CN-116599608-B - Efficient spectrum sensing realization method based on ratio of energy to minimum characteristic value

CN116599608BCN 116599608 BCN116599608 BCN 116599608BCN-116599608-B

Abstract

The invention relates to an efficient spectrum sensing realization method based on the ratio of energy to minimum eigenvalue, which comprises the steps of firstly sampling M antenna received signals for N times continuously to calculate a sampling covariance matrix R x (N), then calculating the minimum eigenvalue lambda of the sampling covariance matrix by using an inverse iteration method based on initial value estimation, and calculating a sensing decision quantity t=tr (R x (N))/lambda, wherein tr (R x (N)) is the received signal energy, secondly, calculating a sensing decision threshold gamma by using a cubic spline interpolation method based on the natural boundary condition, and finally, performing sensing decision, wherein when t > gamma, a main user signal is judged to exist, and when t < gamma, the main user signal is judged not to exist. The method has the advantages of high calculation efficiency, convenience for hardware realization and reliable sensing result, and is suitable for the main user signal detection problem with high real-time requirements in the cognitive radio system.

Inventors

YANG XI
LEI KEJUN
TAN YUHAO
LIAO YUTING
ZHANG GENG
WANG XUMING

Assignees

吉首大学

Dates

Publication Date: 20260505
Application Date: 20230407

Claims (4)

1. A high-efficiency frequency spectrum sensing realization method based on the ratio of energy to minimum characteristic value is characterized in that the method utilizes an inverse iteration method based on initial value estimation to efficiently solve the minimum characteristic value of a received signal sampling covariance matrix, and utilizes a cubic spline interpolation method based on natural boundary conditions to simplify the sensing judgment threshold calculation, firstly, continuously receiving signals of multiple antennas Subsampling, calculating a covariance matrix of the received signal samples And secondly, estimating the value by using the initial value based on the minimum characteristic value And calculating a minimum eigenvalue of the sampling covariance matrix by an inverse iteration method of (2), and calculating a perception judgment quantity according to the minimum eigenvalue, wherein, Representation of Is used for the track of (a), Indicating the number of antennas to be used, Representation of The sum of squares of all the elements of (a), Representation of Again, calculating the perception judgment threshold by a cubic spline interpolation method based on natural boundary conditions , wherein, The probability of the target false alarm is represented, Representation of And finally, carrying out perception judgment, namely judging that the main user signal exists when the perception judgment quantity is larger than a threshold, and judging that the main user signal does not exist when the perception judgment quantity is smaller than the threshold.
2. The efficient energy-to-minimum feature value-based spectrum sensing implementation method according to claim 1, characterized by the specific steps of: step 1. At the moment For a pair of Sampling the signal on the root receiving antenna to obtain Maintaining received signal vectors Wherein the superscript " "Representing matrix transpose operator, continuous sampling Next, obtain From which the sampling covariance matrix of the received signal is calculated ; Step 2, calculating the minimum eigenvalue of the sampling covariance matrix by using an inverse iteration method based on initial value estimation And calculates the perception decision quantity , Step 3, calculating a perception judgment threshold by using a cubic spline interpolation method based on natural boundary conditions , wherein, For the probability of a false alarm of the target, Is about A cubic spline interpolation polynomial based on natural boundary conditions; Step 4, making a perception judgment, if the perception judgment quantity is the same Greater than the perception decision threshold And if the primary user signal is not present, judging that the primary user signal is not present.
3. The efficient energy-to-minimum eigenvalue ratio-based spectrum sensing implementation method according to claim 2, wherein the specific step of calculating the minimum eigenvalue of the sampling covariance matrix by using the initial value estimation-based inverse iteration method in step 2 is as follows: Step 2.1. Initializing Non-zero vector dimension Setting the iteration precision And maximum number of iterations ; Step 2.2. matrix of pairs Performing triangular decomposition into , wherein, Representation of Is arranged in the lower triangular matrix of the (c), Representation of Upper triangular matrix of (a) and initial value estimated value of minimum eigenvalue , wherein, Representation of The sum of squares of all the elements of (a), Representation of Squaring the trace; Step 2.3. Placing , , wherein, Representing the approximate feature vector; step 2.4. , Representing taking absolute values for all elements of the vector, Representing the maximum value of all elements in the selected vector; step 2.5. ; Step 2.6 when Executing the steps 2.7-2.12; Step 2.7. Use of Obtaining vectors , wherein, Representing a temporary vector introduced in the calculation process; Step 2.8. Use of Obtaining vectors , wherein, Representing a temporary vector introduced in the calculation process; step 2.9. ; Step 2.10. Indicating the amount of orientation The corresponding value of the largest absolute value of the medium element; Step 2.11. If Then Outputting the minimum characteristic value and ending the calculation; step 2.12. ; Step 2.13 when Outputting the minimum eigenvalue of the sampling covariance matrix 。
4. An efficient implementation method of spectrum sensing based on the ratio of energy to minimum eigenvalue according to claim 2, characterized in that in said step 3 Is based on a free boundary condition The range of values and the corresponding coefficients are shown in Table 1: table 1 values ranges and interpolation coefficients 。

Description

Efficient spectrum sensing realization method based on ratio of energy to minimum characteristic value Technical Field The invention relates to a high-efficiency frequency spectrum sensing realization method based on the ratio of energy to a minimum characteristic value, belonging to the field of cognitive radio in a wireless communication technology. Background The spectrum sensing algorithm based on the ratio of energy to the minimum characteristic value is an important main user signal detection algorithm in cognitive radio, the detection process does not depend on prior information such as main user signals, wireless channels, noise variances and the like, and the spectrum sensing algorithm has the full-blind detection characteristic and is wide in application range. The algorithm utilizes the ratio of the received signal energy to the minimum eigenvalue of the sampling covariance matrix to construct a statistical decision quantity, and the theoretical decision threshold relates to the calculation of the Tracy-Widom inverse cumulative distribution function. Since the calculation of the minimum eigenvalue involves the decomposition of the covariance matrix of the received signal samples, the complexity of solving the minimum eigenvalue by the conventional method is the third power of the received signal dimension. In this case, the computational complexity of the algorithm will increase significantly with the number of receive antennas, the number of cooperating nodes, and the oversampling rate. Meanwhile, the calculation of the Tracy-Widom inverse cumulative distribution function cannot be simply solved through a closed expression, and a decision threshold value corresponding to the designated target false alarm probability can be obtained through a table look-up method or complex numerical calculation. Under the condition of limited computing and storage resources, the real-time requirement of spectrum sensing is difficult to meet. In this case, how to effectively improve the computational efficiency of the spectrum sensing algorithm based on the ratio of energy to minimum eigenvalue is of great research value. Disclosure of Invention The invention provides a high-efficiency frequency spectrum sensing realization method based on the ratio of energy to a minimum characteristic value. The method has the advantages of high calculation efficiency, convenience for hardware realization and reliable sensing result, and has good application value for detecting the real-time main user signal in the cognitive radio system. The improved frequency spectrum sensing realization method based on the ratio of energy to minimum eigenvalue utilizes an inverse iteration method based on initial value estimation of the minimum eigenvalue to solve the minimum eigenvalue of a received signal sampling covariance matrix so as to accelerate convergence and improve calculation accuracy, utilizes a cubic spline interpolation method based on free boundary conditions to simplify sensing judgment threshold calculation, firstly, continuously samples a multi-antenna received signal for N times to calculate a received signal sampling covariance matrix, secondly, utilizes the inverse iteration method based on initial value estimation to calculate the minimum eigenvalue of the sampling covariance matrix and construct sensing judgment quantity, thirdly, calculates a sensing judgment threshold through a cubic spline interpolation method based on free boundary conditions, and finally, carries out sensing judgment, namely, judges that a main user signal exists when the sensing judgment quantity is larger than the threshold, and judges that the main user signal does not exist when the sensing judgment quantity is smaller than the threshold. The method comprises the following specific steps: Step 1, sampling signals on M receiving antennas at a time N to obtain M multiplied by 1-dimensional receiving signal vectors x (N) = [ x 1(n),x2(n),...,xM(n)]T, wherein the superscript "T" represents a matrix transposition operator, and continuously sampling N times to obtain N receiving signal vectors x (1), x (2), Step 2, calculating a minimum eigenvalue lambda of a sampling covariance matrix by using an inverse iteration method based on minimum eigenvalue initial value estimation, and calculating a perception decision quantity t=tr (R x (N))/lambda, wherein tr (R x (N)) represents received signal energy; Step 3, calculating a perception judgment threshold by using a cubic spline interpolation method based on free boundary conditions Wherein P f is the target false alarm probability, and ψ (1-P f) is a cubic spline interpolation polynomial based on natural boundary conditions about 1-P f; And 4, performing perception judgment, namely judging that the main user signal exists if the perception judgment quantity t is larger than the perception judgment threshold gamma, and otherwise, judging that the main user signal does not exist. Wherein, the The specifi