Detection of wideband interference
A method of detecting interference in a received sample vector using hidden Markov modelling by first estimating noise variance, where estimating noise variance comprises the steps of receiving a sample vector of noise and interference, sorting the sample vector in the frequency domain by order of increasing magnitude to produce an ordered vector, finding a sub-vector of the ordered vector that minimizes the distance from a noise measure, and estimating the noise variance.
1. A method of detecting unknown wideband interference, the method implemented by a blind radio receiver having a processor that is configured to carry out the steps of: receiving a sample vector X of noise and interference, finding a set of parameters that maximise p(X|T, P, I), where T is the transition probability between states, P is the observation probability and I is the initial state, and wherein the step of finding the set of parameters that maximise p(X|T, P, I), includes a step of estimating noise variance, determining an optimal state sequence from the set of parameters, and determining wideband interference from the optimal state sequence.
2. The method according to claim 1 wherein the step of estimating noise variance comprises receiving a sample vector of noise and interference, sorting the sample vector in the frequency domain by order of increasing magnitude to produce an ordered vector, finding a sub-vector of the ordered vector that minimises the distance from a noise measure, and estimating the noise variance, wherein both the noise and interference are zero mean.
3. The method according to claim 1 wherein the step of finding a set of parameters that maximise p(X|T, P, I) further comprises the steps of producing a nominal interference variance and initialising the optimal state sequence based on the nominal interference variance.
4. The method according to claim 3 wherein the nominal interference variance is
{circumflex over (σ)} I 2 r{circumflex over (σ)} Z 2 where {circumflex over (σ)} I 2 is the nominal interference variance, r is a threshold value and {circumflex over (σ)} Z 2 is the noise variance.
5. The method according to claim 1 wherein the optimal state sequence is found as
S
=
arg
max
S
p
(
X
S
,
P
,
T
,
I
)
p
(
S
P
,
T
,
I
)
.
6. The method according claim 1 wherein the optimal state sequence is produced using the Viterbi algorithm.
7. The method according to claim 2 wherein the sub-vector is found as
{right arrow over ({circumflex over (X)} Z [{right arrow over (X)} 1 . . . {right arrow over (X)} N Z ] where {right arrow over ({circumflex over (X)} Z is the sub-vector and {right arrow over (X)} 1 . . . {right arrow over (X)} N Z are elements of the sub-vector.
8. The method according to claim 2 or claim 7 wherein the sub-vector is found as the vector that minimises the distance from Gaussianity, such that
{circumflex over (N)} z =argmin N Z D ( {right arrow over (X)} N Z )
where D(•) is some measure of Gaussianity.
9. The method according to claim 8 wherein the measure of Gaussianity is kurtosis.
10. The method according to claim 8 wherein the measure of Gaussianity comprises a Kolmogorov-Smirnov distance.