Bit detection using code spectrum transform techniques for interference cancellation
View Patent ↗Spreading code detection methods applying autoregressive spectrum estimation (ARSE) techniques to code division multiple access (CDMA) wireless communication systems allowing the subscriber device to estimate the other users' codes. According to one aspect, ARSE techniques are provided using an augmented Wiener-Hopf solution on a received forward link CDMA signal. The ARSE techniques produce a code spectrum transform yielding a power spectral density versus spreading code index relationship from which the spreading codes in the received forward link signal may be identified. According to another aspect, a reduced rank auto-regression (AR) implementation is provided to produce a code spectrum transform efficiently with improved false code detection rate. With knowledge of other users' spreading codes, interference cancellation can be performed on the forward link of CDMA systems.
1. A method in a communication device for estimating spreading codes for other destination devices which are present in a signal that has been transmitted using code division multiple access techniques, and wherein the communication device has knowledge only of its own spreading code, comprising:
a. receiving said signal at said communication device and generating samples thereof;
b. computing a weight vector using auto-regression techniques based on samples of the received signal over time, and computing a power spectral density from the weight vector as a function of the full spreading code matrix to determine correlations among spreading codes used for bits present in the received signal, wherein computing the weight vector comprises computing a M+1 =∥ε∥ 2 (R M+1 ) −1 u M+1 , where R M+1 is a correlation matrix Y M+1 (Y M+1 ) H , Y M+1 is a matrix representing the received signal at time M+1, u M+1 is a unit vector and ∥ε∥ 2 is a mean-square prediction error; and
c. identifying from a peak in the power spectral density a spreading code associated with bits present in the received signal for another destination device.
2. The method of claim 1 , wherein (b) computing
comprises computing from the weight vector the power spectral density
P
y
(
N
)
=
∈
2
S
^
N
H
a
M
+
1
2
,
where S N =[Ŝ 1 Ŝ 2 . . . Ŝ N ] represents a steering matrix for all spreading codes in a set, those spreading codes in the set which are present in the received signal being indicated by a peak in the power spectral density Py(N).
3. The method of claim 2 , wherein (c) identifying comprises identifying from a peak in the power spectral density a spreading code associated with bits for another destination device present in the received signal.
4. The method of claim 1 , wherein (b) computing comprises computing the weight vector using auto-regression techniques in a reduced rank form with a multistage Wiener filter.
5. The method of claim 4 , wherein (b) computing comprises computing the weight vector using multistage Wiener filter recursive equations according to a minimum mean square error (MMSE) solution c MMSE =R −1 h 0 , where h 0 is a spreading code assigned to a device at which the signal is received, R is a correlation matrix YY H , and Y is a matrix representing the received signal.
6. A method for canceling interference in a signal transmitted using code division multiple access techniques comprising the method of claim 1 , and further comprising canceling interference with respect to bits for other destination devices in the received signal using the one or more spreading codes associated with other destination devices.
7. A device comprising:
a. a radio receiver that receives a wireless radio that has been transmitted using code division multiple access techniques and produces a baseband signal therefrom;
b. a baseband signal processor coupled to the radio receiver, wherein the baseband signal processor:
i. computes a weight vector a M+1 using auto-regression techniques from samples of the received signal over time, where a M+1 =∥ε∥ 2 (R M+1 ) −1 u M+1 , R M+1 is a correlation matrix Y M+1 (Y M+1 ) H , Y M+1 is a matrix representing the received signal at time M+1, u M+1 is a unit vector and ∥ε∥ 2 is a mean-square prediction error, and computes a power spectral density as a function of the full spreading code matrix from the weight vector to determine correlations among spreading codes used for bits in the received signal; and
ii. identifies from a peak in the power spectral density a spreading code associated with bits present in the received signal for another destination device.
8. The receiver device of claim 7 , wherein the baseband signal processor computes the power spectral density
P
y
(
N
)
=
∈
2
S
^
N
H
a
M
+
1
2
,
where S N =[Ŝ 1 Ŝ 2 . . . Ŝ N ] represents a steering matrix for all spreading codes in a set, those spreading codes in the set which are present in the received signal being indicated by a peak in the power spectral density Py(N).
9. The receiver device of claim 8 , wherein the baseband signal processor identifies spreading codes by identifying a peak in the power spectral density, wherein a peak corresponds to a spreading code associated with bits for another destination device present in the received signal.
10. The receiver device of claim 7 , wherein the baseband signal processor computes the weight vector using auto-regression techniques in a reduced rank form with a multistage Wiener filter.
11. The receiver device of claim 10 , wherein the baseband signal processor computes the weight vector using multistage Wiener filter recursive equations according to a minimum mean square error (MMSE) solution c MMSE =R −1 h 0 , where h 0 is a spreading code assigned to a device at which the signal is received, R is a correlation matrix YY H , and Y is a matrix representing the received signal.
12. The receiver device of claim 7 , wherein the baseband signal processor cancels interference with respect to bits for other destination devices in the received signal using the one or more spreading codes associated with other destination devices.
13. A method in a communication device for estimating spreading codes for other destination devices which are present in a signal that has been transmitted using code division multiple access techniques, and wherein the communication device has knowledge only of its own spreading code, comprising:
a. receiving said signal at said communication device and generating samples thereof;
b. computing a weight vector a M+1 using auto-regression techniques from samples of the received signal over time, where a M+1 =∥ε∥ 2 (R M+1 ) −1 u M+1 , R M+1 is a correlation matrix Y M+1 (Y M+1 ) H , Y M+1 is a matrix representing the received signal at time M+1, u M+1 is a unit vector and ∥ε∥ 2 is a mean-square prediction error;
c. computing a power spectral density as a function of the full spreading code matrix based on the weight vector to determine correlations among spreading codes used for bits in the received signal; and
d. identifying from a peak in the power spectral density a spreading code associated with bits for another destination device present in the received signal.
14. The method of claim 13 , wherein (c) computing comprises computing the power spectral density
P
y
(
N
)
=
∈
2
S
^
N
H
a
M
+
1
2
,
where S N =[Ŝ 1 Ŝ 2 . . . Ŝ N ] represents a steering matrix for all spreading codes in a set, those spreading codes in the set which are present in the received signal being indicated by a peak in the power spectral density Py(N).
15. The method of claim 13 , wherein (b) computing comprises computing the weight vector using multistage Wiener filter recursive equations according to a minimum mean square error (MMSE) solution c MMSE =R −1 h 0 , where h 0 is a spreading code assigned to a device at which the signal is received, R is a correlation matrix Y(Y) H , and Y is a matrix representing the received signal.
16. The method of claim 13 , wherein (c) computing comprises computing the power spectral density
P
y
(
N
)
=
∈
2
S
^
N
H
a
M
+
1
2
,
where S N =[Ŝ 1 Ŝ 2 . . . Ŝ N ] represents a steering matrix for all spreading codes in a set, those spreading codes in the set which are present in the received signal being indicated by a peak in the power spectral density Py(N).
17. The method of claim 13 , wherein (b) computing comprises computing the weight vector using multistage Wiener filter recursive equations according to a minimum mean square error (MMSE) solution c MMSE =R −1 h 0 , where h 0 is a spreading code assigned to a device at which the signal is received, R is a correlation matrix Y(Y) H , and Y is a matrix representing the received signal.
18. A method for canceling interference in a signal transmitted using code division multiple access techniques comprising the method of claim 13 , and further comprising canceling interference with respect to bits for other destination devices in the received signal using the one or more spreading codes associated with other destination devices.
19. The method of claim 13 , wherein (b) computing comprises computing the weight vector using auto-regression techniques in a reduced rank form with a multistage Wiener filter.