WO2007094622A1 - A wireless communication system - Google Patents

Abstract

A novel diversity method of hybrid selection/equal-gain combining is invented. This scheme dynamically selects the best combination of branches by a simple test and combines them in equal-gain combining (EGC) manner. As a result, the scheme always shows better performance than conventional EGC and selection combining (SC), and close to maximum ratio combining (MRC) in noise-limited channel environment.

Description

[DESCRIPTION]

[Invention Title]

A WIRELESS COMMUNICATION SYSTEM

[Technical Field]

<i> This invention relates to a wireless communication system, and more specifically, to a diversity or beamforming technology, which may be equipped in a base station or a mobile station, so as to significantly improve the performance of the wireless communication system.

[Background Art]

<2> Diversity technique has been considered as an effective method to combat multipath fading phenomenon when independent multiple channels, for instance, multiple antennas or time-distinguishable multiple paths are provided. In particular, if the multiple channels are provided through multiple antennas, the diversity technique is often called beamforming, smart antenna or multiple antenna system. Conventionally, three types of diversity combining have been widely used: maximum-ratio combining (MRC), selection combining

(SC) and equal-gain combining (EGC). Assuming a beamforming system with ^ antennas, the MRC shown in Figure 1 (a) receives multiple input signals (11.1-11.N), and carries out both phase adjustment (13.1-13.N) and amplitude adjustment (15.1-15.N) through phase control block (21) and amplitude control block (23), respectively. The phase and amplitude adjustment is made in the way that generates the maximum output signal-to-noise ratio (SNR) ratio at the combined output (19). The EGC shown in Figure 1 (b) receives multiple input signals (31.1-31.N), and carries out only phase adjustment (33.1-33.N) through phase control block (39). The phase adjustment is made in a way that aligns the phases of input signals to generate a coherently combined output (37). The SC shown in Figure 1 (c) receives multiple input signals (41.1- 41.N), and carries out ON/OFF switching (43.1-43.N) of each branch through the switching control block (49). The switching is made in the way that selects the input signal with the highest SNR among total branches.Among the three diversity methods, it is well known that MRC shows the best performance in the noise-limited channel environment, and SC requires the least complexity among the above options. EGC is a compromise of the high performance of MRC and the low complexity of SC.

<3> Recently, various hybrid types of diversity were invented in order to further compromise between performance and complexity. Taking one example of the hybrid types of diversity which have been published, a hybrid selection/equal-gain combining (HS/EGC) scheme [Ref 1] selects a

"predetermined" number of the strongest branches out of the ^ branches, and combine them in EGC manner. [Ref I]: C. D. Iskander, performance of coherent receiver with hybrid selection/equal-gain combining in Nakagami-m fading, IEEE Radio and Wireless Conference, pp. 111-114, 2004. [Disclosure] [Technical Problem]

<4> However, the performance of the EGC and the conventional HS/EGC in [Ref ^' 1] varies according to the statistics of the channel in spite of its advantage of low complexity. Although its performance is known to be better than SC in almost cases, it may be worse than SC in some case where signal level of a certain branch remains very weak for a long time due to, for example, human blocking over the branch. The drawback of the EGC and the conventional HS/EGC is caused due to the fact that the number of branches to select is predetermined. [Technical Solution]

<5> According to the invention, there is provided a wireless communication system comprising a plurality of branches and a combiner, wherein the system comprises a switching control block which dynamically selects branches for equal gain combining of the selected branches and phase control block which aligns the phases of the selected branches.

<6> In one embodiment, the branches are dynamically selected and combined in EGC manner for maximizing signal-to-noise ratio at the combiner output.

<7> In one embodiment, the switching control block by default includes a branch having highest signal-to-noise ratio and subsequently adds branches according in descending order of signal magnitude until the signal-to-noise ratio does not increase.

<8> In one embodiment, the switching control block utilizes a dynamically chosen set of numbers as proportional coefficients to compare input signal magnitudes and to determine inclusion of a branch or branches in said equal gain combining.

<9> In one embodiment, the phase control block aligns the phase of the each branch to a reference phase according to the phase-locked loop principle.

<io> In another aspect, the invention provides a signal combining method in a multiple channel system configured to select a subset of all branches so as to result in the maximum output SNR out of all possible selections when the selected branches are combined in EGC manner, and to combine the selected subset of branches in EGC manner.

<ii> In another aspect, the invention provides a signal transmission method in a multiple channel system configured to select a subset of all branches so as to result in the maximum reception SNR out of all possible selections when the signals transmitted from the selected antennas are received and superimposed at the other communication side.

<i2> In one embodiment, the method comprises the steps of adjusting the phase of each branch and selecting branches for transmission based on the information of phase control and switching control obtained in receiving mode. [Advantageous Effeetsi

<14> In this invention, a new type of HS/EGC scheme is presented which employs the best combination of branches "dynamically", not predetermined as in the conventional HS/EGC scheme, so that whose performance is always better than those of EGC and SC, and close to that of MRC. At the same time, a hybrid selection/equal-gain transmission (HS/EGT) scheme is presented as the transmission counter part of the HS/EGC scheme. In particular, when the underlying communication system is based on time-division duplexing (TDD) for receiving (RX) and transmission (TX) duplexing, the combination of HS/EGC and HS/EGT according to the invention has a great merit in terms of phase information re-use, and this feature will be described in detail.

<15> In a system with multiple input channels (or branches), the invented hybrid selection/equal-gain combining (HS/EGC) scheme dynamically selects a subset of the total branches according to the wireless channel conditions so that when the selected branches are combined in EGC manner, the output signal-to- noise ratio (SNR) should be maximized over all possible selections. The system is composed of multiple antennas, phase adjustment blocks, switches, a phase control block, a switching control block and a combiner. The scheme can be translated to the transmission version, so called hybrid select ion/equal- gain transmission (HS/EGT), where the information of phase control and switching control are reused for transmission. [Description of Drawings]

<i6> The invention will be more clearly understood from the following description of some embodiments thereof, given by way of example with reference to the accompanying drawings in which:

<i7> Figure 1 is a block diagram of conventional diversity combining schemes: (a) MRC, (b) EGC, (c) SC.

<18> Figure 2 is a flow chart of the invented HS/EGC algorithm.

<i9> Figure 3 is a flow chart of the invented HS/EGC algorithm for dual branches.

<20> Figure 4 is another flow chart of the invented HS/EGC algorithm.

<2i> Figure 5 is another flow chart of the invented HS/EGC algorithm for dual branches.

<22> Figure 6 is a comparison of average output SNR performance between various diversity combining schemes.

<23> Figure 7 is a block diagram of the generic RX beamforming system with the invented HS/EGC scheme.

<24> Figure 8 is a block diagram of the generic TX beamforming system with the invented HS/EGT scheme.

<25> Figure 9 is a block diagram of another embodiment example of the RX beamforming system with the invented HS/EGC scheme.

<26> Figure 10 is a block diagram of another embodiment example of the TX. beamforming system with the invented HS/EGT scheme. <27> Figure 11 is a block diagram of another embodiment example of the TDD-based beamforming system^' with the invented HS/EGC and HS/EGT schemes. <28> Figure 12 is a block diagram of another embodiment example of the TDD-based beamforming system with the invented HS/EGC and HS/EGT schemes, and dual branches. <29> Figure 13 is a block diagram of an embodiment example of the phase-locked loop. <30> Figure 14 is a block diagram of an embodiment of dual-branch beamformer module and its tree-structured extension.

[Best Mode]

<32> HS/EGC algorithm: <33> This invention will be described taking the beamforming system as a typical example of the multiple channel system. However, it should be understood that many other examples can be realized based on the idea disclosed in this invention. For reference, in the beamforming system, the word "antenna branch" (or, simply "branch", is often used for the word "channel", In this invention, a beamforming system with ^N branches are assumed. Let ^a ' be the input signal magnitude and ^{n l} be an independently and identically distributed

<35> (i.i.d.) noise with power ™ for the ^z"-th branch, *^'"* respectively. Then, the input signal to noise ratio (SNR) is expressed as

y _--⁰≠^-

* . In this invention, a subset of total branches out of all possible combinations is selected for combining which will result in the maximum output SNR when the selected branches are combined in equal-gain combining (EGC) manner. Such a scheme and the associated output SNR are

{HS/EGC) called hybrid selection/equal-gain combining (HS/EGC) and ' respectively. <36> Because the EGC scheme combines the branch signals coherently (or, with aligned phases), it is obvious that, given the number of branches to choose l_c \≤L<N l_c

(say , ), choosing branches with the largest input SNRs will give the largest output SNR than any other combinations. Based on the

observation, the largest branches are combined and the associated output

(BS/EGC)

SNR γ' (L) is checked. This will be continued from L=I until output SNR does not increase any more. This HS/EGC scheme can be summarized in an algorithmic manner as follows:

<37> 1. Measure input signal magnitudes for each branch.

QL ₁ > - ■ ■» C£ _w Ot r in j, . . . ₅ Ct r _jyr

<38> 2. Arrange in the descending order ^{L J L J} with

^α [ i] > α

<39> 3 . Set L=I .

<40> 4 . For L=I , 2 , . . . ,N-I ,

(HS/EGC) {HS/EGC)

<41> If T (L+l) > ^ (L) , (1)

<42> then L <- L+l and εrc

<43> Otherwise , L ^"^₆ =L and stop .

L_c ^α[i]^?'""'^α[i_σ] .

<44> 5. Combine branches corresponding to in EGC manner, and go to 1.

<45> The flow chart of the invented HS/EGC algorithm according to the above is shown in Figure 2. First, measure input signal magnitude of each branch,

^{1 w}(101). Second, arrange ^{1 w}in the descending order α_ril,..._s α_rAn ot_ril>...> α_rAn T

^{L1J LiVJ}with ^{L1J [N1} (103). Third, determine the ^L< strongest branches according to the way explained in the -algorithm above

(105, 107, 109, 111). Fourth, select the strongest ^L° branches (113).

Finally, combine the selected branches in EGC manner (115). The above flow repeats. Note that because the above process repeats during operation, the branch selection is dynamically carried out according to the wireless channel conditions, not predetermined as in the conventional methods. Instead of deciding the branch combining by testing output SNRs as described in Equation (1), the decision may be made by testing the input signal magnitudes

as described as follows: Suppose the strongest branches were selected for combining. Then, the signal component power of the combined output is

expressed as

because the signal components would be combined with phase being aligned (in other words, coherently). On the other hand, the noise component power of the combined output is express as LP" because the noise component would be combined with random phase (in other words, noncoherentIy). Therefore, the output SNR is given by

<48> From Equat ion (1) and Equat ion (2) , a test condi t ion for inclusion of a branch in combining can be easi ly obtained such as , after default inclusion ct _rz1 of the strongest branch, the branches corresponding to ^{L J}with

<52> are included in the combining, and the resultant output SNR of the invented HS/EGC scheme is expressed as

(HS/EGC) _ (HSIEGC)(T N _

_<53_> L₀P_n (4)

<55> The test condition (107) in Figure 2 can be replaced by that in

Equation (3).

<56> In particular, the invented algorithm can be written as follows for dual branches (^=2):

Ct₁, Ci

<57> 1. Measure input signal magnitude for each branch

<59> 2. Arrange in the descending order ^{L J L J} with

_<60_> 3. I f ^a P L^^]J^{>"K" GL [} L^ii]j _ffffi_it_th_h K A ≡s JVΪ2 -- 1l (5)

<6i> Select both branches .

<62> Otherwi se , select the strongest branch only.

<63> 4. Combine the selected branches in EGC manner , and go to 1.

<64> The f low chart of the invented HS/EGC algori thm for dual branches according to the above i s shown in Figure 3. First , measure input signal magni tude of

^Ot₁, ^CC₂ ^αl_* ^a2 each branch (151). Second, arrange in descending order

ot m-, ct _C-_J1 Ct _ril> a. r₉₁ a_r,i>Za_rn

^{[1] [2]} with ^{[1] [2]} (153). Third, if ^{[2] [1]} (155), then, select both branches (157). Otherwise, select the strongest branch only (159). Finally, combine the selected branches in EGC manner (161). The above flow repeats. Note that because the process repeats during operation, the

threshold level setting (here, ^L ) and the branch selection are dynamically carried out according to the wireless channel conditions, not predetermined as in the conventional methods.

<65> By the way, the HS/EGC algorithms in Figure 2 and Figure 3 based on the input signal levels may have a difficulty in measuring them in the realistic situation. This is because that the signal is normally received together with noise, and it may not be easy to separate the signal component from the combined signal and noise. As a solution to that,, the algorithm may be modified in such a way replacing the input signal magnitudes by input signal plus noise powers. Here, it is assumed that the noise is independently and p identically distributed (i.i.d) with power " for each branch. Then, the input signal plus noise power for each branch can be expressed as

<66> P = _Αi ²+P_κXi=l,...,X). (6)

<68> This is the power which can be easily measured for each branch and arranged

in the descending order such as r^L11^J F^LiVi^J . Then, the HS/EGC algorithm in Figure 2 can be replaced by another version employing measurement of input signal plus noise power as follows:

<69> 1. Measure input signal plus noi se powers for each branch.

.

<71> 2. Arrange n the descending _order

<72> 3. Set ^L 1 <73> 4. For L = I, 2, . . . ,N-I _γ (HStSGC) ζ_£+l)_{> y} (HStSGC) ^

<74> I f

(7)--

L^L + l

<75> then and go to 4.

<76> Otherwise, ^ and stop.

<77> 5. Combine ^c branches corresponding to ^{L J J} in EGC manner, and go to 1. <78> The flow chart of the invented HS/EGC algorithm according to the above is shown in Figure 4. First, measure input signal plus signal power of each

P P P P branch, (201). Second, arrange in the

descending order [1]^*""* ^_with ^{[1] """ [iV]}(203). Third,

L_c determine the strongest branches according to the way explained in the

∑_c algorithm above (205, 207, 209, 211). Fourth, select the strongest

branches (213). Finally, combine the selected ^c branches in EGC manner (215). The above flow repeats. Note that because the above process repeats during operation, the branch selection is dynamically carried out according to the wireless channel conditions, not predetermined as in the conventional methods.

<79> In the same way, the HS/EGC algorithm for dual branches ( JV=2) in Figure 3 can be replaced by another version employing measurement of input signal plus noise power as follows:

<80> 1. Measure input signal plus noise powers " 1* ^■*^■ 2 and noi■se power P" . <82> 2. Arrange 1 * 2 in the descending order r ^L 11 ^{J 5 L}T 21 ^J with

^JP [ 1] ^{> JP} [2] _

<83> 3 _o . I _{T f}f ^ [ ^L2] ^J> *l ^ [ ^L l] ^J+ *2 ^ wi -_tth_h ^ ≡ (V2 - 1)² and

<84> Select both branches.

<85> Otherwise, select the strongest branch only.

<86> 4. Combine the selected branches in EGC manner, and go to 1.

<87> The condition in Equation (8) is easily obtained from Equation (5) and

P

Equation (6). Note that this version requires measurement of ⁿ as well

as ^l and ^F 2. However, when the input SNR is high enough (i.e.,

P "^■ -*KPⁿ), the condition in Equation (8) can be replaced by

^pW=^{* κ}i ^pm ₍₉₎

<89> This saves the necessity for measuring

<90> The flow chart of the invented HS/EGC algorithm for dual branches according to the above is shown in Figure 5. First, measure input signal plus noise

power of each branch P_1^P₂ and noise power P_n (251). Second,

arrange P l* P 2 in the descending order ^m^{L J}* F^L2i^J with

^{L1J [Zi} (253). Third, if ^{μj X L1J l} (255), then, select both branches (257). Otherwise, select the strongest branch only (259). Finally, combine the selected branches in EGC manner (261). The above flow repeats. As explained in the above, the condition in (255) in Figure 5 can be replaced by Equation (9) when the input SNR is high enough, Note that because the process repeats during operation, the threshold level setting for

^P [2] and the branch selection are carried out dynamically according to the wireless channel conditions, not predetermined as in the conventional methods. <9i> To confirm the performance of the invented scheme, a comparison of the average output SNR (^r ) performance between various diversity schemes (MRC, EGC, SC and the invented HS/EGC) in dual Rayleigh fading wireless channel conditions is shown in Figure 6. Here, the χ-axis is the power correlation

O . . . coefficient ^ between the input branch signals, y-axis is the average

output SNR. The noise power per branch is assumed 1 for all branches, respectively. Two cases of branch input signal power levels are taken as typical examples: the case of the same power levels where the average input

powers are for both branches (Figure 6 (a)) and the case of a large power level difference where the average input powers are

Q₁=O-Ol Ω₂=l and , respectively (Figure 6 (b)). Although the former is dominant, the latter is sufficiently likely to happen. For example, for a mobile with dual antennas, when one antenna is blocked by the hand during conversation, the input power of the blocked branch will be so weak that performance of the conventional EGC will be degraded. This is because the conventional EGC would combine all the branches regardless of noisiness of some weak branches. In this case, the solution described in this invention can make a difference.

<92> As shown in Figure 6 (a) and Figure 6 (b), MRC always shows the best performance of all diversity methods for any wireless channel conditions.

<MRO)_{=2 Q} -Q ₌₁

The ⁷ when (Figure 6 (a)) and -⁽MR^C^_{1 01} Ω^O.υi Ω₂=i when and (Figure 6 Cb)), which are the sums of input SNRs of both branches, respectively. The ^{γ "}of the invented HS/EGC is a little less than that of MRC. However, it is always greater than that of EGC and SC because HS/EGC takes the best of the two according to the wireless channel conditions. On the other hand, the performance order for EGC and SC changes according to ' and ² conditions. It is known that EGC shows better performance than SC in most cases. However, EGC can be inferior to SC when average power of one branch is significantly smaller than the other as shown in Figure 6 (b). Considering the poor performance of EGC in the case of a large power level difference, the HS/EGC can be a good solution which takes advantage of both the schemes.

<93> Although the MRC is the best solution regardless of any wireless channel conditions, the scheme requires a control means on both amplitudes and phases of input branch signals, which increases the system complexity. In particular, the amplitude control becomes a serious problem when the dynamic range of the wireless channel fading is large (e.g., > 60 dB) because it will require an additional automatic gain control (AGO system with as large dynamic range for each branch. Moreover, if fast acquisition is required to the beamforming system, the design of the AGC gets more difficult because the AGC normally tracks the average envelope of the input signal. On the contrary, the invented HS/EGC requires only phase control means with some switching control means as described in the above. Nontheless, the performance of HS/EGC is close to MRC for any channel conditions as shown in Figure 6 (a) and Figure 6 (b) . This is a merit of the invented HS/EGC compared with other diversity schemes.

<94> Embodiment examples'

<95> As an embodiment example of the invented HS/EGC scheme, a beamforming system with the HS/EGC scheme is shown in Figure 7 in a generic form. This is for

the receiving (RX) mode. The system is composed of JV antennas (301.1- 301.N), TV phase adjustment blocks (303.1-303.N), JV switches (305.1-

305.N), a phase control block (309), a switching control block (311) and a combiner (307). For the purpose of brevity, any other blocks in usual RF chains like low noise amplifier (LNA), down-converter, filters are omitted in

the diagram. The antennas receive RF signals from an external signal source. If the receiver is a mobile station, the signal source can be a base

station, and vice versa. The phase control block (309) generates phase control signals (315.1-315.N) according to an appropriate method. The phase control signals are applied to the phase adjustment blocks (303.1-303.N) and adjust the phase of input signals, respectively, so that all its output signals have aligned phases. The phase-locked loop (PLL) is a typical example of the phase control block and the phase adjustment block. The switching

control block (311) generates switching control signals (317.1-317.N) to select appropriate branches for combining according to the invented HS/EGC algorithm as explained in the previous descriptions and Figure 2 to Figure 5. The combiner (307) combines the selected branches, and generates the combined output (313). Here, the order of the phase adjustment blocks and the switches does not matter. Although the phase adjustment blocks precede the switches in the diagram, it should be recognized that it is equally possible to configure so that the switches precede the phase control blocks.

<96> The principle of the invented HS/EGC scheme can be easily extended to the transmission (TX) mode. Particularly, the transmission version of the HS/EGC scheme is called hybrid selection/equal gain transmission (HS/EGT) in this invention, and shown in Figure 8 in a generic form. The structure of HS/EGT is the same as that of HS/EGC except that the combiner (307) of HS/EGC in Figure 7 is replaced by a common transmission input point (363). Particularly, if the underlying communication system operates in the time- division duplexing (TDD) environment for RX and TX duplexing, the phase control information obtained in RX mode can be re-used for TX mode because wireless channel conditions are the same for both RX and TX modes in TDD. This is a great merit of TDD system, and this invented HS/EGT scheme can take the most of it. In other words, the phase control information (315.1-315.N) and the switching control information (317.1-317.N) generated in HS/EGC can be reused for the phase control information (365.1-365.N) and the switching control information (367.1-367.N) from the phase control block (357) and the switching control block (359), respectively, for HS/EGT. The phase control information may be calibrated, if necessary, to compensate for non-negligible and non-uniform phase delay difference between the RX path and TX path for each branch. The effect of HS/EGT appears as an enhanced reception SNR at the opposite communication side when the signals transmitted from the selected branches are received and coherently superimposed at the opposite communication side. Here, the order of the phase adjustment blocks and the switches does not matter. Although the phase adjustment blocks precede the switches in the diagram, it should be recognized that it is equally possible to configure so that the switches precede the phase control blocks. <97> As another embodiment example of the invented HS/EGC scheme, a RX beamforming system with the HS/EGC scheme is shown in Figure 9 in a more detailed form than the generic one in Figure 7. Note that there can be many other examples which may be realized based on the idea disclosed in this invention.

<98> The system is composed of JV antennas (601.1-601.N), a router (603),

TV- 1 phase-locked loops (605.2-605.N), JV- 1 switches (607.2-607.N) , a switching control block (611) and a combiner (609). For the purpose of brevity, any other blocks in usual RF chains like low noise amplifier (LNA), down-converter, filters are omitted in the diagram. Comparing Figure 7 and Figure 9, it is recognized that the phase adjustment blocks (303.1-303.N) and the phase control block (309) in Figure 7 are replaced by phase-locked loops (605.2-605.N) in the dashed box (606), and a router (603) is introduced in Figure 9 and controlled by the routing control signals (621.1-621.N) from the switching control block (611), and the /V switches (305.1-305.N) in Figure

7 is reduced to JV-I switches (607.2-607.N) in Figure 9. The other parts remain the same. However, this difference does not matter as far as the basic idea of the invention is concerned. It should be recognized that the configuration in Figure 9 is just an embodiment example to realize the generic form in Figure 7.

<99> For convenience,we call the branches before the router the "pre-router" branches (615.1-615.N) and those after the router the post-router branches (617.1-617.N), respectively. According to the command from the switching control block (611), the router connects the pre-router branches to the post- router branches so that the first post-router branch has the largest signal level and the second post-router branch has the second largest, and so on. The signals from the post-router branches are combined at the combiner (609) after phase being aligned through the operation of the phase-locked loops (605.2-605.N). The first post-router branch (617.1) is combined automatically without any phase adjustment, and it is applied to the reference input ports (620.2-620.N) of the phase-locked loops for other branches so that its phase acts as a reference phase. This is reasonable because the first post-router has the strongest signal power. Other post-router branches try to align the phases of their PLL outputs (619.2-619.N) to that of the first post-router branch (617.1), respectively, according to the phase-locked loop principle. -

<ioo>

<ioι> At the same time, the switching control block (611) determines which post- router branches to include in the combining, and this is realized through the switches (607.2-607.N). The switching control block realizes the invented HS/EGC algorithm according to the flow charts shown in Figure 2 to Figure 5. It takes in the pre-router branch inputs (615.1-615.N) and arranges them in the descending order in terms of signal magnitude (or, power) and determines which branches to include in the combining. This is carried out by sending the routing control signals (621.1-621.N) to the router and sending the switching control signals (623.2-623.N) to the switches, respectively. Finally, the combiner (609) combines the selected and phase-aligned branch signals to produce the coherently combined output (613) with enhanced SNR. <1O2> As another embodiment example of the invented HS/EGT scheme, a TX beamforming system with the HS/EGT scheme is shown in Figure 10 in a more detailed form than the generic one in Figure 8. Note that there can be many other examples which may be realized based on the idea disclosed in this invention.

<1O4> The system is composed of antennas (651.1-651.N), a router (655),

TV- 1 phase adjustment blocks (657.2-657.N), JV- 1 switches (658.2-

658.N), a phase control block (661), and a switching control block (659). For the purpose of brevity, any other blocks in usual RF chains like power amplifier (PA), up-converter, filters are omitted in the diagram. Comparing

Figure 8 and Figure 10, it is recognized that the phase adjustment

blocks (353.1-353.N) in Figure 8 are reduced to JV-I phase adjustment blocks (657.2-657.N) in Figure 10. Also, a router (655) is introduced in Figure 10 and controlled by the routing control signals (675.1-675.N) from

the switching control block (659), and the Tv switches (355.1-355.N) in

Figure 8 is reduced to JV-I switches (658.2-658.N) in Figure 10. The other parts remain the same. However, this difference does not matter as far as the basic idea of the invention is concerned. It should be recognized that the configuration in Figure 10 is just an embodiment example to realize the generic form in Figure 8. <iO6> For convenience, we call the branches before the router the

"pre-router" branches (669.1-669.N) and those after the router the post- router branches (671.1-671.N), respectively. A transmission input signal

(663) branches to JV identical pre-router signals (667.1-667.N) . The first pre-router signal (667.1) is applied to the router automatically without any phase adjustment, but the second to the N-th pre-router signals are cήecked for connection by the switches (658.2-658.N) under ^'the control of the switching control signals from the switching control block (659), and phase- adjusted by the phase adjustment blocks (657.2-657.N) under the control of the phase control signals (673.2-673.N) from the phase control block (661).

The router connects the JV pre-router signals to the JV post-router signals under the control of the routing control signals (675.1-675.N) from the switching control block (659). Finally, the signals from the post-router branches are radiated from the antennas and received at the opposite communication side producing a coherently superimposed signal with enhanced SNR.

<iO8> In the TDD environment, the commanding signals such as the phase control signals (673.2-673.N) from the phase control block (661), and the routing control signals (675.1-675.N) and the switching control signals (674.2-674.N) from the switching control block (659) may be inherited from those stored in the phase-locked loops (605.2-605.N) and the switching control block (611) in Figure 9, respectively, as will be explained shortly.

<no> As another embodiment example of the invented HS/EGC scheme, a RX and TX beamforming system with both HS/EGC and HS/EGT schemes is shown in Figure 11 in a more detailed form than the generic ones in Figure 7 and Figure 8. This system is a composition of the RX beamforming system with the HS/EGC in Figure 9 and the TX beamforming system with the HS/EGT in Figure 10 sharing common multiple antennas for both receiving and transmission. This configuration is particularly useful in time-division duplexing (TDD) environment where receiving and transmission modes are separated by time while using the same frequency band, thus, the channel conditions are the same for both modes. This means that the phase control and switching control information obtained in the RX beamforming can be re-used for the TX beamforming. The information can be re-used without any modification if the receiving path and transmission path in the beamforming system are wel l matched. Even i f not , the informat ion can be re-used after a proper cal ibrat ion.

<ii2> The system i s composed of ^"^ antennas (701.1-701. N) , RF swi tch boxes

(703.1-703.N), 7V low noise amplifiers (LNA) (705.1-705.N), a switching

control block (719), a RX router (707), JV-I phase-locked loops (709.2-

709.N), /V- 1 RX switches (711.2-711.N) , a combiner (713), TV- 1 TX

switches (720.2-720.N), 7V- 1 phase adjustment blocks (721.2-721.N), a TX

router (723), JV power amplifiers (PA) (725.1-725.N) and an optional dummy LNA (715) with a load (717). Here, extra blocks such as RF switch boxes, LNAs, power amplifiers equipped in a typical TDD system are added to the simpler block diagrams in Figure 9 and Figure 10 in order to present more realistic TDD-based beamforming system. The dummy LNA and the load are introduced to measure the noise power which is necessary for some HS/EGC algorithm as shown in Figure 5. The TX switches (720.2-720.N) may be moved to the power amplifiers (725.1-725.N) with turn-on/off capability. Because this variation is obvious, we avoid describing the details for this case. Instead, we stick to the current configuration shown in Figure 11.

<ii4> In the receiving mode, the RF switch boxes (703.1-703.N) are configured by an external RX/TX selection signal (749) so that the RX signals received at the antennas can be routed to the LNAs, respectively. The low noise-amplified output signals (731.1-731.N) from LNAs are applied to the switching control block together with the dummy LNA output (737). By terminating the input of the dummy LNA with the same load (717) as would be seen by other LNAs and matching the same electrical characteristics for all LNAs, the dummy ..LNA produces the independently and identically distributed noise (737) as other LNAs would produce (731.1-731.N). Note that the output from the dummy LNA contains only noise component while the outputs of the other LNAs contain both signal and noise components. Therefore, it is much easier to measure the noise power using the dummy LNA. The switching control block (719) determines how to arrange the branch signals in the descending order and which branches to include in the combining based on the measurement of the magnitude or power of the input branch signals and the optional noise. This operation has already been explained algorithmically in Figure 2 and Figure 4. To achieve the goal, the switching control block generates suitable commanding signals for the routers and switches as will be explained below.

<ii5> According to the RX routing control signals (739.1-739.N) from the switching control block, the RX router (707) connects the pre-router branches (731.1- 731.N) to the post-router branches (733.1-733.N) so that the first post- router branch (733.1) has the largest signal level and the second post-router branch (733.2) has the second largest signal level, and so on. The signals from the post-router branches are combined at the combiner (713) after phase being aligned through the operation of the phase-locked loops (709.2-709.N) . The first post-router branch is combined automatically without any phase adjustment, and it is applied to the reference input ports (736.2-736.N) of the phase-locked loops for other branches so that its phase acts as a reference phase. This is reasonable because the first post-router has the strongest signal power. Other post-router branches try to align the phases of their PLL outputs (735.2-735.N) to that of the first post-router branch (733.1) according to the phase-locked loop principle. At the same time, under the control of the RX switching control signals (741.2-741.N) from the switching control block, the RX switches (711.2-711.N) connect or disconnect the PLL outputs to the combiner. Finally, the combiner (713) combines the selected phase-aligned signals to produce the coherently combined output signal (727) with enhanced SNR.

<ii7> In the transmission mode, the RF switch boxes (703.1-703.N) are configured by an external RX/TX selection signal (749) so that the TX signals from the power amplifiers can be routed to the antennas, respectively. At first, a transmission input signal (729) branches to -^ identical pre-router signals (751.1-751.N). The first pre-router signal (751.1) is applied to the TX router automatically without any phase adjustment, but the second to the N-th pre-router signals are checked for connection at the TX switches (720.2- 720.N) under the control of the TX switching control signals (745.2-745.N) from the switching control block (719), and phase-adjusted at the phase adjustment blocks (721.2-721.N) under the control of the TX phase control signals (747.2-747.N) from the PLLs (709.2-709.N). Because each phase-locked loop keeps the phase information for the RX beamforming in a proper storage device (for example, a capacitor), the information can be re-used for the TX beamforming under the TDD environment. The phase adjustment can be achieved by using a voltage-controlled phase shifter (VCPS), for example. Then, the TX router (723) routes the ** pre-router signals (753.1-753.N) to the ^ post- router signals (755.1-755.N) under the control of the TX routing control signals (743.1-743.N) from the switching control block (719). The commanding signals such as RX routing control signals (739.1-739.N), RX switching control signals (741.2-741.N) and RX phase control signals (inside the PLL) used for the RX beamforming are re-used for those such as TX routing control signals (743.1-743.N), TX switching control signals (745.2-745.N) and TX phase control signals (747.2-747.N) in TX beamforming. Finally, the signals from the post-router branches go through the RF switch boxes, and are radiated from the antennas and received at the opposite communication side producing a coherently superimposed signal with enhanced SNR.

<ii9> As another embodiment of the invented HS/EGC scheme, a RX and TX beamforming system with both HS/EGC and HS/EGT schemes is shown in Figure 12 in a more detailed form than that in Figure 11. As the same as in Figure 11, the TDD environment is assumed. For the purpose of brevity, the system employs two

branches only ( ). However, it should be recognized that the idea

disclosed in this invention can be easily extended to any integer JV case in the form shown in Figure 11.

<i2i> The system is composed of two antennas (801.1-801.2), two RF switch boxes (803.1-803.2), two LNAs (805.1-805.2), a switching control block (817), a RX router (807), a phase-locked loop (811), a RX switch (813), a combiner (815), a TX switch (834), a phase adjustment block (835), a TX router (837), two power amplifiers (843.1-843.2) and an optional dummy LNA (829) with a load (831). The switching control block (817) generates some switching control signals such as RX routing control signal (847), RX switching control signal (849), TX routing control signal (848) and TX switching control signal (850) which are necessary for the HS/EGC and HS/EGT operations. The switching control block is composed of power detector 1 (819.1), power detector 2 (819.2), an optional power detector 3 (833), comparator 1 (821.1), comparator 2 (821.2), and an internal router (823). The internal router inside the switching control block carries the same operation for the detected power levels (869.1-869.2) of the RX branch signals as the RX router (807) does for the RX branch signals (853.1-853.2). It is composed of the multiplexer 3 (825.1), multiplexer 4 (825.2) and an inverter (827). The algorithmic operation of the switching control block has been explained in Figure 5. Reiterating its operation in association with the engaged components, first, the power detector 1 (819.1) and power detector 2 (819.2) measures the input

signal plus noise power and from the branch 1 and branch 2 (853.1-853.2), respectively. Optionally, at the same time, the optional power

detector 3 (833) measures the noise power from the dummy LNA (829)

^l ^F2 with a load (831). Then, the comparator 1 (821.1) compares and , and generates the RX routing control signal (847) which is HIGH if

p{>p₂ and LOW if

Then, the RX routing control signal is applied to the internal router (823). The multiplexers are designed to connect the first input (InI) to the output (Out) if Select is HIGH, and to connect the second input (In2) to the output (Out) if Select is LOW. In this invention, the logic level HIGH activates "the associated function block, and vice versa. As a result, the internal router generates the largest power

level signal P _r ^L _i1 ^J at the first post-router branch (871.1) and the second

power level signal P \^L2^Jλ at the second post-router branch (871.2). Then,

the comparator 2 (821.2) compares "^• , "^{■ ■}" and ^ra in order to check the combining conditions as expressed in Equation (8) (or, Equation (9)). Then, the comparator 2 generates the RX switching control signal (849)

-P r-π> Jv i P r 1-]+ K.₇ P _m P τy{> ELΛ P _{\ \}Λ which is HIGH if ^L2J * UJ 2 ^B _(or> L^{j 1} L^lJ)

with- ^{1 ~}^ * and ^{2 =} ^ ■^•• and LOW otherwise. When

input branch signals has sufficiently high SNR so that P[1] » P^₇ the

Pn can be ignored and the optional dummy LNA and associated power detector 3 may be omitted to simplify the system. The TX routing control signal (848) and TX switching control signal (850) are obtained by holding the RX routing control signal (847) and RX switching control signal (849), respectively, at some stage when the RX mode turns over to the TX mode. This can be realized by using latch 1 (822.1) and latch 2 (822.2) controlled by an external hold signal (824). The RX router (807), which is composed of the multiplexer 1 (809.1), multiplexer 2 (809.2) and an inverter (810), connects the first and the second pre-router signals (853.1-853.2) to the first and the second post- router signals (855.1-855.2), respectively, when the RX routing control signal (847) from the switching control block (817) is HIGH. Similarly, the RX router connects the first and the second pre-router signals (853.1-853.2) to the second and the first post-router signals (855.2-855.1), respectively, when the RX routing control signal is LOW. The TX router (837), which is composed of the multiplexer 5 (839.1), multiplexer 6 (839.2) and""an inverter (841), carries out the similar operation for the TX pre-router signals (865.1-865.2) as the RX router does for the RX pre-router signals under the control of the TX routing control signal.

<125> In the receiving mode, the RF switch boxes (803.1-803.2) are configured so that the signals received at the antennas can be routed to the LNAs. The branch signals are low noise-amplified at the LNAs, and are applied to the RX router as pre-router signals (853.1-853.2). The RX routing control signal (847) is applied to the RX router (807) in order to connect the pre-router branch signal with the largest power to the first post-router branch signal (855.1), and to connect the pre-router branch signal with the second largest power to the second post-router branch signal (855.2), respectively. The first post-router branch (855.1) is automatically included in the combining without phase adjustment, and is applied to the reference input port (858) of the phase-locked loop (811) in the second post-router branch. The second post-router branch (855.2) tries to align the phase of its PLL output (857) to that of the first post-router branch (855.1) according to the phase-locked loop principle. This issue will be explained in more detail shortly. Then, the RX switching control signal (849) is applied to the RX switch (813) in the second post-router branch in order to determine the inclusion of the branch in the combining. Finally, the combiner (815) combines the selected and phase-aligned branch signals to produce the coherently combined output (845) with enhanced SNR.

<127> Block diagram of an embodiment of the phase-locked loop (811) in Figure 12 is shown in Figure 13. It is composed of a phase adjustment block (901), a phase detector (903), a loop filter (905), an integrator (907) and a latch (909). The phase adjustment block can be realized with a voltage-controlled phase shifter (VCPS), for example. The VCPS is a phase shifter which shifts the phase of the input signal (911) as much as commanded by a control voltage signal (917). As a result, the output signal (913) of the VCPS is a copy of the input signal (911) whose phase is shifted by the amount commanded by the control signal (917). An embodiment example of the VCPS is found in the following literature [Ref 2]:

<i29> [Ref 2]: B. Chun, "A circular voltage-controlled phase shifter? filed to Irish Patent Office, 2006.

<i3i> The input signal (911) may have any kind of modulated waveform around the center frequency. Assuming a certain value of the phase control signal (917), the output signal (913) has the same waveform as the input signal but a certain amount of phase shift (or, time delay). On the other hand, the output signal (913) and the reference signal (915) from the first post-router branch (855.1) are applied to the phase detector (903). Here, the phase of the output signal is compared with that of the reference signal, and the phase detector generates a phase error signal (921). The phase error signal is filtered through the loop filter (905) and applied to the integrator (907). Then, the output of the integrator is applied to the latch (909). In the RX beamforming mode, the hold signal (919) from an external source keeps LOW normally. Then, the latch passes the signal from the integrator to the phase adjustment block as it is so that the whole phase-locked loop operates in a closed loop. As a result, the phase of output signal (913) can be aligned to that of the reference signal (915). However, the hold signal switches to HIGH at some stage when the mode turns over to TX mode or during RX mode. Although, even in the RX mode, there is a possibility that the hold signal goes to HIGH to hold the PLL operation when the beamformer reaches the steady state, it is assumed that the hold signal goes to HIGH when the mode changes to the TX mode just for convenience. Then, the latch holds the phase control signal value in some memory device. As a result, the PLL opens and the phase control signal value keeps its value regardless of the input signal variation (e.g., no input signal in the TX mode).

<133> During the closed loop operation of the PLL, because the integrator accumulates the filtered phase error signal, the output (925) of -the integrator increases or decreases as long as the phase error signal is nonzero. In other words, the output of the integrator have a constant value only when the phase error signal is settled to zero, which means that the phase of the output signal (913) is aligned (or, phase-locked) to that of the reference signal (915). Note that the phase-locking operation happens normally regardless of the modulation of the input signal. The constant value (or, voltage) of the integrator output (925) contains the phase information for RX beamforming, and can be stored in a memory device such as a capacitor in the integrator or the latch by activating the hold signal (919). This phase information signal (917) will be re-used for TX beamforming in the TDD environment. In the transmission mode, the RF switch boxes (803.1-803.2) are configured by an external RX/TX selection signal (869) so that the signals from the power amplifiers can be routed to the antennas, respectively. At first, a transmission input signal (851) branches to two pre-router branch signals (863.1-863.2). The first pre-router signal (863.1) is applied to the TX router (837) automatically without any phase adjustment, but the second pre- router signal (863.2) is selected under the control of the TX switching control signal (850), and phase-adjusted by the TX phase control signal (861) from the phase-locked loop (811). Because the phase-locked loop keeps the phase information for the RX beamforming in an appropriate storage device, the information can be re-used for the TX beamforming under the TDD environment. The phase adjustment can be realized by using the same VCPS as the phase adjustment block (901) in the PLL (811). Then, the TX router connects the two pre-router signals (865.1-865.2) to the two post-router signals (867.1-867.2) under the control of the TX routing control signal (848) from the switching control block. The commanding signals used for the RX beamforming are re-used for the TX beamforming. The commanding signals such as RX routing control signal (847), RX switching control signal (849) and RX phase control signal (917) used for the RX beamforming are re-used for those such as TX routing control signal (848), TX switching control signal (850) and TX phase control signal (861) in TX beamforming mode. These control signals for TX mode is obtained by holding the corresponding control signals for RX mode at some stage when the RX mode turns over to the TX mode. This is controlled by an external hold signal (824). Finally, the signals from the selected post-router branches are power-amplifed by the power amplifiers (843.1-843.2) and go through the RF switch boxes. Then, they are radiated from the antennas and received at the opposite communication side producing a coherently superimposed signal with enhanced SNR.

<i37> The dual branch beamformer can be extended to the beamforming system with any number of branches. While the extension can take a parallel structure as shown in Figure 11, a tree structure of extension is possible as well. That is, the dual branch beamformer can be designed in a modular way so that it can be systematically extended to the beamforming system with any number of branches. The extension is realized by organizing the dual branch beamformer modules in a tree structure as shown in Figure 14.

<139> In Figure 14 (a), the block diagram of the dual branch beamformer module is shown. This corresponds to the subsystem inside the thick dashed box (800) in Figure 12. The beamformer module (1051) has two RX inputs (1053.1-1053.2), one RX output (1055), two TX outputs (1057.1-1057.2), one TX input (1059) and one hold signal input (1061). As an embodiment of the beamformer with invented HS/EGC and HS/EGT schemes, a four-branch tree-structured beamformer based on the dual-branch beamformer module is shown in Figure 14 (b) . The operation of the dual branch module has been explained before in detail, and the operation of the four-branch beamforming system is quite obvious. Explaining its operation in short, in the RX mode where the hold signal (1023) is LOW, signals from antenna 1 and 2 (1001.1-1001.2) are combined in the HS/EGC way through the first dual branch beamformer module (1031.1). The module produces an intermediate combined RX signal (1015.1). At the same time, signals from antenna 3 and 4 (1001.3-1001.4) are combined in the HS/EGC way through the second dual branch beamformer module (1031.2). The module produces another intermediate combined RX signal (1015.2). The intermediate combined RX signals are again combined in the HS/EGC way through the third dual branch beamformer module (1031.3). The module produces the final combined RX signal (1019). In- the TX mode where the hold signal (1023) is HIGH, the TX beamforming is carried out based on the pieces of information obtained in the RX mode. Because its operation is obvious, its explanation is omitted. Meanwhile, recalling that each dual branch beamformer module has an effect of reducing the number of branches by one, any number of branches (say, N) can be combined in a tree-structure using (N-I) dual branch beamformer modules.

<14O> The invention is not limited to the embodiments described but may be varied in construction and detail. For example, the invented HS/EGC and HS/EGT algorithms can be applied to the RAKE receiver and the pre-RAKE transmitter, respectively, in code-division multiple access (CDMA) communication system where the multiple channels are defined as distinguishable multipaths, instead of antenna branches. The similar configurations can be derived based on the same idea as the beamforming system which has been mainly described in this invention. [Industrial Applicability]

<i4i> This invention relates to a wireless communication system, and more specifically, to a diversity or beamforming technology, which may be equipped in a base station or a mobile station.

Claims

[CLAIMS] [Claim 1] <143> A signal combining method in a multiple channel system configured to select a subset of total branches dynamically during operation so as to result in the maximum output SNR out of all possible selections when the selected branches are combined in EGC manner, and to combine the selected branches in EGC manner .

[Claim 2]

<i44> Said signal combining method in Claim 1 configured to operate according to ' the way that: first, measuring input signal magnitude or power for each branch, <145> next, arranging said input signal magnitudes or powers in the descending order so as to generate an ordered branch signal magnitudes or powers, <146> next, starting from the branch with the largest magnitude, including said ordered branch signals subsequently until said output SNR which would be obtained by combining the selected ordered branch signals in EGC manner does not increase anymore, <147> next, selecting said ordered branch signals which results in the highest output SNR. <i4S> finally, combining said selected ordered branch signals in EGC manner, and repeating the above process.

[Claim 3] <i49> Said signal combining method in Claim 1 configured to operate according to the way that :

<15O> first, measuring input signal magnitude or power for each branch, <i5i> next, arranging said input signal magnitudes or powers in the descending order so as to generate an ordered branch signal magnitudes or powers, <i52> next, checking if said ordered branch signal magnitude or power exceeds a threshold level subsequently from the one with the largest magnitude or power until said ordered branch signal magnitude or power does not exceed said threshold level, <i53> next, selecting said ordered branch signals which exceeds said threshold level,

<i54> finally, combining said selected ordered branch signals in EGC manner, and repeating the above process.

[Claim 4]

<155> Said threshold level in Claim 3 for the _-th largest branch signal is

determined by

where ^[LA^]J stands for the k-th largest signal magnitude in the total branches.

[Claim 5] <i56> A signal combining method with two branches configured to operate according to the way that :

<i57> first, measuring input signal magnitude or power for each branch, <i58> next, choosing the branch with the largest signal magnitude or power. for inclusion in the combining automatically, and <i59> next, including the branch with the second largest signal magnitude or power in the combining if the second largest signal magnitude or power is larger than a threshold value, and rejecting said branch with the second largest magnitude or power otherwise. <i60> finally, combining said selected branch signals in EGC manner, and repeating the above process.

[Claim 6] <i6i> Said threshold value in Claim 5 in terms of magnitude for inclusion of the

(γ2-l)α_ril α _ril second largest branch is determined by where ^L means the largest branch signal magnitude.

[Claim 7] <i62> Said threshold value in Claim 5 in terms of power for inclusion of the second largest branch is determined by

where

rπ means the largest branch signal plus noise power and n means the noise power.

[Claim 8] <i63> Said threshold value in Claim 5 in terms of power for inclusion of the second

largest branch is determined by

the input SNR is

sufficiently high where ^[1] means the largest branch signal plus noise power.

[Claim 9]

<164> A signal transmission method in a multiple channel system configured to select a subset of all branches so as to result in the maximum reception SNR out of all possible selections when phases of said selected branches are adjusted in EGT manner, and the signals from the selected branches are transmitted from said system, and received and superimposed at the opposite communication side.

[Claim 10]

<165> Said signal transmission method in Claim 9 is configured to adjust the phase of each branch and select branches based on the information of phase control and switching control obtained in receiving mode.

[Claim 11]

<166> A signal combining device in a multiple channel system such as RX beamformer or RAKE receiver or any equivalent composed of multiple RX input ports, multiple RX phase adjustment blocks, multiple RX switches, a RX phase control ^• block, a RX switching control block, a combiner and a RX output port wherein:

<i67> Said RX input ports are configured to receive multiple channel input signals, and

<168> Said RX switching control block is configured to generate RX switching control signals so that a subset of total branches be selected according to Claim 1, and <169> Said RX phase control block is configured to generate RX phase control signals so that phases of said RX input signals be aligned according to Claim 1, and <i70> Said RX switches are configured to switch said RX input signals under the control of said RX switching control signals, and to provide selected RX input signals for said combiner, and <i7i> Said RX phase adjustment blocks are configured to adjust phases of said RX input signals under the control of said RX phase control signals, and <i72> Said combiner is configured to combine said selected RX input signals, and <i73> Said RX output port is configure to output said combined RX output .

[Claim 12] <i74> A signal transmission device in a multiple channel system such as TX beamformer or pre-RAKE transmitter or any equivalent composed of multiple TX output ports, multiple TX phase adjustment blocks, multiple TX switches, a TX phase control block, a TX switching control block and a TX input port wherein: <175> Said TX input port is configured to receive a TX signal, and branch it to multiple identical TX signals, and <i76> Said TX switching control block is configured to generate TX switching control signals so that a subset of total branches be selected according to Claim 9, and <i77> Said TX phase control block is configured to generate TX phase control signals so that phases of said TX signals be adjusted according to

Claim 9, and <178> Said TX switches are configured to switch said TX signals under the control of said TX switching control signals, and to provide selected TX signals for said TX output ports, and <179> Said TX phase adjustment blocks are configured to adjust phases of said TX signals under the control of said TX phase control signals, and <i80> Said TX output ports are configured to transmit said selected TX output signals to the wireless channel or some next stage.

[Claim 13] <i8i> A signal combining device in a multiple channel system such as RX beamformer or RAKE receiver or any equivalent composed of multiple RX input ports, a RX router, multiple phase-locked loops, multiple RX switches, a RX switching control block, a combiner and a RX output port wherein: <i82> Said RX input ports are configured to receive multiple channel input signals, and <1S3> Said RX switching control block is configured to generate RX routing control signals and RX switching control signals so that a subset of total branches be selected according to Claim 1, and <i84> Said RX router is configured to route said RX input signals (RX pre-router signals) to RX post-router signals in the descending order in terms of signal magnitude or power under the control of said RX routing control signals, and <i85> Said phase-locked loops are configured to align their output phases to a reference phase, and <186> Said RX switches are configured to switch said RX post-router signals under the control of said RX switching control signals, and to provide selected RX post-router signals for the said combiner, and <187> Said combiner is configured to combine said selected RX post- router signals, and <188> Said RX output port is configure to output said combined RX output .

[Claim 14] <189> Said signal combining system in Claim 13 wherein said reference phase is provided from the first post-router signal with the strongest signal power out of said RX post-router signals.

[Claim 15] <19O> Said signal combining device in Claim 13 wherein said first post-router signal with the strongest signal power is combined in said combiner without any phase adjustment.

[Claim 16] <i9i> A signal transmission device in a multiple channel system such as TX beamformer or pre-RAKE transmitter or any equivalent composed of multiple TX output ports, a TX router, multiple phase adjustment blocks, multiple TX switches, a TX switching control block, a TX phase control block and a TX input port wherein: <I92> Said TX input port is configured to receive a TX signal, and branch it to multiple identical TX signals, and <i93> Said TX switches are configured to switch said TX signals under the control of said TX switching control signals, and to provide selected TX signals for said TX output ports, and <194> Said TX phase control block is configured to generate TX phase control signals so that phases of said TX signals be adjusted according to

Claim 9, and <195> Said TX phase adjustment blocks are configured to adjust phases of said TX signals under the control of said TX phase control signals, and <196> Said TX switching control block is configured to generate TX routing control signals and TX switching control signals so that a subset of total branches be selected according to Claim 9, and <197> Said TX router is configured to route said TX signals (TX pre- router signals) to TX post-router signals in the descending order in terms of signal magnitude or power under the control of said TX routing control signals, and <198> Said TX output ports are configured to transmit said selected TX post-router signals to the wireless channel or some next stage.

[Claim 17] <199> A signal combining and transmission device in a multiple channel system such as TDD-based beamformer or TDD-based RAKE system or any equivalent composed of multiple RX input ports, a RX router, multiple phase-locked loops, multiple RX switches, a switching control block, a combiner and a RX output port, and multiple TX output ports, a TX router, multiple phase adjustment blocks, multiple TX switches and a TX input port wherein:

<2oo> Said RX input ports are configured to receive multiple channel input signals, and

<20i> Said switching control block is configured to generate RX routing control signals and RX switching control signals so that a subset of total branches be selected according to Claim 1, as well as to generate TX routing control signals and TX switching control signals so that a subset of total branches be selected according to Claim 9, and

<202> Said RX router is configured to route said RX input signals (RX pre-router signals) to RX post-router signals in the descending order in terms of signal magnitude or power under the control of said RX routing control signals, and

<203> Said phase-locked loops are configured to align their output phases to a reference phase as well as to generate RX phase control signals, and

<204> Said RX switches are configured to switch said RX post-router signals under the control of said RX switching control signals, and to provide selected RX post-router signals for the said combiner, and

<205> Said combiner is configured to combine said selected RX post- router signals, and

<206> Said RX output port is configure to output said combined RX output , and

<207> Said TX input port is configured to receive a TX signal, and branch it to multiple identical TX signals, and

<208> Said TX switches are configured to switch said TX signals under the control of said TX switching control signals, and to provide selected TX signals for said TX output ports, and <209> Said TX phase adjustment blocks are configured to adjust phases of said TX output signals under the control of some TX phase control signals, and <2io> Said TX router is configured to route said TX output signals (TX pre-router signals) to TX post-router signals in the descending order in terms of signal magnitude or power under the control of said TX routing control signals, and <2ii> Said TX output ports are configured to transmit said selected TX post-router signals to the wireless channel or some next stage, and <2i2> Said RX phase control signals, RX routing control signals and RX switching control signals are re-used for TX phase control signals, TX routing control signals and TX switching control signals, respectively.

[Claim 18] <2i3> Said signal combining and transmission device in Claim 17 wherein said reference phase is provided from said first RX post-router signal with the strongest signal power out of said RX post-router signals.

[Claim 19] <2i4> Said signal combining and transmission device in Claim 17 wherein said first post-router signal with the strongest signal power is combined in said combiner without any phase adjustment.

[Claim 20] <2i5> Said signal combining and transmission device in Claim 17 wherein said RX switches, said RX switching control signals, said TX switches and said TX switching control signals are omitted or ignored.

[Claim 21] <2i6> Said signal combining and transmission device in Claim 17 wherein said TX control signals such as TX phase control signals, TX routing control signals and TX switching control signals are obtained by storing said RX control signals such as RX phase control signals, RX routing control signals and RX switching control signals in appropriate memory devices such as capacitors and latches under the control of an external hold signal .

[Claim 22]

<218> Said signal combining and transmission device in Claim 21 wherein said TX phase control signals are obtained by calibrating said RX phase control signals if there is non-negligible and non-uniform phase delay difference between the RX path and TX path for each branch.

[Claim 23]

<2i9> A signal combining and transmission device in a dual channel system such as TDD-based beamformer or TDD-based RAKE system composed of two RX input ports, one RX output, port, two TX output ports, one TX input port, a hold signal port, and a RX signal processing block, a switching control block and a TX signal processing block wherein:

<220> said RX signal processing block is composed of said two RX input ports, said one RX output port, a RX router, a phase-locked loop, a RX switch and a combiner, and

<22i> said switching control block is composed of two RX input ports, control signal output ports such as a RX routing control signal, a RX switching control signal, a TX routing signal and a TX switching control signal, two power detectors (the first power detector and the second power detector), an optional power detector (the third power detector), an internal router, two comparators (the first comparator and the second comparator) and two latches (the first latch and the second latch), and

<222> said TX signal processing block is composed of two TX output ports, one TX input port, a TX switch, a phase adjustment block and a TX router.

[Claim 24]

<223> Said switching control block in Claim 23 is configured so as"-

<224> to generate said RX routing control signal in order to route said RX input signals (RX pre-router signals) to RX post-router signals according to Claim 5, and

<225> to generate said RX switching control signal in order to switch said second RX post-router signal according to Claim 5, and

<226> to generate said TX routing control signal in order to route said TX input signals (TX pre-router signals) to TX post-router signals, and

<227> to generate said TX switching control signal in order to switch said second TX pre-router signal.

[Claim 25]

<228> Said RX signal processing block in Claim 23 is configured so that:

<229> said RX input ports receive two RX input signals, and

<230> said RX router routes said RX pre-router signals to RX post-router signals under the control of said RX routing control signal, and

<23i> said phase-locked loop aligns the output phase of the second post-router signal to a reference phase and stores the phase control information in a memory device for re-use in TX mode, and

<232> said RX switch switches the second post-router signal under the control of said RX switching control signal, and

<233> said combiner combines said the first post-router signal and the second post-router signal to generate said RX output, and

<234> said RX output port outputs said combined RX output.

[Claim 26]

<235> Said TX signal processing block in Claim 23 is configured so that:

<236> said TX input port receives a TX signal and branches it to two identical TX signals (the first TX pre-router signal and the second TX pre- router signal), and

<237> said TX switch switches said second TX pre-router signal under control of said TX switching control signal, and

<238> said phase adjustment block adjusts the phase of the second TX pre-router signal under the control of said phase control information from said RX signal processing block, and

<239> said TX router routes said TX pre-router signals to TX post- router signals under control of said TX routing control signal, and

<240> said TX output ports transmits said TX post-router signals to the wireless channel or some next stage.

[Claim 27] <24i> Said phase reference in Claim 25 is provided from said first RX post-router signal .

[Claim 28] <242> Said phase-locked loop in Claim 25 is composed of an input signal port, an output signal port, a phase reference signal port, a hold signal port, a phase information signal port, a phase adjustment block, a phase detector, a loop filter, an integrator and a latch wherein: <243> said phase adjustment block adjusts the phase of said input signal according to said phase information signal, and generates said output signal, and <244> said phase detector detects the phase error between said output signal and said phase reference signal, and

<245> said loop filter filters said phase error signal, and

<246> said integrator integrates said phase error signal, and

<247> said latch passes said integrator output (said phase information signal) to said phase adjustment block in normal operation so that said phase-locked loop operates in a closed loop mode, and holds said integrator output (said phase information signal) under the control of said hold signal so as to store said integrator output (said phase information signal) in a memory device.

[Claim 29] <248> Said memory device in Claim 28 is a capacitor.

[Claim 30] <249> Said TX switching control signal and said TX routing control signal in Claim

26 are obtained by employing said first latch and said second latch in said switching control block in Claim 23 under the control of said hold signal.

[Claim 31] <250> A signal combining and transmission device in a dual channel system in Claim

23 wherein said RX switch, said TX switch, said RX switching control signal, said TX switching control signal, said internal router, said optional power detector (said third power detector), said second comparator and said second latch are omitted or ignored.

[Claim 32] <25i> A signal combining device in a dual channel system in Claim 23 wherein said

TX signal processing block is omitted or ignored.

[Claim 33] <252> A signal combining and transmission device in a multiple channel system composed of said dual channel signal combining and transmission devices in

Claim 23 configured in a tree structure.