CN109412658A - A kind of improved B B search tree detection method based on shade domain - Google Patents
A kind of improved B B search tree detection method based on shade domain Download PDFInfo
- Publication number
- CN109412658A CN109412658A CN201811383148.2A CN201811383148A CN109412658A CN 109412658 A CN109412658 A CN 109412658A CN 201811383148 A CN201811383148 A CN 201811383148A CN 109412658 A CN109412658 A CN 109412658A
- Authority
- CN
- China
- Prior art keywords
- vector
- search tree
- improved
- shade
- detection method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/0413—MIMO systems
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/391—Modelling the propagation channel
- H04B17/3912—Simulation models, e.g. distribution of spectral power density or received signal strength indicator [RSSI] for a given geographic region
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/32—Carrier systems characterised by combinations of two or more of the types covered by groups H04L27/02, H04L27/10, H04L27/18 or H04L27/26
- H04L27/34—Amplitude- and phase-modulated carrier systems, e.g. quadrature-amplitude modulated carrier systems
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/32—Carrier systems characterised by combinations of two or more of the types covered by groups H04L27/02, H04L27/10, H04L27/18 or H04L27/26
- H04L27/34—Amplitude- and phase-modulated carrier systems, e.g. quadrature-amplitude modulated carrier systems
- H04L27/36—Modulator circuits; Transmitter circuits
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Electromagnetism (AREA)
- Digital Transmission Methods That Use Modulated Carrier Waves (AREA)
Abstract
The present invention proposes a kind of improved B B search tree detection method based on shade domain;The unreliable symbol for falling into shadow region is obtained optimal integer solution with branch and bound search tree method after the step qp method of progress by present invention combination qp and branch-bound method;Complexity is reduced with suitable trimming strategy in search tree;Under extensive MIMO scene, the improved B B search tree detection method proposed in this paper based on shade domain not only shows good performance but also is suitable for high order modulation.
Description
Technical field
The improved B B search tree detection method based on shade domain that the present invention relates to a kind of is mainly used in extensive how defeated
Enter multi output (MIMO) scene, belongs to mobile communication technology field.
Background technique
Currently, the correlative study of the 5th Generation Mobile Communication System (5G) is being actively developed, and the shape in 2018
At first edition 5G standard.Wherein, extensive MIMO technology is the hot spot of research and discussion.It is a large amount of by being used in base station side
Dual-mode antenna, extensive mimo system can use additional freedom degree, the multiple data flows of parallel transmission, while improve diversity increasing
Benefit, so as to the energy efficiency for greatly increasing the availability of frequency spectrum, improving transmission reliability and improving system.
Since base station uses a large amount of dual-mode antenna, the high-performance detection used under suitable large scale scene how is designed
Algorithm is at the significant challenge faced in application.Although many detection methods have been proposed for traditional mimo system, such as
MMSE (Minimum Mean Square Error, least mean-square error), ML (Maximum Likelihood, maximum likelihood),
(zero) Zero Forcing is broken, spherical decoding method etc. ZF, but these detection methods are with antenna under extensive MIMO scene
Its detection performance of the increase of number is not optimal.Based on problem above, proposed in extensive mimo system new
Detection method such as RTS (Reactive Tabu Search, initiative TABU search), LAS (LikelihoodAscent
Search, likelihood rise search) detection method, the LBTS based on BB (Branch andbound, branch-and-bound) search tree
(Likelihood-Based Tree Search, maximum likelihood searching tree) method and 2qp (2-stage quadratic
Programming, the order two planning) detection method.These new detection algorithms compare its performance with detection algorithm before
It is promoted, still, in the extensive mimo system that number of antennas increases, these new algorithms are in high order modulation and higher noise
Than under scene, it is higher that detection performance is unable to reach optimal and its detection complexity.
Summary of the invention
Based on the above issues, the invention proposes a kind of modifieds based on shade domain applied to extensive MIMO scene
BB search tree detection method realizes BER (Bit ErrorRate, bit error rate) better performances and is suitable for high-order orthogonal width
Degree modulation, while greatly reducing computation complexity.
The present invention proposes a kind of improved B B search tree detection method based on shade domain, as shown in Figure 1, including following step
It is rapid:
S1: using known channel matrix and symbol construction qp (quadratic programming, quadratic programming) is sent
Model;
S2: with shade domain algorithm to qp model carry out solve respectively obtain fall into the unreliable symbol in shade domain vector and
The vector of the reliability symbols in shade domain is not fallen within;
S3: branch is carried out using branch-bound method to the unreliable symbol for falling into shade domain, and with suitable trimming plan
Slightly obtain the detection vector of needs;
S4: it will carry out integrating and being detected after demodulation with obtained detection vector after the vector quantization of reliability symbols
As a result.
Present invention assumes that the transmission antenna number of mimo system is NT, receiving antenna number is NR, and is using M rank just in receiving end
The modulation of friendship amplitude;
Preferably, the qp model in S1 is to be obtained ML model conversation based on relaxation factor;
Preferably, the qp model in S1 are as follows:
It is constrained in
Preferably, the unreliable symbol that shade domain is fallen into described in S2 be will solve solution vector that qp model obtains and its most
The absolute value of close integer obtains compared with threshold delta compared to obtained difference, is considered if difference is greater than threshold value
It is insecure, it is otherwise reliable.
Preferably, the value range of the threshold delta is 0.2-0.3.
Further, branching method described in step S3 are as follows: using first unreliable signal as first node branch
Variable simultaneously carries out branch, further solves the branch problem sees the solution vector for whether obtaining meeting integer condition, if do not obtained
It then needs this layer of all nodes carrying out branch.
Preferably, trimming strategy described in S3 is depth trimming, and width trimming and the approximation based on cost function value are repaired
It cuts.
Method of the invention not only shows preferable performance in low-order-modulated, but also is suitable for high order modulation,
When number of antennas is more, performance is still good, is also greatly reduced computation complexity.
Detailed description of the invention
Fig. 1 is the improved B B search tree detection method process signal under the extensive MIMO scene of the present invention based on shade domain
Figure;
Fig. 2 is NT=32, NR=32, when modulation system is 64QAM, the performance comparison figure of the present invention and the prior art.
Specific embodiment
In order to which the purpose, technological means and advantage of the application is more clearly understood, the application is done below in conjunction with attached drawing
It is further described.
64QAM in order to better illustrate the specific implementation step of this method, under MIMO scene, when with NT=NR=32
It is illustrated for modulation.
Improved B B search tree detection method embodiment under extensive MIMO scene proposed by the present invention based on shade domain,
Include:
Step 1: in mimo systems, it is assumed that transmission antenna number is NT, and receiving antenna number is NR, and channel is static flat
Fading channel, transmitting terminal will transmit symbolic vector x through M rank QAM (Quadrature amplitude modulation, it is orthogonal
Amplitude modulation) after mode modulates, transferred out by wireless channel H;H indicates the gain matrix of static flat fading channel, clothes
It is just being distributed very much from standard.
Step 2: received vector is received in receiving end, converts real value domain ML model for complex value domain ML method model, and
It is converted into optimization problem, using relaxation factor, Optimized model is converted to qp model.
The detailed process of the ML model conversion optimization problem of real value described in step 2 is:
Pass through formulaBy real number setIt is converted to whole
Manifold is closedChannel matrix H is converted into positive semidefinite matrix Q, Q=HTH, real value ML detection model
It is converted toWherein,
Further, with relaxation thought, problem is converted into qp model:It is constrained inM is the constellation size of QAM modulation, I=[1,1,1 ..., 1]TIt is the column vector of 2NT × 1,0 table
Showing the full null vector of 2NT × 1, the lower limit that transmitting terminal vector x is converted to each element (symbol) of positive integer vector z is 0,
The upper limit isY indicates the data vector that receiving end receives;
Step 3: qp model is solved, with shade domain thought, obtain the unreliable symbol for falling into shade domain to
Vector γ, the η for measuring and not falling within the reliability symbols in shade domain, to the reliability symbols vector η amount of progress for not falling within shade domain
Change;
Qp model in step 3 is solved using interior point method.After obtaining its solution vector to qp model solution, solution vector is judged
Element whether fall into shade domain, the specific standards in shade domain are: the absolute value of solution vector element and its immediate integer
Whether difference is greater than a threshold delta (range are as follows: 0.2~0.3), then think the corresponding symbol of the element if more than this threshold value
It is insecure, is reliable otherwise.
Step 4: use for reference branch-bound method, first symbol in shade domain will be fallen into as root node, d=1, by node into
Row branch, each node can be divided into the condition node of two mutual exclusions, specific expression formula are as follows:
It is constrained inOr Represent the solution vector of node 0;
Step 5: whether solution d node layer subproblem, the cost function value for calculating each node meet Integer constrained characteristic, use
Trimming strategy, retains this layer and optimal solution most possibly occurs and obtain M node.
Trimming strategy includes: 1) width trimming, and the preceding M symbol of optimal solution is occurred in most probable in every layer with M method
Progress branch again is extracted, other nodes of this layer are deleted;2) depth is trimmed: as long as one layer of decline, at least one node is more
Close to optimal integer solution, suitable depth is selected, complexity can be reduced;3) the approximation strategy based on cost function value, this
Strategy depends on the difference of the cost function value of the cost function value of relaxation factor and the integer vectors of quantization, and difference is smaller, closely
Like continuous solution closer to integer solution.For accessed node m, decision threshold is provided, judges the difference of cost function, if meeting item
Part is considered as and has found optimal solution, this can be considered a kind of trimming strategy.
Optimal continuous solution for node m isIts cost function value isWherein m=0,1,2 ..., Nv,
NvIt is accessed node quantity, while defines the optimal continuous solution of quantization in identical position, usesIt indicates, corresponding cost
Functional value isApproximate solution are as follows:
It is the value of a very little greater than zero, passes throughA balance can be done between performance and complexity,More
Greatly, complexity is lower, and performance is poorer, standardEqual to 0.
Step 6: number of plies d adds 1, if meets depth capacity, 7 is entered step if meeting, if being unsatisfactory for returning to step 5;
Step 7: from finding optimal solution in this node layer in solution vector and quantify, the detection after the quantization obtained with step 2
Vector integration, demodulation obtain the result of signal detection.
In order to embody the method for the present invention performance advantage, modulation system 64QAM, as shown in Fig. 2, by MMSE, qp, 2qp,
BB and method proposed in this paper (the Improved-BB curve in Fig. 2) carry out performance comparison.Because qp method does not account for examining
The reliability of symbol is surveyed, 2qp considers reliability, screen to the solution vector of qp, extraction is unsatisfactory on the basis of qp method
The symbol of threshold value, carries out qp detection again, and 2qp (see 2qp curve in Fig. 2) is improved with respect to qp (see qp curve in Fig. 2) method
Reliability.Wherein, δ=0.25 can guarantee that the symbol quantity for falling into shade domain is more appropriate, because of (the δ < when δ is very small
0.1) it, will lead to big quantity symbol and fallen into search tree process, in this case, under especially relatively low state of signal-to-noise, search
Tree process can not optimize the detection performance of first time qp.When δ is very big (δ > 0.4), even if some symbolic distances are nearest
Integer is far, but numerical symbol all it is poor to will lead to detection performance by Integer constrained characteristic mostly;For 64QAM, whenWhen can guarantee that the number of iterations when using interior point method solution node problems is minimum, substantially reduce complexity.Tradition
BB method ratio qp, 2qp and MMSE method performance are more excellent, but improved method performance is stablized herein, as in Fig. 2 in BER=10-4, and (see BB in Fig. 2 (16,2) curve) when d=16, w=2, improved BB search tree method and tradition BB search tree have about
The performance gain of 1dB, there is the performance gain of about 6dB compared with qp method, there is the performance gain of about 4dB compared with 2qp.Together
When, depth increases, and performance is more excellent, for another example: working as d=32, w=4, BER=10-4When (see BB in Fig. 2 (32,4) curve), improve BB
Search tree detection method has the performance gain of about 10dB compared with qp method, there is the performance gain of 7dB compared with 2qp method,
With d=16, the improved BB searching method of w=2 compares the performance gain for having 3dB.Depth increases, and width increases, and performance is more preferable,
This can be weighed between performance and complexity by width and depth taking complexity to exchange performance in fact.
The foregoing is merely preferred embodiments of the present invention, are not intended to limit the invention, all in the method for the present invention
Within principle, any modification, same replacement, improvement for being made etc. be should be included within the scope of the present invention.
Claims (7)
1. a kind of improved B B search tree detection method based on shade domain, which comprises the following steps:
S1: using known channel matrix and symbol construction quadratic programming qp model is sent;
S2: qp model is carried out solving to respectively obtain falling into the vector of the unreliable symbol in shade domain and not having with shade domain algorithm
Fall into the vector of the reliability symbols in shade domain;
S3: branch is carried out using branch-bound method to the unreliable symbol for falling into shade domain, and uses and suitably trims tactful obtain
The detection vector needed out;
S4: the detection vector obtained after the vector quantization of reliability symbols with step 3 integrate and examined after demodulation
Survey result.
2. a kind of improved B B search tree detection method based on shade domain according to claim 1, which is characterized in that step
In rapid S1, the qp model is to be obtained ML model conversation based on relaxation factor.
3. a kind of improved B B search tree detection method based on shade domain according to claim 1, which is characterized in that step
In rapid S1, the qp model are as follows:
It is constrained inQ=HTH,
Wherein, Q indicates positive semidefinite matrix, and M is the constellation size of QAM modulation,0Indicate the full null vector of 2NT × 1, I=
[1,1,1,...,1]TIt is the column vector of 2NT × 1, H indicates the gain matrix of static flat fading channel, obeying standard just
It is distributed very much;Y indicates that the data vector that receiving end receives, NT indicate transmission antenna number.
4. a kind of improved B B search tree detection method based on shade domain according to claim 1, which is characterized in that step
In rapid S2, the unreliable symbol for falling into shade domain includes that will solve the solution vector and its immediate integer that qp model obtains
Absolute value obtained with threshold delta compared with compared to obtained difference, if difference greater than being considered insecure if threshold value.
5. a kind of improved B B search tree detection method based on shade domain according to claim 4, which is characterized in that institute
The value range for stating threshold delta is 0.2-0.3.
6. a kind of improved B B search tree detection method based on shade domain according to claim 1, which is characterized in that S3
The branching method are as follows: first unreliable signal as first node branch variable and is subjected to branch, is further asked
It solves the branch problem sees the solution vector for whether obtaining meeting integer condition, needs to be divided this layer of all nodes if not obtaining
Branch.
7. a kind of improved B B search tree detection method based on shade domain according to claim 1, which is characterized in that S3
The trimming strategy is depth trimming, width trimming and the approximation trimming based on cost function value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811383148.2A CN109412658A (en) | 2018-11-20 | 2018-11-20 | A kind of improved B B search tree detection method based on shade domain |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811383148.2A CN109412658A (en) | 2018-11-20 | 2018-11-20 | A kind of improved B B search tree detection method based on shade domain |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109412658A true CN109412658A (en) | 2019-03-01 |
Family
ID=65474064
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811383148.2A Pending CN109412658A (en) | 2018-11-20 | 2018-11-20 | A kind of improved B B search tree detection method based on shade domain |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109412658A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110971548A (en) * | 2019-12-06 | 2020-04-07 | 重庆邮电大学 | PDSCH signal detection method for 5G system |
CN111371794A (en) * | 2020-03-09 | 2020-07-03 | 北京金睛云华科技有限公司 | Shadow domain detection model, detection model establishing method, detection method and system |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106357312A (en) * | 2016-09-12 | 2017-01-25 | 南京信息工程大学 | Lattice reduction assisted breadth-first tree searching MIMO detecting method |
US20180032964A1 (en) * | 2016-08-01 | 2018-02-01 | Nec Europe Ltd. | Transportation system and method for allocating frequencies of transit services therein |
-
2018
- 2018-11-20 CN CN201811383148.2A patent/CN109412658A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180032964A1 (en) * | 2016-08-01 | 2018-02-01 | Nec Europe Ltd. | Transportation system and method for allocating frequencies of transit services therein |
CN106357312A (en) * | 2016-09-12 | 2017-01-25 | 南京信息工程大学 | Lattice reduction assisted breadth-first tree searching MIMO detecting method |
Non-Patent Citations (2)
Title |
---|
ALI ELGHARIANI 等: "Branch and Bound with M algorithm for near optimal MIMO detection with higher order QAM constellation", 《MILCOM 2012 - 2012 IEEE MILITARY COMMUNICATIONS CONFERENCE》 * |
ALI ELGHARIANI 等: "Low Complexity Detection Algorithms in Large-Scale MIMO Systems", 《 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110971548A (en) * | 2019-12-06 | 2020-04-07 | 重庆邮电大学 | PDSCH signal detection method for 5G system |
CN111371794A (en) * | 2020-03-09 | 2020-07-03 | 北京金睛云华科技有限公司 | Shadow domain detection model, detection model establishing method, detection method and system |
CN111371794B (en) * | 2020-03-09 | 2022-01-18 | 北京金睛云华科技有限公司 | Shadow domain detection model, detection model establishing method, detection method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101039137B (en) | Method and system for reducing codebook search-based precoding feedback bits of MIMO-OFDM system | |
CN105554865B (en) | A kind of MIMO-SCMA system down link design method based on STBC | |
CN106209187B (en) | A kind of self-adapting generalized modulating method using combined optimization | |
CN107302766B (en) | Method for balancing and optimizing energy efficiency and spectral efficiency in distributed antenna system | |
CN105933971B (en) | Energy efficiency optimization method suitable for large-scale multi-input multi-output system | |
CN106060917B (en) | One kind is based on the matched antenna of grouping and power combined allocation method | |
CN109412658A (en) | A kind of improved B B search tree detection method based on shade domain | |
CN105162503B (en) | Multi-user beam forms the co-design method with day line options in a kind of extensive mimo system | |
CN109286587B (en) | Multi-active generalized spatial modulation detection method | |
CN101964667B (en) | High-efficiency multi-antenna detection method for long term evolution scheme | |
Qian et al. | Antenna location design for distributed antenna systems with selective transmission | |
CN101958875B (en) | Detecting method of high order modulated MIMO system in mobile environment | |
CN108337024A (en) | A kind of extensive mimo system efficiency optimization method based on energy acquisition | |
CN112188605A (en) | Design method of network-assisted full-duplex system transceiver in wireless energy-carrying communication scene | |
CN104821840B (en) | A kind of anti-interference method of extensive multiple-input and multiple-output downlink system | |
CN101483467A (en) | Method for MIMO multiple access channel throughput maximization | |
CN110190880A (en) | A kind of incremental emitting antenna combination selection method of generalized space modulation | |
CN103326825B (en) | A kind of quasi-orthogonal space time block code low-complexity decoding method | |
CN109818891A (en) | A kind of lattice about subtracts the low complex degree greed globular decoding detection method of auxiliary | |
Wu et al. | Direct transmit antenna selection for transmit optimized spatial modulation | |
CN109167748A (en) | A kind of part maximum likelihood detection method based on energy ordering | |
CN104038457A (en) | Soft output sphere decoding method in coding MIMO system based on initial sphere radius | |
CN102055566A (en) | Dynamic MIMO (multiple input multiple output) detection algorithm and detector structure based on signal channel analysis | |
CN109347772A (en) | Improved QRD-M detection algorithm in generalized spatial modulation system | |
CN105119869A (en) | Space-shift keying method based on time-space matrix constellation diagram |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190301 |