Method for allocation of power in multiuser orthogonal frequency division multiplexing
View Patent ↗Power allocation methods for multiuser orthogonal frequency division multiplexing (OFDM) are described. Arbitrary partitioning of the OFDM tones among users and arbitrary modulation formats, possibly different for every user, are considered. Two systems of receivers and transmitters and corresponding methods address power allocation, respectively for slow fading channels tracked instantaneously by the system and for fast fading channels known only statistically thereby.
1. An apparatus for allocating transmission power among a plurality of users in an orthogonal frequency division multiplexing (OFDM) wireless communication system comprising:
a receiver for receiving transmissions and providing channel and noise estimates which fit a zero-mean complex Gaussian random process; and
a processor for receiving the channel and noise estimates, assigning tones and priorities among the plurality of users, and generating a channel state for each user, wherein the channel state is related to a complex gain value representing a fading state of a corresponding tone divided by a noise estimate for the corresponding tone, and allocating power to tones whose channel state is above a threshold, wherein the threshold is directly proportional to a bandwidth fraction assigned to a corresponding user and inversely proportional to the priority assigned to the corresponding user.
2. The apparatus of claim 1 further comprises:
a transmitter for transmitting signals to each user at a power level based on the power allocation, wherein the power allocation maximizes the region of simultaneously achievable mutual informations.
3. The apparatus of claim 1 wherein no power is allocated to tones whose channel state is below or equal to the threshold.
4. The apparatus of claim 1 wherein the threshold for the channel state for user j and parameter η is β j /w j η, wherein β j is the bandwidth fraction and w j is the priority of user j.
5. The apparatus of claim 4 wherein the parameter η is obtained by solving
∑
i
=
1
n
j
∑
j
=
1
J
1
n
γ
i
,
j
MMSE
j
-
1
(
min
{
1
,
β
j
η
w
j
γ
i
,
j
}
)
=
P
wherein for user j MMSE j −1 is the inverse minimum mean square error, γ ij is the channel state of tone i of user j, n is the total number of tones, and nP is the maximum power that can be supplied by a transmitter.
6. The apparatus of claim 5 wherein the power allocated to a tone i for a user j, p i,j , is given by
p
i
,
j
=
1
γ
i
,
j
MMSE
j
-
1
(
β
j
η
w
j
γ
i
,
j
)
γ
i
,
j
>
threshold
.
7. The apparatus of claim 4 wherein the channel state is an average channel state γ j =E[γ i,j (k)] i=1, . . . , n j .
8. The apparatus of claim 4 wherein the parameter η and the power allocated to every tone i for a user j, p j , is obtained by solving
E
[
γ
i
,
j
(
k
)
MMSE
j
(
p
_
j
γ
i
,
j
(
k
)
)
]
=
β
j
w
j
η
.
9. A method for allocating transmission power among a plurality of users in an orthogonal frequency division multiplexing (OFDM) wireless communication system comprising:
receiving channel and noise estimates which fit a zero-mean complex Gaussian random process;
generating a channel state for each user, wherein the channel state is related to a complex gain value representing a fading state of a corresponding tone divided by a noise estimate for the corresponding tone; and
allocating power to tones whose channel state is above a threshold, wherein the threshold is directly proportional to a bandwidth fraction assigned to a corresponding user and inversely proportional to a priority assigned to the corresponding user.
10. The method of claim 9 further comprises:
transmitting signals to each user at a power level based on the power allocation, wherein the power allocation maximizes the region of simultaneously achievable mutual informations.
11. The method of claim 9 further comprises:
allocating zero power to tones whose channel state is below or equal to the threshold.
12. The method of claim 9 wherein the threshold for the channel state for user j and parameter η is β j /w j η, wherein β j is the bandwidth fraction and w j is the priority of user j.
13. The method of claim 12 further comprises:
solving
∑
i
=
1
n
j
∑
j
=
1
J
1
n
γ
i
,
j
MMSE
j
-
1
(
min
{
1
,
β
j
η
w
j
γ
i
,
j
}
)
=
P
to obtain parameter η, wherein for user j MMSE j −1 is the inverse minimum mean square error, γ ij is the channel state of tone i of user j, n is the number of tones, and nP is the maximum power that can be supplied by a transmitter.
14. The method of claim 13 wherein allocating power further comprises:
allocating power p i,j to a tone i for a user j based on
∑
i
=
1
n
j
∑
j
=
1
J
1
n
γ
i
,
j
MMSE
j
-
1
(
min
{
1
,
β
j
η
w
j
γ
i
,
j
}
)
=
P
15. The apparatus of claim 12 wherein the channel state is an average channel state γ j =E[γ i,j (k)] i=1, . . . , n j wherein γ i,j (k) is the channel state for tone i user j.
16. The method of claim 12 wherein allocating power further comprises:
allocating power p j allocated to every tone i of user j based on E[γ i,j (k)MMSE j ( p j γ i,j (k))]=β j /w j η when the average channel state exceeds the threshold.
17. A non-transitory electronic medium storing a computer program, the computer program comprising instructions for allocating transmission power among a plurality of users in an orthogonal frequency division multiplexing (OFDM) wireless communication system, the instructions when executed causing a processor to perform steps comprising:
receiving channel and noise estimates which fit a zero-mean complex Gaussian random process;
generating a channel state for each user, wherein the channel state is related to a complex gain value representing a fading state of a corresponding tone divided by a noise estimate for the corresponding tone; and
allocating power to tones whose channel state is above a threshold, wherein the threshold is directly proportional to a bandwidth fraction assigned to a corresponding user and inversely proportional to a priority assigned to the corresponding user.
18. The non-transitory electronic medium of claim 17 further storing instructions when executed causing a processor to perform a step comprising:
transmitting signals to each user at a power level based on the power allocation, wherein the power allocation maximize the region of simultaneously achievable mutual informations.
19. The non-transitory electronic medium of claim 17 further storing instructions when executed causing a processor to perform steps comprising:
solving
∑
i
=
1
n
j
∑
j
=
1
J
1
n
γ
i
,
j
MMSE
j
-
1
(
min
{
1
,
β
j
η
w
j
γ
i
,
j
}
)
=
P
to obtain parameter η, wherein, for user j, MMSE j −1 is the inverse minimum mean square error, γ i,j is the channel state of tone i of user j, n is the total number of tones, and nP is the maximum power that can be supplied by a transmitter and wherein the threshold for the channel state for user j and parameter η is
β
j
w
j
η
,
wherein β j is the bandwidth fraction and w j is the priority of user j; and
allocating power p i,j to a tone i for a user j based on
E
[
γ
i
,
j
(
k
)
MMSE
j
(
p
_
j
γ
i
,
j
(
k
)
)
]
=
β
j
w
j
η
.
20. The non-transitory electronic medium of claim 17 further storing instructions when executed causing a processor to perform a step comprising:
allocating power p j allocated to every tone i for user j based on
E
[
γ
i
,
j
(
k
)
MMSE
j
(
p
_
j
γ
i
,
j
(
k
)
)
]
=
β
j
w
j
η
when an average channel state exceeds the threshold wherein the channel state is an average channel state γ j =E[γ i,j (k)] i=1, . . . , n j and γ i,j (k) is the channel state for tone i of user j.