Robust iterative linear system solvers
View Patent ↗In receiving equipment such as a mobile terminal ( 1 ) for decoding received signals coded according to a CDMA method are decoded in an intermediate processor ( 9 ). In the decoding, coefficients or elements of a vector w are calculated in a unit or module ( 13 ) by an iterative method as a solution to a system of linear equations represented by the matrix equation Rw=h, where R is a matrix of known elements and h is a known vector. For terminating the iteration, a value of the signal to interference plus noise ratio SINR for the received signals is calculated in each iteration step and then this calculated value or a quantity derived therefrom is used as a stopping criterion or generally in a stopping algorithm.
1. A method, in a receiving equipment for demodulating received signals, the method comprising:
performing a demodulation calculation, the performing further comprising
iteratively calculating, in a plurality of successive iteration steps, coefficients or elements of a vector w to obtain a solution to a system of linear equations represented by a matrix equation Rw=h, wherein
R is a matrix of known elements,
h is a known vector, and
the values of the coefficients or elements of the vector w are used by demodulating circuits in the receiving equipment to produce demodulated signals, and
for each successive iteration step among the plurality of successive iteration steps, calculating a signal to interference plus noise ratio (SINR) for the received signals, after the received signals are demodulated using the values of the coefficients or elements of vector w calculated in each successive iteration step; and
stopping the demodulation calculation using the SINR in a stopping condition or in a stopping algorithm.
2. The method according to claim 1 , wherein the demodulating is performed in a RAKE receiver, wherein that the vector h is composed of net channel coefficients, the matrix R is an impairment covariance matrix and the vector w is a vector of combining weights.
3. The method according to claim 1 , wherein the calculating the SINR further comprises:
calculating the value of the SINR is, in each successive iteration step among the plurality of successive iteration steps, according to the ratio of signal power, computed as a function of w i and h, to interference plus noise power, computed as a function of w i and R, where i is the order number of the iteration step and the vector w i is the approximate value of the vector w calculated in this iteration step.
4. The method according to claim 1 , further comprising
comparing the value of the signal to interference plus noise ratio SINR i calculated in the i th iteration step to the value of the signal to interference plus noise ratio SINR i−1 calculated in the directly preceding (i−1) th iteration step; and
stopping the demodulation calculation further comprises
stopping the demodulation calculation in the case where said value of the signal to interference plus noise ratio SINR i calculated in the i th iteration step is not greater than the value of the signal to interference plus noise ratio SINR i−1 calculated in the directly preceding (i−1) th iteration step, and
assigning the value of vector w i calculated in the i th iteration step is taken as a solution to the matrix equation.
5. The method according to claim 1 , wherein
a predetermined number N iter of iteration steps are performed, and
that the value of the vector w best corresponding to the maximum SINR max of the values of the signal to interference plus noise ratio SINR calculated in all of the successive iteration steps is taken as the solution to the matrix equation.
6. A receiving equipment for receiving signals, the receiving equipment comprising:
a decoding unit for demodulating received signals, the decoding unit further comprises
a unit for performing a demodulation calculation, the unit for performing further comprising
a unit for iteratively calculating, in a plurality of successive iteration steps, coefficients or elements of a vector w to obtain a solution to a system of linear equations represented by a matrix equation Rw=h, wherein
R is a matrix of known elements,
h is a known vector, and
the values of the coefficients or elements of the vector w are used by demodulating circuits in the receiving equipment to produce demodulated signals, and
a unit for calculating, for each successive iteration step among the plurality of successive iteration steps, a signal to interference plus noise ratio (SINR) for the received signals, after the received signals are demodulated using the values of the coefficients or elements of vector w calculated in each successive iteration step; and
a control unit for stopping the demodulation calculation using the SINR in a stopping condition or in a stopping algorithm.
7. The receiving equipment according to claim 6 , wherein the unit for calculating is adapted to calculate a value of the SINR, in each successive iteration step among the plurality of successive iteration steps, according to the ratio of signal power, computed as a function of w i and h, to interference plus noise power, computed as a function of w i and R, where i is the order number of the iteration step and the vector w i is the approximate value of the vector w calculated in this iteration step.
8. The receiving equipment according to claim 6 , wherein the decoding unit is a RAKE receiver, the vector h being composed of net channel coefficients, the matrix R being an impairment covariance matrix and the vector w being a vector of combining weights.
9. The receiving equipment according to claim 6 , further comprising:
a comparator coupled to or included in the control unit and coupled to the unit for calculating, the comparator adapted to compare the value of the signal to interference plus noise ratio calculated in the i th iteration step to the value of the signal to interference plus noise ratio calculated in the directly preceding (i−1) th iteration step;
the control unit being adapted to stop the iterative method in the case where the result of the comparing is that said value of the signal to interference plus noise ratio calculated in the i th iteration step is not greater than the value of the signal to interference plus noise ratio calculated in the directly preceding (i−1) th iteration step; and
the control unit being adapted to take the value of vector w i calculated in the i th iteration step as the solution to the matrix equation.
10. The receiving equipment according to claim 6 , further comprising
a memory for storing the value of the signal to interference plus noise ratio calculated in each iteration step.
11. The receiving equipment according to claim 6 , wherein
the control unit is adapted to command that always a predetermined number N iter of successive iteration steps are performed, and
the control unit is adapted to take the value of the vector w best corresponding to the maximum SINR max of the values of the signal to interference plus noise ratio calculated in all of said predetermined number N iter of successive iteration steps as a solution to the matrix equation.
12. The receiving equipment according to claim 11 , further comprising
a memory for holding the predetermined number N iter of iteration steps;
a memory for holding the maximum SINR max of the values of the signal to interference plus noise ratio SINR calculated in the iteration steps up to the current iteration step; and
a memory for holding the value of the vector w best corresponding to the maximum SINR max of the values of the signal to interference plus noise ratio.