CN103368621A - Signal processing method and device - Google Patents
Signal processing method and device Download PDFInfo
- Publication number
- CN103368621A CN103368621A CN2012100827890A CN201210082789A CN103368621A CN 103368621 A CN103368621 A CN 103368621A CN 2012100827890 A CN2012100827890 A CN 2012100827890A CN 201210082789 A CN201210082789 A CN 201210082789A CN 103368621 A CN103368621 A CN 103368621A
- Authority
- CN
- China
- Prior art keywords
- vector
- covariance matrix
- matrix
- channel covariance
- weighing
- 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.)
- Granted
Links
Images
Abstract
The invention provides a signal processing method and device. The method includes: acquiring a channel estimation value and acquiring a covariance matrix C of a first uplink channel according to the channel estimation value, wherein the covariance matrix C of the first uplink channel is k-th power of a covariance matrix R of a second uplink channel and the covariance matrix R of the second uplink channel is a covariance matrix generated from the channel estimation value; according to the covariance matrix C of the first uplink channel, an input initial vector V0 (n) and a configured iteration m/k, acquiring a weight vector U; and carrying out a weighting processing on downlink signals according to the weight vector U, wherein n represents the element number in the initial vector V0 (n); m, k and m/k are all positive integers; and m is an initial iteration. The signal processing method and device enable convergence of feature vectors to be speeded up and calculation speed of the feature vectors to be improved so that processing of the whole link is speeded up and streamline realization is facilitated.
Description
Technical field
The present invention relates to the communication technology, relate in particular to a kind of processing method and equipment of signal.
Background technology
Wave beam forming is a kind of closely spaced aerial array multi-antenna transmitting transferring technology that is applied to, its cardinal principle is to utilize the principle of interference of the strong correlation of space channel and ripple to produce the antenna pattern of highly directive, make the main lobe of antenna pattern point to adaptively arrival bearing user, thereby the raising signal to noise ratio, and improve power system capacity or coverage.The algorithm of wave beam forming has multiple, and is wherein fairly simple based on the figuration algorithm of characteristic value and be widely used, and concrete principle is weight coefficient of superposition on transmitted signal, so that the received power of receiving terminal is maximum.
Weight coefficient obtains according to the up channel covariance matrix, and at present, a kind of mode of finding the solution weight coefficient according to the up channel covariance matrix is to adopt power method, namely utilizes formula (1) iteration to carry out computing and tries to achieve final weight coefficient:
U＝R
m*V(n) (1)
Wherein, R represents the up channel covariance matrix, and V (n) represents that n ties up non-zero vector arbitrarily, and m represents the primary iteration number of times.But, when initial iterations m numerical value is larger, treatment facility is often because the restriction of disposal ability, the iterative process of asking for weighing vector (or being called characteristic vector) will spend the more time, computational speed is also slower, thereby the Calculation bottleneck that causes the whole piece link concentrates on this treatment facility iteration and asks for the weighing vector part, and then is unfavorable for the flowing water realization.
Summary of the invention
The embodiment of the invention provides a kind of processing method of signal on the one hand, comprising:
Obtain channel estimation value, and according to described channel estimation value, obtain the first up channel covariance matrix C; Wherein, described the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, and described the second up channel covariance matrix R is the covariance matrix that is generated by described channel estimation value;
According to described the first up channel covariance matrix C, the initial vector V of input
0(n) and the configuration iterations m/k, obtain weighing vector U;
Generate weight coefficient according to described weighing vector U, downstream signal is weighted processing, and send the downstream signal after weighting is processed;
Wherein, n represents described initial vector V
0The number of the element (n), m, k and m/k are positive integer, and m is the primary iteration number of times.
The treatment facility that a kind of signal is provided on the other hand of the embodiment of the invention comprises:
Matrix calculation unit is used for obtaining channel estimation value, and according to described channel estimation value, obtains the first up channel covariance matrix C; Wherein, described the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, and described the second up channel covariance matrix is the covariance matrix that is generated by described channel estimation value;
The weighing vector processing unit is used for according to described the first up channel covariance matrix C, the initial vector V of input
0(n) and the configuration iterations m/k, obtain weighing vector U;
Weighting processing unit is used for generating weight coefficient according to described weighing vector U, and downstream signal is weighted processing, and sends the downstream signal after weighting is processed;
Wherein, n represents described initial vector V
0The number of the element (n), m, k and m/k are positive integer, and m is the primary iteration number of times.
Technique effect of the present invention is: because the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, the iterations that disposes in the process that treatment facility is processed the first up channel covariance matrix has become the 1/k of primary iteration number of times, alleviated the processing load of this processor, accelerated the convergence of characteristic vector, improved the speed of calculated characteristics vector, thereby accelerate the processing speed of whole piece link, and be beneficial to the flowing water realization.
Description of drawings
Fig. 1 is the flow chart of an embodiment of the processing method of signal of the present invention;
Fig. 2 is the flow chart of another embodiment of the processing method of signal of the present invention;
Fig. 3 is the flow chart of another embodiment of the processing method of signal of the present invention;
Fig. 4 is the also flow chart of an embodiment of the processing method of signal of the present invention;
Fig. 5 is the structural representation of an embodiment of the treatment facility of signal of the present invention;
Fig. 6 is the structural representation of another embodiment of the treatment facility of characteristic vector of the present invention;
Fig. 7 is the structural representation of another embodiment of the treatment facility of signal of the present invention;
Fig. 8 is the concrete structure schematic diagram of buf_ctrl.
Embodiment
Fig. 1 is the flow chart of an embodiment of the processing method of signal of the present invention, and as shown in Figure 1, the method for present embodiment comprises:
Alternatively, channel estimation value can pass through the channel condition informations acquisitions such as channel conditions, multidiameter delay, Doppler shift.
In the present embodiment, n represents this initial vector V
0The number of the element (n), m, k and m/k are positive integer, and m is the primary iteration number of times.
In the present embodiment, for instance, work as m=4, and during k=4, according to this channel estimation value, the time of obtaining the first up channel covariance matrix C is 1s, and the time of obtaining weighing vector U is 1s (m/k=1), the time of obtaining weighing vector U in the prior art is 4s (m=4), and the present invention has improved the speed of calculated characteristics vector effectively.
In the present embodiment, because the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, the iterations that disposes in the process that treatment facility is processed the first up channel covariance matrix has become the 1/k of primary iteration number of times, alleviated the processing load of this processor, accelerated the convergence of characteristic vector, improve the speed of calculated characteristics vector, thereby accelerated the processing speed of whole piece link, and be beneficial to the flowing water realization.
Fig. 2 is the flow chart of another embodiment of the processing method of signal of the present invention, and as shown in Figure 2, the method for present embodiment comprises:
V
i+1(n)＝C
m/k*V
i(n) (2)
M/k power and primary vector V with this first up channel covariance matrix C
i(n) carry out matrix multiple and process, obtain secondary vector V
I+1(n).
In the present embodiment, obtain the initial vector V of input
0(n) and primary iteration number of times m, and this primary iteration number of times m is configured processing, obtains the iterations m/k of configuration.In addition, when initialization, i=0, and primary vector V
i(n)=V
0(n).
V
i+1(n)＝V
i+1(n)/|V
i+1(n)| (3)
To this secondary vector V
I+1(n) carry out normalized.Step 203 can reduce signal handling equipment owing to the limited impact that causes resource to be overflowed of signal handling capacity (for example processing bit wide) in actual the use; If do not consider the signal handling capacity of signal handling equipment, but skips steps 203.
In the present embodiment, because the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, the iterations that disposes in the process that treatment facility is processed the first up channel covariance matrix has become the 1/k of primary iteration number of times, alleviated the processing load of this processor, accelerated the convergence of characteristic vector, improve the speed of calculated characteristics vector, thereby accelerated the processing speed of whole piece link, and be beneficial to the flowing water realization.
Fig. 3 is the flow chart of another embodiment of the processing method of signal of the present invention, and as shown in Figure 3, the method for present embodiment comprises:
V
i+1(n)＝C
m/k*V
i(n) (2)
M/k power and and primary vector V with this first up channel covariance matrix C
i(n) carry out matrix multiple and process, obtain secondary vector V
I+1(n).
In the present embodiment, obtain the initial vector V of input
0(n) and primary iteration number of times m, and this primary iteration number of times m is configured processing, obtains the iterations m/k of configuration.In addition, when initialization, i=0, and primary vector V
i(n)=V
0(n).
U＝U/|U| (4)
U carries out normalized to this weighing vector.Step 305 can reduce since signal handling equipment because the limited impact that causes resource to be overflowed of signal handling capacity (for example processing bit wide) in the actual use; If do not consider the signal handling capacity of signal handling equipment, but skips steps 305.
In the present embodiment, because the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, the iterations that disposes in the process that treatment facility is processed the first up channel covariance matrix has become the 1/k of primary iteration number of times, alleviated the processing load of this processor, accelerated the convergence of characteristic vector, improve the speed of calculated characteristics vector, thereby accelerated the processing speed of whole piece link, and be beneficial to the flowing water realization.
Further, Fig. 4 is the flow chart of an also embodiment of the processing method of signal of the present invention, as shown in Figure 4, on above-mentioned Fig. 2 or basis embodiment illustrated in fig. 3, can also comprise after step 206 or step 305:
λ＝(R*U)
H*U (5)
Obtain eigenvalue λ.
R′＝R-λ*(U*U
H) (6)
Obtain the 3rd up channel covariance matrix R ';
In the single current situation, calculate weighing vector U and can be used as the weight coefficient that finally is superimposed upon on the transmitted signal, thereby so that the receiving terminal received power satisfies the index (desirable index is that received power reaches maximum) of systemic presupposition here.In double-current situation, the base station is judged when current channel circumstance is fit to stream transmission, can calculate the second up channel covariance matrix R characteristic of correspondence vector (for example weighing vector U) to implementing three any described method steps according to embodiment one, then will calculate the 3rd up channel covariance matrix R ' in this characteristic vector substitution formula (5) and (6).Again with the R in the 3rd up channel covariance matrix R ' alternative embodiment one to embodiment three any described method step, thereby calculate the weighing vector (i.e. time weighing vector) of R ' correspondence, this weighing vector of R ' correspondence can be the sub-eigenvector of the second up channel covariance matrix R.Thus, the weight coefficient that finally is superimposed upon on the transmitted signal can be the matrix that is made of the weighing vector weighing vector corresponding with R ' corresponding to R.It will be understood by those skilled in the art that in the situation that is generalized to multithread that final weight coefficient can be the matrix of a plurality of characteristic vectors compositions corresponding to the second up channel covariance matrix R.
Initial vector in the embodiment of the invention when obtaining characteristic vector is initial vector V
0(n), iterations is m/k, has further realized the unification of characteristic vector account form, namely adopts identical initial vector and iterations, thereby has improved hard-wired convenience.
Fig. 5 is the structural representation of an embodiment of the treatment facility of signal of the present invention, as shown in Figure 5, the treatment facility of the signal of present embodiment comprises: matrix calculation unit 11, weighing vector processing unit 12 and weighting processing unit 13, wherein, matrix calculation unit 11 is used for according to the up DMRS and the sounding signal acquisition channel estimation value that receive, and according to this channel estimation value, obtain the first up channel covariance matrix C; Wherein, described the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, and described the second up channel covariance matrix is the covariance matrix that is generated by described channel estimation value; Weighing vector processing unit 12 is used for according to this first up channel covariance matrix C, the initial vector V of input
0(n) and the configuration iterations m/k, obtain weighing vector U; Weighting processing unit 13 is used for generating weight coefficient according to this weighing vector U downstream signal is weighted processing, and sends the downstream signal after weighting is processed; Wherein, n represents described initial vector V
0The number of the element (n), m, k and m/k are positive integer, and m is the primary iteration number of times.
The treatment facility of the signal of present embodiment can be carried out the technical scheme of embodiment of the method shown in Figure 1, and for to carry out the as shown in Figure 1 technical scheme of embodiment of the method, present embodiment is also to comprise some nonrestrictive electronic circuits and structure etc.
In the present embodiment, because the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, the iterations that disposes in the process that treatment facility is processed the first up channel covariance matrix has become the 1/k of primary iteration number of times, alleviated the processing load of this processor, accelerated the convergence of characteristic vector, improve the speed of calculated characteristics vector, thereby accelerated the processing speed of whole piece link, and be beneficial to the flowing water realization.
Fig. 6 is the structural representation of another embodiment of the treatment facility of signal of the present invention, on above-mentioned basis embodiment illustrated in fig. 5, as shown in Figure 6, this weighing vector processing unit 12 comprises: matrix-vector multiply each other subelement 121, normalized subelement 122 and judgment sub-unit 123.
Preferably, matrix-vector multiplies each other subelement 121 for adopting formula V
I+1(n)=C
M/k* V
i(n) with this first up channel covariance matrix C and primary vector V
i(n) carry out the matrix-vector processing of multiplying each other, obtain secondary vector V
I+1(n); Normalized subelement 122 is used for adopting formula V
I+1(n)=V
I+1(n)/| V
I+1(n) | to this secondary vector V
I+1(n) carry out normalized; Judgment sub-unit 123 is used for judging whether i equals m/k-1; If this judgment sub-unit 123 is judged as be not equal to, i is added 1, and adopt multiply each other subelement 121 and this normalized subelement 122 of this matrix-vector to process accordingly; If being judged as, this judgment sub-unit 123 equals, with this secondary vector V
I+1(n) as weighing vector U; Wherein, i is positive integer, and i=0 during initialization, primary vector V
i(n)=V
0(n).
Further, the treatment facility of this signal can also comprise: characteristic value processing unit 14 and depression of order processing unit 15, and wherein, characteristic value processing unit 14 is used for according to this weighing vector U, employing formula λ=(R*U)
H* U obtains eigenvalue λ; Depression of order processing unit 15 is used for according to this weighing vector U and this eigenvalue λ, adopts formula R '=R-λ * (U*U
H), obtain the 3rd up channel covariance matrix R ', then weighing vector processing unit 12 also is used for according to the 3rd up channel covariance matrix R ', initial vector V
0(n) and iterations m/k, obtain time weighing vector U '; The arranged in matrix that weighting processing unit 13 also is used for weighing vector U and time weighing vector U ' composition is weight coefficient.
The treatment facility of the signal of present embodiment can execution graph 2 to Fig. 4 arbitrary shown in the technical scheme of embodiment of the method, and for carrying out the technical scheme to embodiment of the method shown in Figure 4 such as Fig. 2, present embodiment is also to comprise some nonrestrictive electronic circuits and structure etc.
In the present embodiment, because the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, the iterations that disposes in the process that treatment facility is processed the first up channel covariance matrix has become the 1/k of primary iteration number of times, alleviated the processing load of this processor, accelerated the convergence of characteristic vector, improve the speed of calculated characteristics vector, thereby accelerated the processing speed of whole piece link, and be beneficial to the flowing water realization.Initial vector when obtaining characteristic vector each time in the embodiment of the invention is initial vector V
0(n), iterations is m/k, and the embodiment of the invention has further realized the unification of characteristic vector account form, namely adopts identical initial vector and iterations, thereby has improved hard-wired convenience.
Further, in another embodiment of invention, matrix-vector multiplies each other subelement 121 for adopting formula V
I+1(n)=C
M/k* V
i(n), with this first up channel covariance matrix C and primary vector V
i(n) carry out the matrix-vector processing of multiplying each other, obtain secondary vector V
I+1(n); Judgment sub-unit 123 is used for judging whether i equals m/k-1; If this judgment sub-unit 123 is judged be not equal to, i is added 1 and adopt this matrix-vector subelement 121 that multiplies each other to process accordingly; If judging, this judgment sub-unit 123 equals, with this secondary vector V
I+1(n) as described weighing vector U; Normalized subelement 122 is used for adopting formula U=U/|U|, and U carries out normalized to this weighing vector; I=0 when wherein i is positive integer and initialization, primary vector V
i(n)=V
0(n).
The treatment facility of characteristic vector provided by the present invention can realize that by multiple hardwares wherein a kind of preferred embodiment is described below:
Fig. 7 is the structural representation of another embodiment of the treatment facility of signal of the present invention, in the present embodiment, with the treatment facility of this signal based on field programmable gate array (Field-Programmable Gate Array; Be called for short: hardware FPGA) is realized, and k=8 is example, introduce in detail the technical scheme of present embodiment, as shown in Figure 7, the treatment facility of the characteristic vector of present embodiment comprises: matrix computations device 21 (being equivalent to Fig. 5 or matrix calculation unit 11 shown in Figure 6), characteristic vector calculation element 22 (being equivalent to Fig. 5 or weighing vector processing unit 12 shown in Figure 6), matrix reduction and characteristic value calculation element 23 (comprising above-mentioned characteristic value processing unit 14 shown in Figure 6 and depression of order processing unit 15) and output buffer storage 24.
Need to prove, the treatment facility of signal of the present invention is not limited to realize based on the hardware of FPGA, can also realize based on hardware such as central processing unit or application-specific integrated circuit (ASIC)s, particularly can realize based on hardware such as the central processing unit of parallel processing or application-specific integrated circuit (ASIC)s.
In the present embodiment, this matrix computations device 21 can comprise three matrix square computing modules, is respectively the first matrix square computing module U0, the second matrix square computing module U1 and the 3rd matrix square computing module U2.Concrete, the first matrix square computing module U1 mainly is comprised of 3 functional units: buffer memory prime input data soldier pang buffer (Ping-Pong Buffering; Be called for short: PPBUF) group 0, buf_ctrl unit and algorithm process unit U0_SQU.These buffer memory prime input data PPBUF group 0 comprises RAM0, RAM1 and RAM2.This algorithm process unit U0_SQU comprises: subelement is processed in matrix square operation subelement, transparent transmission and bypass.The second matrix square computing module U1 also mainly is comprised of 3 functional units: buffer memory prime input data PPBUF group 1, buf_ctrl unit and algorithm process unit U1_SQU.These buffer memory prime input data PPBUF group 1 comprises RAM3, RAM4 and RAM5.This algorithm process unit U1_SQU comprises: subelement is processed in matrix square operation subelement, transparent transmission and bypass.The 3rd matrix square computing module U2 also mainly is comprised of 3 functional units: buffer memory prime input data PPBUF group 2, buf_ctrl unit and algorithm process unit U2_SQU.These buffer memory prime input data PPBUF group 2 comprises RAM6, RAM7 and RAM8.This algorithm process unit U2_SQU comprises: subelement is processed in matrix square operation subelement, transparent transmission and bypass.
Characteristic vector calculation element 22 comprises: buffer memory prime input data PPBUF group 3, buf_ctrl module and algorithm processing module GET_VECTOR.Wherein, buffer memory prime input data PPBUF group 3 comprises RAM9 and RAM10.Algorithm processing module GET_VECTOR comprises acquiring unit, calculated characteristics vector location and transparent transmission unit.Wherein, acquiring unit is connected with the calculated characteristics vector location, and processes subelement and algorithm processing module GTE_R be connected respectively (not drawing in the drawings) with matrix square operation subelement, transparent transmission and bypass among each algorithm process unit U0_SQU to U2_SQU.Need to prove that also this transparent transmission and bypass are processed in the subelement and also be provided with not gate, an end of this not gate is connected with this acquiring unit.
Matrix reduction and characteristic value calculation element 23 comprise: table tennis register, buffer memory prime input data PPBUF group 4, buf_ctrl module and algorithm processing module GTE_RUSB.Wherein, buffer memory prime input data PPBUF group 4 comprises RAM11.Algorithm processing module GTE_RUSB comprises computation of characteristic values and R ' unit and transparent transmission unit.
Need to prove that RAM0 to RAM12 can read a dual port RAM of writing and for PPBUF for one.
In the present embodiment, the operation principle of the treatment facility of characteristic vector is mainly: RAM0, RAM1 and RAM2 in the buffer memory prime input data PPBUF group 0 all store covariance matrix R (being equivalent to the second up channel covariance matrix shown in above-described embodiment), algorithm process unit U0_SQU copies the second up channel covariance matrix R among the RAM0, and by transparent transmission and bypass processing subelement this covariance matrix R is write among the RAM3.Simultaneously, the matrix square operation subelement among the algorithm process unit U0_SQU reads the covariance matrix R among RAM1 and the RAM2, the row matrix of going forward side by side square operation.In addition, because matrix square operation subelement receives the not by-passing signal bypass0 that acquiring unit sends, transparent transmission and bypass are processed subelement and are received the not by-passing signal bypass0 that this acquiring unit sends, and transparent transmission and bypass process in the subelement not gate to this not by-passing signal bypass0 carry out conversion process, so that by-passing signal bypass0 does not convert by-passing signal to, thereby so that not bypass of matrix square operation subelement, the subelement bypass is processed in transparent transmission and bypass, therefore, so that matrix square operation subelement is result of calculation, i.e. covariance matrix R
2Export among RAM4 and the RAM5.
Algorithm process unit U1_SQU copies the covariance matrix R among the RAM3, and by transparent transmission and bypass processing subelement this covariance matrix R is write among the RAM6.Simultaneously, matrix square operation subelement reads covariance matrix R among RAM4 and the RAM5 among the algorithm process unit U1_SQU
2, the row matrix of going forward side by side square operation.In addition, because matrix square operation subelement receives the not by-passing signal bypass1 of acquiring unit input, transparent transmission and bypass are processed subelement and are received the not by-passing signal bypass1 that this acquiring unit sends, and transparent transmission and bypass process in the subelement not gate to this not by-passing signal bypass1 carry out conversion process, so that by-passing signal bypass1 does not convert by-passing signal to, then so that the not bypass of matrix square operation subelement, the subelement bypass is processed in transparent transmission and bypass, therefore, matrix square operation subelement is result of calculation, i.e. covariance matrix R
4Export among RAM7 and the RAM8.
Algorithm process unit U2_SQU copies the second up channel covariance matrix R among the RAM6, and by transparent transmission and bypass processing subelement this covariance matrix R is write among the RAM9.Simultaneously, matrix square operation subelement reads covariance matrix R among RAM7 and the RAM8 among the algorithm process unit U1_SQU
4, the row matrix of going forward side by side square operation.In addition, because matrix square operation subelement receives the not by-passing signal bypass2 of acquiring unit input, transparent transmission and bypass are processed subelement and are received the not by-passing signal bypass2 that this acquiring unit sends, and transparent transmission and bypass process in the subelement not gate to this not by-passing signal bypass2 carry out conversion process, so that by-passing signal bypass2 does not convert by-passing signal to, then so that the not bypass of matrix square operation subelement, the subelement bypass is processed in transparent transmission and bypass, therefore, matrix square operation subelement is result of calculation, i.e. covariance matrix R
8Export among the RAM10.
Algorithm processing module GET_VECTOR copies the second up channel covariance matrix R among the RAM9, and this covariance matrix R is write among the RAM11 by the transparent transmission unit.Simultaneously, the calculated characteristics vector location receives the not by-passing signal bypass3 of acquiring unit input, then so that the covariance matrix R among the RAM10 is read in the not bypass of calculated characteristics vector location
8(being equivalent to the first up channel covariance matrix shown in above-described embodiment), and from acquiring unit, obtain initial vector V
0(n) and the iterations m/k of configuration, adopt formula U=C
M/k* V
0(n) iteration is carried out calculation process, obtains characteristic vector U (being equivalent to the weighing vector shown in above-described embodiment), and this characteristic vector U is exported to the table tennis register of matrix reduction and characteristic value calculation element 23.
Algorithm processing module GTE_R ' reads the characteristic vector U in the table tennis register, compute matrix depression of order, i.e. covariance matrix R ' (being equivalent to the 3rd up channel covariance matrix shown in above-described embodiment), and computation of characteristic values λ.In addition, owing to receive the not by-passing signal bypass4 of acquiring unit input, therefore, covariance matrix R ', characteristic vector U and eigenvalue λ are written among the RAM12.
Further, when needs obtain sub-eigenvector and sub-eigenvalue, only need all to store covariance matrix R ' and repeat above-mentioned processing among RAM0, RAM1 in buffer memory prime input data PPBUF group 0 and the RAM2, namely can obtain sub-eigenvector and sub-eigenvalue.
It should be noted that, above-mentioned buf_ctrl unit and the major function of buf_ctrl module are the rights to use of distributing buffer, so that the unit in the PPBUF group or the side of writing in the subelement and the side of reading do not operate respectively unit or subelement in this PPBUF group simultaneously.In addition, because the buffer of control is PPBUF, so the buf_ctrl unit manages table tennis BUF and pang the BUF of PPBUF simultaneously.
Fig. 8 is the concrete structure schematic diagram of buf_ctrl, as shown in Figure 8, the control of PPBUF distributed and switches mainly pass through Flag[1:0] and wr_rsl[1:0], rd_rls[1:0] three signals finish, Flag[1 wherein] and wr_rsl[1], rd_rls[1] signal controlling pang BUF, Flag[0] and wr_rsl[0], rd_rls[0] signal controlling table tennis BUF.Take control table tennis BUF as example, it distributes the machine-processed as follows of BUF control:
When resetting, Flag[0] be 0, the control of table tennis BUF is in the side of writing at this moment, only finish the data writing to table tennis BUF this moment in the side of writing, after control is write in release (be wr_rsl[0] be 1, this signal is pulse signal), Flag[0] become 1, this moment, the control of table tennis BUF was given the side of reading, and only finished the reading out data to table tennis BUF in the side of reading, and the release read control (be rd_rsl[0] be 1, this signal is pulse signal) after, Flag[0] become 0, the control of table tennis BUF is given the side of writing, and so far finishes the circulation of a read-write.
In this equipment, all comprised a plurality of PPBUF in PPBUF group 0, PPBUF group 1, PPBUF group 2 and the PPBUF group 3, but these PPBUF can unify to be controlled by 1 buf_ctrl unit or buf_ctrl module, carry out together the switching of PPBUF control.
When external unit calls this equipment and calculates the characteristic vector of covariance matrix, after need to writing first covariance matrix R to RAM0, RAM1 and RAM2, pass through again register configuration initial vector and iterations to the calculated characteristics vector location, notify this equipment by the wr_rls signal of buf_ctrl unit or buf_ctrl module at last; This equipment will read covariance matrix R, finishes calculating, and exports the as a result R ' of this calculating and eigenvalue λ and characteristic vector U and export buffer storage 24 to, and discharges the control of writing of output output buffer storage 24; Output buffer storage 24 is after finishing receiving data, and the Flag signal of buf_ctrl module can become 1, thereby the notice external unit obtains final result of calculation.
One of ordinary skill in the art will appreciate that: all or part of step that realizes above-mentioned each embodiment of the method can be finished by the relevant hardware of program command.Aforesaid program can be stored in the computer read/write memory medium.This program is carried out the step that comprises above-mentioned each embodiment of the method when carrying out; And aforesaid storage medium comprises: the various media that can be program code stored such as ROM, RAM, magnetic disc or CD.
It should be noted that at last: above each embodiment is not intended to limit only in order to technical scheme of the present invention to be described; Although with reference to aforementioned each embodiment the present invention is had been described in detail, those of ordinary skill in the art is to be understood that: it still can be made amendment to the technical scheme that aforementioned each embodiment puts down in writing, and perhaps some or all of technical characterictic wherein is equal to replacement; And these modifications or replacement do not make the essence of appropriate technical solution break away from the scope of various embodiments of the present invention technical scheme.
Claims (8)
1. the processing method of a signal is characterized in that, comprising:
Obtain channel estimation value, and according to described channel estimation value, obtain the first up channel covariance matrix C; Wherein, described the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, and described the second up channel covariance matrix R is the covariance matrix that is generated by described channel estimation value;
According to described the first up channel covariance matrix C, the initial vector V of input
0(n) and the configuration iterations m/k, obtain weighing vector U;
Generate weight coefficient according to described weighing vector U, downstream signal is weighted processing, and send the downstream signal after weighting is processed;
Wherein, n represents described initial vector V
0The number of the element (n), m, k and m/k are positive integer, and m is the primary iteration number of times.
2. the processing method of signal according to claim 1 is characterized in that, described initial vector V according to described the first up channel covariance matrix C, input
0(n) and the configuration iterations m/k, obtain weighing vector U, comprising:
Adopt formula V
I+1(n)=C
M/k* V
i(n) with described the first up channel covariance matrix C and primary vector V
i(n) carry out matrix multiple and process, obtain secondary vector V
I+1(n);
Adopt formula V
I+1(n)=V
I+1(n)/| V
I+1(n) | to described secondary vector V
I+1(n) carry out normalized;
Judge whether i equals m/k-1;
If be not equal to, i is added 1, adopt formula V
I+1(n)=C
M/k* V
i(n) with described the first up channel covariance matrix C and primary vector V
i(n) carry out matrix multiple and process, obtain secondary vector V
I+1(n); And employing formula V
I+1(n)=V
I+1(n)/| V
I+1(n) | to described secondary vector V
I+1(n) carry out normalized;
If i equals m/k-1, with described secondary vector V
I+1(n) as weighing vector U;
Wherein, i is positive integer, and i=0 during initialization, primary vector V
i(n)=V
0(n).
3. the processing method of signal according to claim 1 is characterized in that, described initial vector V according to described the first up channel covariance matrix C, input
0(n) and the configuration iterations m/k, obtain weighing vector U, comprising:
Adopt formula V
I+1(n)=C
M/k* V
i(n) with described the first up channel covariance matrix C and primary vector V
i(n) carry out the matrix-vector processing of multiplying each other, obtain secondary vector V
I+1(n);
Judge whether i equals m/k-1;
If be not equal to, i is added 1, adopt formula V
I+1(n)=C
M/k* V
i(n) with described the first up channel covariance matrix C and primary vector V
i(n) carry out the matrix-vector processing of multiplying each other, obtain secondary vector V
I+1(n);
If i equals m/k-1, with described secondary vector V
I+1(n) as weighing vector U;
Adopt formula U=U/|U|, described weighing vector U is carried out normalized;
Wherein, i is positive integer, and i=0 during initialization, primary vector V
i(n)=V
0(n).
4. the processing method of signal according to claim 1 is characterized in that, described weighing vector U generates described weight coefficient, comprising:
According to described weighing vector U, employing formula λ=(R*U)
H* U obtains eigenvalue λ;
According to described weighing vector U and described eigenvalue λ, adopt formula R '=R-λ * (U*U
H) obtain the 3rd up channel covariance matrix R ';
According to described the 3rd up channel covariance matrix R ', described initial vector V
0(n) and described iterations m/k, obtain time weighing vector U ';
Be described weight coefficient with the arranged in matrix of described weighing vector U and time weighing vector U ' composition.
5. the treatment facility of a signal is characterized in that, comprising:
Matrix calculation unit is used for obtaining channel estimation value, and according to described channel estimation value, obtains the first up channel covariance matrix C; Wherein, described the first up channel covariance matrix C is the k power of the second up channel covariance matrix R, and described the second up channel covariance matrix is the covariance matrix that is generated by described channel estimation value;
The weighing vector processing unit is used for according to described the first up channel covariance matrix C, the initial vector V of input
0(n) and the configuration iterations m/k, obtain weighing vector U;
Weighting processing unit is used for generating weight coefficient according to described weighing vector U, and downstream signal is weighted processing, and sends the downstream signal after weighting is processed;
Wherein, n represents described initial vector V
0The number of the element (n), m, k and m/k are positive integer, and m is the primary iteration number of times.
6. the treatment facility of signal according to claim 5 is characterized in that, described weighing vector processing unit comprises:
The matrix-vector subelement that multiplies each other is used for adopting formula V
I+1(n)=C
M/k* V
i(n) with described the first up channel covariance matrix C and primary vector V
i(n) carry out the matrix-vector processing of multiplying each other, obtain secondary vector V
I+1(n);
The normalized subelement is used for adopting formula V
I+1(n)=V
I+1(n)/| V
I+1(n) | to described secondary vector V
I+1(n) carry out normalized;
Judgment sub-unit is used for judging whether i equals m/k-1;
If described judgment sub-unit is judged as be not equal to, i is added 1, and adopt multiply each other subelement and described normalized subelement of described matrix-vector to process accordingly;
If being judged as, described judgment sub-unit equals, with described secondary vector V
I+1(n) as weighing vector U;
Wherein, i is positive integer, and i=0 during initialization, primary vector V
i(n)=V
0(n).
7. the treatment facility of signal according to claim 5 is characterized in that, described weighing vector processing unit comprises:
The matrix-vector subelement that multiplies each other is used for adopting formula V
I+1(n)=C
M/k* V
i(n), with described the first up channel covariance matrix C and primary vector V
i(n) carry out the matrix-vector processing of multiplying each other, obtain secondary vector V
I+1(n);
Judgment sub-unit is used for judging whether i equals m/k-1;
If described judgment sub-unit is judged be not equal to, i is added 1, and adopt the described matrix-vector subelement that multiplies each other to process accordingly;
If judging, described judgment sub-unit equals, with described secondary vector V
I+1(n) as described weighing vector U;
The normalized subelement is used for adopting formula U=U/|U|, and described weighing vector U is carried out normalized;
Wherein, i is positive integer, and i=0 during initialization, primary vector V
i(n)=V
0(n).
8. the treatment facility of signal according to claim 5 is characterized in that, also comprises:
The characteristic value processing unit is used for according to described weighing vector U, employing formula λ=(R*U)
H* U obtains eigenvalue λ;
The depression of order processing unit is used for according to described weighing vector U and described eigenvalue λ, adopts formula R '=R-λ * (U*U
H), obtain the 3rd up channel covariance matrix R ', to obtain time weighing vector U ';
Then described weighing vector processing unit also is used for according to described the 3rd up channel covariance matrix R ', described initial vector V
0(n) and described iterations m/k, obtain time weighing vector U ';
The arranged in matrix that described weighting processing unit also is used for described weighing vector U and time weighing vector U ' composition is described weight coefficient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210082789.0A CN103368621B (en) | 2012-03-26 | 2012-03-26 | The processing method and equipment of signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210082789.0A CN103368621B (en) | 2012-03-26 | 2012-03-26 | The processing method and equipment of signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103368621A true CN103368621A (en) | 2013-10-23 |
| CN103368621B CN103368621B (en) | 2017-12-15 |
Family
ID=49369256
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210082789.0A Active CN103368621B (en) | 2012-03-26 | 2012-03-26 | The processing method and equipment of signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103368621B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1925362A (en) * | 2005-08-29 | 2007-03-07 | 中兴通讯股份有限公司 | Method for realizing intelligent antenna based on even linear array |
| CN101192868A (en) * | 2006-11-24 | 2008-06-04 | 中兴通讯股份有限公司 | Multi-service wave bundle shaping device for wireless communication system |
| CN101488792A (en) * | 2008-01-15 | 2009-07-22 | 大唐移动通信设备有限公司 | Wave beam shaping method and apparatus |
| US20110170521A1 (en) * | 2010-01-14 | 2011-07-14 | Cisco Technology, Inc. | Dynamic Downlink Beamforming Weight Estimation for Beamforming-Space Time Code Transmissions |
-
2012
- 2012-03-26 CN CN201210082789.0A patent/CN103368621B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1925362A (en) * | 2005-08-29 | 2007-03-07 | 中兴通讯股份有限公司 | Method for realizing intelligent antenna based on even linear array |
| CN101192868A (en) * | 2006-11-24 | 2008-06-04 | 中兴通讯股份有限公司 | Multi-service wave bundle shaping device for wireless communication system |
| CN101488792A (en) * | 2008-01-15 | 2009-07-22 | 大唐移动通信设备有限公司 | Wave beam shaping method and apparatus |
| US20110170521A1 (en) * | 2010-01-14 | 2011-07-14 | Cisco Technology, Inc. | Dynamic Downlink Beamforming Weight Estimation for Beamforming-Space Time Code Transmissions |
Non-Patent Citations (1)
| Title |
|---|
| 肖航: ""智能天线中的自适应波束形成算法研究与设计"", 《中国优秀硕士学位论文全文数据库》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103368621B (en) | 2017-12-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP3634696B2 (en) | Space-time diversity receiver for wireless systems | |
| CN104698430B (en) | It is a kind of for carrying the high-precision angle estimating method based on virtual antenna array | |
| US8417758B1 (en) | Left and right matrix multiplication using a systolic array | |
| CN105049097B (en) | Extensive MIMO linearity tests hardware architecture and detection method under non-ideal communication channel | |
| CN108141259A (en) | Scalable extensive MIMO receiver | |
| CN107852387B (en) | Method for reducing precoding matrix calculation and user equipment grouping complexity in large-scale multi-input multi-output system | |
| CN105915477A (en) | Large-scale MIMO detection method based on GS method, and hardware configuration | |
| CN103516643A (en) | MIMO detecting preprocessing device and method | |
| CN106716866A (en) | Ping pong beamforming | |
| US8510364B1 (en) | Systolic array for matrix triangularization and back-substitution | |
| Luong et al. | Dynamic network service selection in IRS-assisted wireless networks: A game theory approach | |
| US8473540B1 (en) | Decoder and process therefor | |
| CN107483090B (en) | Large-scale MIMO system precoding realization method based on LDLT decomposition | |
| CN105978609A (en) | Massive MIMO linear detection hardware architecture and method under correlated channels | |
| CN101384992A (en) | Method and apparatus to perform multiply-and-accumulate operations | |
| CN101998440A (en) | Method and device for detecting signals in multi-input and multi-output system | |
| Liu et al. | An FPGA-based MVDR beamformer using dichotomous coordinate descent iterations | |
| CN103368621A (en) | Signal processing method and device | |
| Thanos et al. | Hardware trade-offs for massive MIMO uplink detection based on Newton iteration method | |
| CN115426012A (en) | Base band chip, hybrid pre-coding method and terminal equipment | |
| CN105553899B (en) | The signal detecting method and device of approximate solution are asked based on system of linear equations | |
| Liu et al. | Architecture design of a memory subsystem for massive MIMO baseband processing | |
| Lin et al. | Accelerating Next-G Wireless Communications with FPGA-Based AI Accelerators | |
| CN111541472A (en) | Low-complexity machine learning assisted robust precoding method and device | |
| Khan et al. | Area & power efficient VLSI architecture for computing pseudo inverse of channel matrix in a MIMO wireless system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |