternative PD of individual PD activates collection, is designated as m_p, have m_p�ʦ�_p;

Step 2: each the alternative PD for producing according to step 1 activates collection m_p, calculate its scale factor

Wherein:WithRepresent that a certain alternative PD activates collection m respectively_pThe channel matrix of corresponding jth user and pre-coding matrix, the product of the two is the equivalent channel matrix of user j;Represent according to m_pCalculate user j equivalent channel matrix minimum singular value square;Represent and activate collection m_pDC loss parameter;

This step is assumed that the channel information of transmitting terminal can be fed back by channel estimating and dedicated signaling and is obtained;

Step 3: select best alternative PD to activate collection m according to following criterion_p:

$m_{p,select}=\underset{m_p\∈{\mathrm{\Ω}}_p}{\mathrm{argmax}}{\mathrm{\κ}}_{m_p,\min }---\left(12\right)$

Collection m is activated according to the PD selected_p,selectAnd (the N of correspondence_r��N_t) dimension channel transfer matrices H, namely can determine that PD that each user activates this symbol period planted agent and carry out information transmission, the result that PD selects can be passed through to control channel transmission PDS configuration signal to each user, receiver user activate corresponding PD accordingly;

Each user needs the bit data flow of transmission to generate the symbol data of corresponding modulating mode through manipulator, obtains (the N after jth user modulation_r,j�� 1) dimension data vector u_j, it is mapped to N again through precoding_tIn individual LED, use method for precoding that the multiple user signals after PDS processes is carried out precoding, obtain following (N_t�� 1) dimension frequency domain data vector:

$f={(f_1,\mathrm{...},f_{N_t})}^H=(P_1,\mathrm{...},P_K){(u_1^H,\mathrm{...},u_K^H)}^H=\underset{j=1}{\overset{K}{\Σ}}\left(P_j{u}_j\right)---\left(13\right)$

Wherein u_jFor the data vector corresponding to user j, P_jFor the pre-coding matrix that user j after PDS is corresponding; After OOFDM modulates, obtain optical signal x (t) in LED further;

At receiving terminal, optical information is received by imaging type optical receiver after VLC dissemination channel, forms hot spot by imaging type lens on the PD activated; Specifically, the reception signal r on PD that q-th activates_qT () is expressed as:

$r_q\left(t\right)=R\underset{i=1}{\overset{N_t}{\Σ}}{h}_{qi}{x}_i\left(t\right)+n_q\left(t\right)---\left(14\right)$

Wherein h_qiRepresent the channel gain between the PD that i-th LED and q-th activate, x_iT () represents the optical signal in i-th LED, n_qT () represents the zero-mean additive white Gaussian noise on q-th PD, R represents the electricity conversion of PD, and the modulation optical signal that LED produces processes through PD and photoelectric conversion, and optical signal is converted into the signal of telecommunication;

Then carry out OOFDM demodulation to received signal and obtain frequency domain data, obtain the equivalent received signals r of jth user_jFor:

$r_j={\mathrm{RH}}_{j}f+n={\mathrm{RH}}_j{P}_j{u}_j+R\underset{j=1,j^{\′}\≠j}{\overset{K}{\Σ}}\left(H_j{P}_{j^{\′}}{u}_{j^{\′}}\right)+n={\mathrm{RU}}_j{\mathrm{\Λ}}_j{u}_j+n---\left(15\right)$

Wherein n represents zero-mean additive white Gaussian noise vector; Use matrix U_jConjugate transpose to obtain signal process, obtain user j data estimation vectorAs follows:

${\tilde{u}}_j=U_j^H{r}_j={\mathrm{RU}}_j^H{U}_j{\mathrm{\Λ}}_j{u}_j+U_j^{H}n={\mathrm{R\Λ}}_j{u}_j+U_j^{H}n---\left(16\right)$

Processing through above, the mimo channel of each user is broken down into multiple SISO channel, obtains estimate vector with maximum-likelihood demodulation

Compared with prior art, technical solution of the present invention provides the benefit that: a kind of precoding MIMO-OOFDM-VLC imaging type communication means based on PDS that the present invention proposes can significantly reduce the VLC channel relevancy between the PD of same user terminal or different user terminals, so that multi-user system obtains stable high performance gains in typical indoor environment, and make multiple user have the transmission link of more robust in typical indoor environment, it is effectively improved the link error rates performance of communication system.

Accompanying drawing explanation

Fig. 1 is based on the multiuser MIMO-OOFDM-VLC imaging type system block diagram of PDS.

Fig. 2 is the imaging type receiver architecture in VLC indoor model and imaging schematic diagram.

Fig. 3 is based on the implementing procedure figure of the multi-user pre-coding MIMO-OOFDM-VLC imaging type system of PDS.

Fig. 4 is user terminal 2 BER Performance comparision schematic diagram of different system when three diverse locations.

Fig. 5 is when signal to noise ratio is 153dB, does not adopt the MU-MIMO-OOFDM imaging type system average BER performance schematic diagram of system with user terminal 2 change in location of PDS.

Fig. 6 is when signal to noise ratio is 153dB, and the MU-MIMO-OOFDM imaging type system based on PDS that the present invention proposes is with the average BER performance schematic diagram of system of user terminal 2 change in location.

Detailed description of the invention

Accompanying drawing being merely cited for property explanation, it is impossible to be interpreted as the restriction to this patent; In order to the present embodiment is better described, some parts of accompanying drawing have omission, zoom in or out, and do not represent the size of actual product;

To those skilled in the art, in accompanying drawing, some known features and explanation thereof are likely to omission and will be understood by. Below in conjunction with drawings and Examples, technical scheme is described further.

Embodiment 1

The focus of the present invention is that imaging receptor and PDS technology are introduced indoor MU-MIMO-VLC system, effectively reduces the channel relevancy between same user or different user receiver PD, thus farthest improving the performance of multiuser MIMO-VLC system. The system block diagram of the present invention is as shown in Figure 1.

Without loss of generality, it is assumed that the ceiling of indoor room is uniformly distributed N_tIndividual LED, has K user terminal in room, wherein jth user terminal is equipped with M_r,jIndividual PD, then the PD total number of this multi-user system isWithin each symbol transmission cycle, the M of jth user_r,jIndividual PD only has N_r,jIndividual PD is activated, then the PD activated in this multi-user system adds up toFig. 2 gives the schematic diagram of above-mentioned typical scene. A kind of detailed description of the invention of the present invention is described for this typical scene below.

First, each user in the present invention is equipped with imaging type optical receiver, can be projected on one or more PD of receiver by the optical signal that transmitting terminal LED array sends by imaging type lens. PD receives after the optical signal of imaging type forming lens, is translated into the signal of telecommunication then through photoelectric conversion, then completes the demodulation of the signal of telecommunication. As in figure 2 it is shown, the imaging type optical receiver that each user terminal is equipped with is made up of imaging type lens and several PD, all PD of each user terminal share same lens.In order to ensure the symmetry that user moves, imaging type lens are arranged on the centre of user's optical receiver. Specifically, the present invention illustrates for the desirable paraxial optics imaging system simplified, but is used as other optical imaging systems with similar functions in reality. In the model, the magnification M of imaging type optical receiver is defined as:

$M=\frac{L}{d_z-L}---\left(1\right)$

Wherein L represents the focal length of lens length in image optics, d_zRepresent the lens place plane vertical dimension to LED place ceiling plane. In desirable paraxial optics imaging model, unrelated through the material elements such as imaging magnification ratio and the angle of incident illumination and optical axis of lens, and consistent with primary source shape and undistorted through the imaging of lens. Adopting imaging type optical receiver, the optical signal in free space is efficiently separated by imaging type lens, provides great space diversity gain for MU-MIMO system, it is possible to decrease the mimo channel dependency between the PD of close together.

For multi-user's VLC system, multi-user pre-coding technology can be adopted to carry out pretreatment to eliminate MUI to launching signal. Block diagonalization method (BlockDiagonalization, BD) is a kind of simple linear pre-coding method, and its decoding complex degree is low, it is possible to be conveniently used in small-sized reception equipment to reduce energy consumption. The ultimate principle of BD method for precoding is the design pre-coding matrix interference for eliminating between user. In the present invention, BD precoding makes the number of data streams q of jth user_jMeet q_j��N_r,j(j=1 ..., K) andUnder the restriction of BD method for precoding, in order to maximally utilize the spatial degrees of freedom of system, the present invention makesAnd q_j=N_r,j(j=1 ..., K), therefore each user can be considered as (N_r,j��N_t) the support N that ties up_r,jThe mimo system of individual data stream, and meet

BD method for precoding concrete grammar is described below: definition comprises except jth user ((N outside_r-N_r,j)��N_t) Wesy's family channel matrixH_jRepresent (the N of jth user_r,j��N_t) dimension channel matrix. Definition rank of matrix isUse singular value decomposition (SingularValueDecomposition, SVD) method that matrix is carried out decomposition to obtain:

${\tilde{H}}_j={\tilde{U}}_j{\widetilde{\mathrm{\Λ}}}_j{\left({\tilde{V}}_j^{\left(1\right)}\right|{\tilde{V}}_j^{\left(0\right)})}^H---\left(2\right)$

Wherein ((N_r-N_r,j)��(N_r-N_r,j)) dimension matrixContain whole left singular vector, ((N_r-N_r,j)��N_t) dimension matrixRepresent singular value diagonal matrix,Dimension matrixBefore comprisingIndividual right singular vector,Dimension matrixContain remaining right singular vector. Usual channel matrix is to have N in row full rank and the present invention_t=N_r, therefore haveThen the equivalent channel matrix defining user j isApplication SVD method continues that matrix is carried out decomposition and obtains:

${\stackrel{\‾}{H}}_j=H_j{\tilde{V}}_j^{\left(0\right)}=U_j{\mathrm{\Λ}}_j{V}_j^{\left(1\right)H}---\left(3\right)$

�� in formula (3)_jIt is (N_r,j��N_r,j) the singular value diagonal matrix tieed up, U_jIt is (N_r,j��N_r,j) matrix for finally demodulating signal tieed up, (N_r,j��N_r,j) matrix tieed upContain right singular vector. Finally obtain (the N of jth user_t��N_r,j) dimension pre-coding matrix P_j, it is calculated as follows:

$P_j={\[p_{1,j}^H,\mathrm{...},p_{N_t,j}^H\]}^H={\tilde{V}}_j^{\left(0\right)}{V}_j^{\left(1\right)}---\left(4\right)$

Wherein p_i,jRepresent jth user pre-coding matrix P_jIn (1 �� N_r,j) tie up the i-th every trade vector.

Without loss of generality, frequency domain data f after the precoding that each LED is exported by multi-user's VLC system employing direct current light offset orthogonal frequency division multiplexing (DC-biasedOpticalOrthogonalFrequencyDivisionMultiplexing, DCO-OFDM) in the present embodiment is modulated. In real system, it is possible to use other OOFDM modulation system, such as ACO-OFDM etc. For producing to meet the real-valued OOFDM time-domain signal of VLC transmission condition, the frequency domain modulation data point in DCO-OFDM system to meet Hermitian conjugate symmetry, namely assumes have N number of carrier wave and frequency domain modulation data X=[X₀,X₁,X₂,��,X_N-1], should meetAnd X₀=X_N/2=0.Frequency domain data f in i-th LED_iOFDM manipulator is modulated on each carrier wave, after inversefouriertransform (InverseFastFourierTransform, IFFT), frequency domain data is converted into the time-domain signal x of correspondence_0,iT (), required direct current biasing is sized toWherein 10log₁₀(��²+ 1) [dB] represents given direct current biasing intensity. x_0,iT () is one has and just has negative real-valued signal, and meets E{x_0,i(t) }=0. The non-negative real transmitting signal x of VLC transmission requirement is finally met after adding direct current biasing and clipping operation_i(t)��x_0,i(t)+B_DC,i, wherein B_DC,iRepresent the size of direct current biasing. Therefore i-th LED launches being desired for of signal:

E{x_i(t)}��E{x_0,i(t)+B_DC,i}=E{x_0,i(t)}+E{B_DC,i}=B_DC,i(5)

When direct current biasing is sufficiently large, clipped noise is less on the desired impact of signal, it is believed that have E{x_i(t) }=B_DC,i. In the VLC system adopting IM/DD technology, the mathematic expectaion of the light emitted signal of LED that is to say the average emitted luminous power of this lamp, and the average luminescence watt level obtaining i-th LED according to formula (5) is P_opt.i=B_DC,i. Under certain direct current biasing benchmark, during such as 7dB and 13dB, the size of direct current biasing is directly proportional to the variance of signal, therefore traditionally the average luminescence power of i-th LED and x under DCO-OFDM direct current biasing interpolation standard_0,iT () is correlated with, x on each LED_0,iT () different variance makes the luminous intensity of every LED under this standard different. And LED is generally symmetrically arranged in indoor ceiling in reality, in order to ensure indoor Uniform Illumination it is generally required to the average light power P that sends of each LED_opt,i(i=1,2 ..., N_t) identical, therefore under certain direct current biasing intensity benchmark, in order to reduce due to the impact of the DCO-OFDM clipped noise caused as far as possible, MU-MIMO-OOFDM-VLC system should meet on the basis of the required maximum LED of direct current biasing, other LED is adopted identical direct current biasing, is also disposed as by the direct current biasing size of all lampsSo while satisfied indoor Uniform Illumination, can also reach the purpose of VLC information transmission. Finally obtain all N_tTransmitting optical signal x (t) in individual LED is:

$x\left(t\right)={\left(x_1\right(t),\mathrm{...},x_{N_t}(t\left)\right)}^H---\left(6\right)$

It is consistent that they meet luminous power, namely:

$P_{opt}=E\left\{x_1\left(t\right)\right\}=E\left\{x_2\left(t\right)\right\}=,\mathrm{...},=E\left\{x_{N_t}\left(t\right)\right\}=B_{DC,ma:}---\left(7\right)$

User side initial data of the present invention adopts bipolarity ON-OFF keying modulation (On-offKeying, OOK). Assuming that the mean power of each user every data stream symbol is equal, the frequency domain data average electrical power after precoding that convolution (4) can obtain in i-th LED is:

$E\left\{f_i^2\right\}=\underset{j=1}{\overset{j=K}{\Σ}}{p}_{i,j}{p}_{i,j}^H\·a=Z_{i}a---\left(8\right)$

WhereinA represents the mean power of bipolarity OOK symbol on the every data stream of each user. Therefore, N_tOn individual LED, the maximum direct current biasing needed for DCO-OFDM is:

$\begin{array}{c}{B}_{DC,\max }=\mathrm{\ξ}\sqrt{\max \left(E\right\{f_1^2\},E\{f_2^2\},\mathrm{...},E\{f_{N_t}^2\left\}\right)}\\ =\sqrt{\max (Z_1,Z_2\mathrm{...},Z_{N_t}){\mathrm{\ζ}}^{2}a}\\ =\sqrt{va}\end{array}---\left(9\right)$

Here definition DC loss factor:

$v=max(Z_1,Z_2\·\·\·,Z_{N_t}){\mathrm{\ζ}}^2---\left(10\right)$

It represents the DC loss factor that system produces due to OOFDM operation, and its value size can change the energy utilization efficiency of system, thus systematic function is produced impact.

Based on above-mentioned BD precoding and OOFDM implementing procedure, the PDS scheme that the present invention introduced below proposes. For maximally utilizing spatial degrees of freedom, it is assumed thatAnd q_j=N_r,j(j=1 ..., K), the PD number M that each user is equipped with_r,jMore than N_r,j. Therefore, when system transfers, each user must activate M_r,jN in individual PD_r,jChange that individual PD, the PD being activated move with customer location and it may happen that change. PDS design in the present invention, based on maximization minimum singular value (MaximumMinimumSingularValue, the MMSV) method improved, is implemented as follows:

Step 1: define all M in this multi-user system_rThe complete or collected works of individual PD are ��_p.M from each user_r,jIndividual PD selects N_r,jDuring individual PD, may make up one and compriseThe alternative PD of individual PD activates collection, is designated as m_p, have m_p�ʦ�_p��

Step 2: each the alternative PD for producing according to step 1 activates collection m_p, calculate scale factor

Wherein:Represent that a certain certain candidate PD activates collection m respectively_pThe corresponding channel matrix of jth user, pre-coding matrix, the product of the two is the equivalent channel matrix of user j;Represent according to m_pCalculate user j equivalent channel matrix minimum singular value (MinimumSingularValue, MSV) square;The DC loss parameter that expression (10) defines. This step is assumed that the channel information of transmitting terminal can be fed back by channel estimating and dedicated signaling and is obtained.

Step 3: select best alternative PD to activate collection m according to following criterion_p:

$m_{p,select}=\underset{m_p\∈{\mathrm{\Ω}}_p}{\mathrm{argmax}}{\mathrm{\κ}}_{m_p,\min }---\left(12\right)$

Collection m is activated according to the PD selected_p,selectAnd (the N of correspondence_r��N_t) dimension channel transfer matrices H, namely system can determine that PD that each user activates this symbol period planted agent and carries out information transmission. The result that PD selects can be passed through system control channels and send PDS configuration signal to each user, receiver user activate corresponding PD accordingly.

Each user needs the bit data flow of transmission to generate the symbol data (such as bipolarity OOK symbol) of corresponding modulating mode through manipulator, obtains (the N after jth user modulation_r,j�� 1) dimension data vector u_j, it is mapped to N again through precoding_tIn individual LED. Use BD method for precoding that the multiple user signals after PDS processes is carried out precoding, obtain following (N_t�� 1) dimension frequency domain data vector:

$f={(f_1,\mathrm{...},f_{N_t})}^H=(P_1,\mathrm{...},P_K){(u_1^H,\mathrm{...},u_K^H)}^H=\underset{j=1}{\overset{K}{\Σ}}\left(P_j{u}_j\right)---\left(13\right)$

Wherein u_jFor the data vector corresponding to user j, P_jFor the pre-coding matrix that user j after PDS is corresponding. After OOFDM modulates, optical signal x (t) in the LED shown in formula (6) can be obtained further. At receiving terminal, optical information is received by imaging type optical receiver after VLC dissemination channel, forms hot spot by imaging type lens on the PD activated. Specifically, the reception signal on PD that q-th activates can be expressed as:

$r_q\left(t\right)=R\underset{i=1}{\overset{N_t}{\Σ}}{h}_{qi}{x}_i\left(t\right)+n_q\left(t\right)---\left(14\right)$

Wherein h_qiRepresent the channel gain between the PD that i-th LED and q-th activate, x_iT () represents the optical signal in i-th LED, n_qT () represents the zero-mean additive white Gaussian noise (AdditionWhiteGaussianNoise, AWGN) on q-th PD, R represents the electricity conversion of PD. The modulation optical signal that LED produces processes through PD and photoelectric conversion, and optical signal is converted into the signal of telecommunication. Then carry out OOFDM demodulation to received signal and obtain frequency domain data, obtain the equivalent received signals r of jth user_jFor:

$r_j={\mathrm{RH}}_{j}f+{\mathrm{nRH}}_j{P}_j{u}_j+R\underset{j^{\′}=1,j^{\′}\≠j}{\overset{K}{\Σ}}\left(H_j{P}_{j^{\′}}{u}_{j^{\′}}\right)+n={\mathrm{RU}}_j{\mathrm{\Λ}}_j{u}_j+n---\left(15\right)$

Wherein n represents zero-mean additive white Gaussian noise vector. Use the matrix U that formula (3) generates_jConjugate transpose to obtain signal process, obtain user j data estimation vectorAs follows:

${\tilde{u}}_j=U_j^H{r}_j={\mathrm{RL}}^H{U}_j{\mathrm{\Λ}}_j{u}_j+U_j^{H}n={\mathrm{R\Λ}}_j{u}_j+U_j^{H}n---\left(16\right)$

Processing through above, the mimo channel of each user is broken down into multiple SISO channel, and we can use maximum-likelihood demodulation (MaximumLikelihoodDetector, MLD) to obtain estimate vector

Designing based on said system, Fig. 3 gives the main implementing procedure of the present embodiment.

By setting forth that the present invention is had the advantage that more fully, below in conjunction with simulation analysis and result, further effectiveness of the invention and advance are explained. Analogue system chooses typical indoor room model, and ceiling is symmetrical arranged N_t=4 LED, user terminal number K=2, each user terminal is equipped with 4 PD namely M_r,1=M_r,2=4.Two data streams of each User support, namely q₁=q₂=2, it is assumed that each user activates 2 PD namely N in any time_r,1=N_r,2=2. In order to meet the demand of device miniaturization, it is set to 1.01cm with the distance between the PD of a user terminal. It is plane coordinates initial point that analogue system chooses room centre. Average height according to common people, it is assumed that terminal is all located in the plane that overhead 0.85m is high. In model use signal to noise ratio definition with document " T.FathandH.Haas; " PerformanceComparisonofMIMOTechniquesforOpticalWirelessC ommunicationsinIndoorEnvironments; " IEEETransactionsonCommunications, vol.61, no.2, pp.733 742, Feb.2013. " consistent, OOFDM modulating part uses 13dB direct current biasing intensity.

Fig. 4 gives PDS-MU-MIMO-OOFDM imaging type system at the user terminal 2 BER performance curve when three diverse locations, and compares with the performance of non-imaging type system and the imaging type system not adopting PDS. Wherein, user terminal 1 is fixed in the middle of room (0,0,0.85), 3 positions respectively (0.2,0.2,0.85) of user terminal 2, (0.9,0.9,0.85) and (1.8,1.8,0.85). As shown in Figure 4, compared with non-imaging type system, imaging type system utilizes the diversity gain of imaging type lens to make the BER performance of system obtain and significantly improves. Such as, when target BER is 10^-3Time, compared with non-imaging type system, imaging type system can obtain the performance gain up to 45dB, 48dB and 18dB respectively in position 1,2 and 3. Further, compared with the imaging type system not adopting PDS, the imaging type system based on PDS that the present invention proposes can obtain the performance gain of 15dB, 5dB and 3dB further respectively three positions. This shows that this programme has merged imaging type receptor and the advantage of PDS technology effectively, and the hybrid system structure formed can be substantially improved the transmission reliability of multiuser MIMO-VLC communication system.

Fig. 5 and Fig. 6 provides further and does not adopt PDS and based on the MU-MIMO-OOFDM imaging type system of PDS with the average BER performance comparison of system of user terminal 2 change in location. User terminal 1 in scene I and scene II is fixing respectively to be positioned on the position of (0,0,0.85) and (0.6,0.6,0.85), and user terminal 2 then moves freely in room model. As shown in Figure 5, in the imaging type system not using PDS, when user terminal 2 is near user terminal 1, owing to the channel relevancy between different user increases, result in the sharply deterioration of systematic function. By contrast, the high bit-error region in Fig. 5 can effectively be removed based on the imaging type system of PDS, it is thus achieved that the raising of overall system performance.

The foregoing is only embodiments of the invention, be not limited to the present invention. The present invention can have various suitable change and change. All make within the spirit and principles in the present invention any amendment, equivalent replacement, improvement etc., should be included within protection scope of the present invention.

Claims (3)

1. precoding MIMO-OOFDM-VLC imaging type the communication means based on PDS, it is characterised in that comprise the following steps:

Step 1: the complete or collected works of all PD in definition multi-user system are ��_p, each user is furnished with M_r,jIndividual PD, j=1,2 ..., K, K represents the user terminal sum in room; M from each user_r,jIndividual PD selects N_r,jDuring individual PD, constitute one and compriseThe alternative PD of individual PD activates collection, is designated as m_p, have m_p�ʦ�_p;

Step 2: each the alternative PD for producing according to step 1 activates collection m_p, calculate its scale factor

Wherein:WithRepresent that a certain alternative PD activates collection m respectively_pThe channel matrix of corresponding jth user and pre-coding matrix, the product of the two is the equivalent channel matrix of user j;Represent according to m_pCalculate user j equivalent channel matrix minimum singular value square;Represent and activate collection m_pDC loss parameter;

This step is assumed that the channel information of transmitting terminal can be fed back by channel estimating and dedicated signaling and is obtained;

Step 3: select best alternative PD to activate collection m according to following criterion_p:

$m_{p,select}=\underset{m_p\∈{\mathrm{\Ω}}_p}{\mathrm{argmax}}{\mathrm{\κ}}_{m_p,\min }---\left(12\right)$

Collection m is activated according to the PD selected_p,selectAnd (the N of correspondence_r��N_t) dimension channel transfer matrices H, namely can determine that PD that each user activates this symbol period planted agent and carry out information transmission, the result that PD selects can be passed through to control channel transmission PDS configuration signal to each user, receiver user activate corresponding PD accordingly;

Each user needs the bit data flow of transmission to generate the symbol data of corresponding modulating mode through manipulator, obtains (the N after jth user modulation_r,j�� 1) dimension data vector u_j, it is mapped to N again through precoding_tIn individual LED, use method for precoding that the multiple user signals after PDS processes is carried out precoding, obtain following (N_t�� 1) dimension frequency domain data vector:

Wherein u_jFor the data vector corresponding to user j, P_jFor the pre-coding matrix that user j after PDS is corresponding; After OOFDM modulates, obtain optical signal x (t) in LED further;

At receiving terminal, optical information is received by imaging type optical receiver after VLC dissemination channel, forms hot spot by imaging type lens on the PD activated; Specifically, the reception signal r on PD that q-th activates_qT () is expressed as:

Wherein h_qiRepresent the channel gain between the PD that i-th LED and q-th activate, x_iT () represents the optical signal in i-th LED, n_qT () represents the zero-mean additive white Gaussian noise on q-th PD, R represents the electricity conversion of PD, and the modulation optical signal that LED produces processes through PD and photoelectric conversion, and optical signal is converted into the signal of telecommunication;

Then carry out OOFDM demodulation to received signal and obtain frequency domain data, obtain the equivalent received signals r of jth user_jFor:

Wherein n represents zero-mean additive white Gaussian noise vector; Use matrix U_jConjugate transpose to obtain signal process, obtain user j data estimation vectorAs follows:

Processing through above, the mimo channel of each user is broken down into multiple SISO channel, obtains estimate vector with maximum-likelihood demodulation

2. method according to claim 1, it is characterised in that be use BD method for precoding that the multiple user signals after PDS processes is carried out precoding in step 3;

Wherein BD method for precoding particularly as follows:

Definition comprises except jth user ((N outside_r-N_r,j)��N_t) Wesy's family channel matrixH_jRepresent (the N of jth user_r,j��N_t) dimension channel matrix, definition rank of matrix isUse singular value decomposition (SVD) method that matrix is carried out decomposition to obtain:

Wherein ((N_r-N_r,j)��(N_r-N_r,j)) dimension matrixContain whole left singular vector, ((N_r-N_r,j)��N_t) dimension matrixRepresent singular value diagonal matrix,Dimension matrixBefore comprisingIndividual right singular vector,Dimension matrixContain remaining right singular vector, Qi ZhongyouThen the equivalent channel matrix defining user j isApplication SVD method continues that matrix is carried out decomposition and obtains:

�� in formula (3)_jIt is (N_r,j��N_r,j) the singular value diagonal matrix tieed up, U_jIt is (N_r,j��N_r,j) matrix for finally demodulating signal tieed up, (N_r,j��N_r,j) matrix tieed upContain right singular vector, finally obtain (the N of jth user_t��N_r,j) dimension pre-coding matrix P_j, it is calculated as follows:

Wherein p_i,jRepresent jth user pre-coding matrix P_jIn (1 �� N_r,j) tie up the i-th every trade vector.

3. method according to claim 2, it is characterized in that, wherein user terminal initial data adopts bipolarity ON-OFF keying modulation, if the mean power of each user every data stream symbol is equal, convolution (4) can obtain the frequency domain data average electrical power after precoding in i-th LEDFor:

WhereinA represents the mean power of bipolarity OOK symbol, then N on the every data stream of each user_tMaximum direct current biasing B needed for DCO-OFDM on individual LED_DC,maxFor:

Here definition DC loss factor: