Method and apparatus for managing interference in wireless backhaul networks through power control with a one-power-zone constraint
Methods and apparatus are provided for managing interference in a wireless backhaul network comprising a plurality of hubs, each hub serving a plurality of remote backhaul modules (RBM), using power control with a one-power-zone (OPZ) constraint. Each hub uses a transmit frame structure comprising a plurality of zones, each RBM is scheduled on a different zone, and the same power level is maintained across all zones within a transmit frame. Under the OPZ constraint, and for scheduling policies under which the number of zones assigned to each RBM is fixed, the power and scheduling sub-problems are decoupled. This enables power control independent of scheduling, using methods having lower computational complexity. Methods are disclosed comprising iterative function evaluation or Newton's method approaches based on a weighted sum-rate maximization across the network, which can be implemented in a distributed fashion. Some of the methods can be implemented asynchronously at each hub.
1. A method for managing interference in a wireless backhaul network comprising a plurality (L) of hubs, each serving a plurality (K) of remote backhaul modules (RBM), the method comprising:
for each hub, maintaining a one-power-zone (OPZ) constraint, comprising
for each transmitted frame comprising a plurality of zones, wherein the number of zones assigned to each RBMs is fixed,
scheduling each RBM on a different zone or a different set of zones,
maintaining the same power level across each of the plurality zones within a transmitted frame,
updating the power level for each zone for every hub frame based on weights λ lk and channel measurements h jlk , determining the appropriate power P l for every hub frame of the lth hub by steps comprising:
maximizing a weighted sum rate across all hub-RBM links, using an iterative function evaluation approach or a Newton's method approach, and wherein the optimization is defined by:
max
∑
l
,
n
λ
lk
r
lk
s
.
t
.
0
≤
P
l
≤
S
max
where
r
lk
=
log
(
1
+
P
l
h
llk
2
Γ
(
σ
2
+
∑
j
≠
l
P
j
h
jlk
2
)
)
where k=f(l,n) is the RBM scheduled at zone n of lth hub frame, Γ is the SNR gap, S max is the maximum power constraint, h jlk is the channel between the jth hub and the kth RBM of the lth hub, and where the optimization is over the set of powers P l .
2. A method according to of claim 1 wherein an objective function is selected to provide an optimization across a set of zones, or alternatively, across all zones for each frame.
3. A method according to 1 wherein the step of updating the power level comprises updating the transmit power for each hub downlink and/or uplink.
4. A method according to claim 1 wherein the step of updating is made after any one of: a change in channel; a change in weights λ ij ; and choosing the best solution among a set of optima resulting from different starting points.
5. A method according to claim 1 wherein the weights λ ij are set:
a) to be proportional to the rate constraints requirements of the system; or
b) to be proportional to the queue length of the data required for each RBM; or
c) to be inversely proportional to the long term average rate of each RBM.
6. A method according to claim 5 , wherein power allocation is based on channel gains for each hub-RBM link using the one-power-zone-constraint and an iterative function evaluation method (OPZ-IFEM).
7. The method of claim 6 (OPZ-IFEM), wherein the value of the power P l for each hub is updated from time t to t+1 is found iteratively using the following update function:
P
l
(
t
+
1
)
=
[
∑
K
λ
lk
(
SINR
l
k
(
t
)
1
+
SINR
l
k
(
t
)
)
∑
j
≠
l
∑
k
′
λ
jk
′
h
ljk
′
2
σ
2
+
∑
i
≠
j
P
i
(
t
)
h
ijk
′
2
SINR
j
k
′
(
t
)
1
+
SINR
j
k
′
(
t
)
]
0
S
max
where
SINR
l
k
=
P
l
h
llk
2
Γ
(
σ
2
+
∑
j
≠
l
P
j
h
jlk
2
)
.
8. A method of according to claim 5 , wherein power allocation is based on the channel gains, using the one-power-zone-constraint and a high SINR iterative function evaluation method (OPZ-HSIFEM) with a high SINR approximation.
9. The method of claim 8 (OPZ-HSIFEM), wherein the power of each hub P l is found by taking the limit as SINR goes to infinity, and wherein the value of the power P l is found iteratively using the following function:
P
l
(
t
+
1
)
=
[
∑
k
λ
lk
∑
j
≠
l
∑
k
′
λ
jk
′
h
ljk
′
2
σ
2
+
∑
i
≠
j
P
i
(
t
)
h
ijk
′
2
]
0
S
max
.
10. A method according to claim 5 , wherein power allocation is based on the channel gains, using the one-power-zone-constraint and a theta iterative function evaluation method (OPZ-theta-IFEM) using a fixed value of SINR.
11. The method of claim 10 (OPZ-theta-IFEM), wherein the power of each hub P l is found by replacing the SINR term by a fixed value of SINR evaluated at maximum allowable power, and wherein the value of the power P l is found iteratively using the following function:
P
l
(
t
+
1
)
=
[
∑
k
λ
lk
θ
l
k
∑
j
≠
l
∑
k
′
λ
jk
′
h
ljk
′
2
σ
2
+
∑
i
≠
j
P
i
(
t
)
h
ijk
′
2
θ
j
k
′
]
0
S
max
where
θ
l
k
=
SINR
~
l
k
1
+
SINR
~
l
k
and
SINR
~
l
k
=
S
max
h
llk
2
Γ
(
σ
2
+
∑
j
≠
l
S
max
h
jlk
2
)
.
12. A method according to claim 5 , wherein power allocation is based on the channel gains using the one-power-zone-constraint and an average iterative function evaluation method (OPZ-AVG-IFEM).
13. The method of claim 12 (OPZ-AVG-IFEM), wherein the power of each hub P l is found by taking one representative value of the power calculated on a per-zone basis P l n , wherein the value of the power P l n is found iteratively using the following iterative function:
P
l
n
(
t
+
1
)
=
[
λ
lk
SINR
l
k
(
t
)
1
+
SINR
l
k
(
t
)
∑
j
≠
l
λ
jk
′
h
ljk
′
2
σ
2
+
∑
i
≠
j
P
i
n
(
t
)
h
ijk
′
2
SINR
j
k
′
(
t
)
1
+
SINR
j
k
′
(
t
)
]
0
S
max
and wherein the value of the power P l is found using the average value of the resulting power P l n .
14. A method according to claim 5 , wherein power allocation is based on the channel gains using the one-power-zone-constraint and a Newton's method approach (OPZ-NM).
15. The method of claim 14 (OPZ-NM), wherein the power of each hub P l is found using an appropriate update of the power P l based on a Newton's method approach wherein the value of the power P l is found iteratively using the following function:
P l ( t +1)=[ P l ( t )+μΔ P l ( t )] 0 S max ,
where μ is the step size, set to one or chosen to guarantee the algorithm convergence, and where ΔP l (t) is found as the ratio of the corresponding gradient and hessian elements of the objective function:
Δ
P
l
(
t
)
=
∂
R
∂
P
l
(
t
)
∂
2
R
∂
P
l
2
(
t
)
,
where
∂
R
∂
P
l
=
∑
n
λ
lk
P
l
(
SINR
l
k
1
+
SINR
l
k
)
-
∑
j
≠
l
∑
n
λ
jk
′
h
ljk
′
2
σ
2
+
∑
i
≠
j
P
i
(
t
)
h
ijk
′
2
SINR
j
k
′
1
+
SINR
j
k
′
and
where
∂
2
R
∂
P
l
2
≈
∑
n
-
λ
lk
(
P
l
)
2
(
1
+
1
SINR
l
k
)
-
2
where k=f(l,n) and k′=f(j,n) are the RBMs served by hubs l and j in zone n, respectively.
16. The method of claim 1 , comprising any one of OPZ-IFEM, OPZ-HSIFEM, OPZ-theta-IFEM, OPZ-AVG-IFEM, and OPZ NM, wherein the steps of the method are implemented in a distributed manner across all hubs.
17. The method of claim 16 , comprising one of OPZ-HSIFEM and OPZ-theta-IFEM, wherein the steps of the method are implemented in an asynchronous manner at every hub.
18. Apparatus in a backhaul network comprising processor means for implementing the methods steps of claim 1 , comprising one of OPZ-IFEM, OPZ-HSIFEM, OPZ-theta-IFEM, and OPZ-AVG-IFEM, and OPZ-NM.
19. A computer program product embodied in a non-transitory transmission medium or non-transitory storage for performing, in processor means of a backhaul network, the method steps of claim 1 , comprising one of OPZ-IFEM, OPZ-HSIFEM, OPZ-theta-IFEM, and OPZ-AVG-IFEM, and OPZ-NM.
20. A system in a wireless backhaul network comprising processor means for implementing the method steps of claim 1 , comprising one of OPZ-IFEM, OPZ-HSIFEM, OPZ-theta-IFEM, and OPZ-AVG-IFEM, and OPZ-NM.