Skip to main content
Log in

Duality Between the Improved Rate Regions of Gaussian Multiple Access and Broadcast Channel for Discrete Constellations

  • Published:
Wireless Personal Communications Aims and scope Submit manuscript

Abstract

In this paper, we first look at the achievable rate regions of the Gaussian Multiple Access Channel (G-MAC) and Gaussian Broadcast Channel (G-BC) when discrete constellations are transmitted. Focusing on the two-user case and assuming uncoded Pulse Amplitude Modulation (PAM), we show that the rate regions of G-MAC and G-BC can be improved by power conservation at the transmitter, a factor ignored in the earlier derived rate regions. We then further investigate duality between the G-MAC and the G-BC and show that a rate pair achieved in the G-MAC can be translated to a rate pair in the dual G-BC, such that the equal sum power constraint be satisfied. Due to the similarity of the rate expressions to Shannon’s capacity formula, for an appropriate choice of signal to noise ratio (SNR) gap, we show that when finite constellations are used for transmission, rate regions of these two channels also have a dual relationship (the known for Gaussian alphabets). The rate region of the G-BC can therefore be characterized from the rate region of the dual G-MAC and vice versa.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9

Similar content being viewed by others

References

  1. Shannon, C. E. (1948). A mathematical theory of communications. Bell Systems Technical Journal, 27, 379–423. (part I) and 623–656 (part II).

    Article  MATH  MathSciNet  Google Scholar 

  2. Cioffi, J. M. (2011). EE379. Digital communication. Lecture notes. Stanford, CA: Stanford University.

    Google Scholar 

  3. Toumpakaris, D., & Lee, J. (2009). The gap approximation for Gaussian multiple access channels. In Global Telecommunications Conference, 2009. GLOBECOM 2009. IEEE, Nov 30 2009–Dec 4 2009 (pp. 1–6). IEEE.

  4. Toumpakaris, D., & Lee, J. (2010). On the use of the gap approximation for the Gaussian broadcast channel. In Global Telecommunications Conference, 2009. GLOBECOM 2009. IEEE, 6–10 Dec 2010 (pp. 1–5). IEEE.

  5. Ghaffar, R., Toumpakaris, D., & Lee, J. (2011). Achievable rates for transmission of discrete constellations over the Gaussian MAC channel. International Conference on ICT Convergence (ICTC), pp. 1–5.

  6. Jindal, N., Vishwanath, S., & Goldsmith, A. (2004). On the duality of Gaussian multiple-access and broadcast channels. IEEE Transactions on Information Theory, 50(5), 768–783.

    Article  MATH  MathSciNet  Google Scholar 

  7. Cover, T. M., & Thomas, J. A. (2006). Elements of information theory. New Jersy: Wiley.

    MATH  Google Scholar 

  8. Gamal, A. E., & Kim, Y.-H. (2012). Network information theory. Cambridge University Press: Cambridge.

  9. Rimoldi, B., & Urbanke, R. (1996). A rate-splitting approach to the Gaussian multiple-access channel. IEEE Transactions on Information Theory, 42(2), 364–375.

    Article  MATH  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Rizwan Ghaffar.

Appendix

Appendix

We can rewrite the rate equations in G-MAC and G-BC as

$$ {R_i}^{mac} = \frac{1}{2}log_2 \left (1+ \frac{h_i{P_i}^{mac}}{A_i} \right), {R_i}^{bc} = \frac{1}{2}log_2 \left(1+ \frac{h_i{P_i}^{bc}}{B_i}\right)$$
(38)

We show that if

$$ {{P_i}^{mac}}{B_i}={{P_i}^{bc}}{A_i}, i=1, ..., K $$
(39)

for all i, where

$$A_i={\tilde{\Gamma }_{i,mac}}\left( {\sum _{j=i+1}^{K}h_j{P_j} ^{mac}}+ N_{0}\right)$$
(40)

and

$$ B_i={\tilde{\Gamma }_{i,bc}}\left( h_i{\sum _{j=1}^{i-1}{P_j}^{bc}}+ N_{0}\right)$$
(41)

then \(\sum _{j=1}^{K}{P_j}^{mac}=\sum _{j=1}^{K}{P_j}^{bc}\). We do this by inductively showing that

$$\sum _{j=1}^{i}{P_j}^{bc}= \frac{{{\tilde{\Gamma }_{1,bc}} N_{0}}\sum _{j=1}^{i}{P_j}^{mac}}{A_i}, $$
(42)

For i = 1,

$${{P_1}^{bc}}=\frac{{P_1}^{mac}{B_1}}{A_1}, $$
(43)

From (41)

$$ B_1={\tilde{\Gamma }_{1,bc}} N_{0} $$
(44)

So (43) becomes

$${{P_1}^{bc}}=\frac{{{\tilde{\Gamma }_{1,bc}} N_{0}}{P_1}^{mac}}{A_1}, $$
(45)

We can generalize (45) for i users as shown in (46).

$$ \sum _{j=1}^{i}{P_j}^{bc}= \frac{{{\tilde{\Gamma }_{1,bc}} N_{0}}\sum _{j=1}^{i}{P_j}^{mac}}{A_i}, $$
(46)

Assume that (46) holds for i users so for i + 1 case we have,

$$\begin{aligned} \sum _{j=1}^{i+1}{P_j}^{bc}= \sum _{j=1}^{i}{P_j}^{bc} + \frac{{B_{i+1}}{P_{i+1}}^{mac}}{A_{i+1}} \end{aligned}$$
(47)

From (41), we can get the equation for \(B_{i+1}\) as

$$\begin{aligned} B_{i+1}={\tilde{\Gamma }_{i+1,bc}}\left( h_{i+1}{\sum _{j=1}^{i} {P_j}^{bc}}+ N_{0}\right) \end{aligned}$$
(48)

Substituting (48) in (47), we get

$$\begin{aligned} \sum _{j=1}^{i+1}{P_j}^{bc}&= \sum _{j=1}^{i}{P_j}^{bc} + \frac{{{\tilde{\Gamma }_{i+1,bc}}\left( h_{i+1}{\sum _{j=1}^{i} {P_j}^{bc}}+ N_{0}\right) }{P_{i+1}}^{mac}}{A_{i+1}} \end{aligned}$$
(49)
$$\begin{aligned} \sum _{j=1}^{i+1}{P_j}^{bc}&=\frac{{\tilde{\Gamma }_{i+1,bc}}N_{0}{P_{i+1}}^{mac}+ \left( P_{i+1} ^{mac}h_{i+1}\tilde{\Gamma }_{i+1,bc}+A_{i+1}\right) \sum _{j=1}^{i}{P_j}^{bc}}{A_{i+1}} \end{aligned}$$
(50)

For all \(i>1\), \(\tilde{\Gamma }_{i,bc}\) is equal to one in high SNR regime. Thus \(\tilde{\Gamma }_{i+1,bc}\) will also equal to one, for \(i\ge 1\).

We can write \( A_i=P_{i+1} ^{mac}h_{i+1}\tilde{\Gamma }_{i+1,mac}+A_{i+1}\) from (40). But we know that the additional gap in the G-MAC is equal to one, for all \(i<K\). Till when \(i<K\), \(\tilde{\Gamma }_{i+1,mac}\) will be equal to one. Thus \(A_i\) could be written as \( A_i=h_{i+1}P_{i+1} ^{mac}+A_{i+1}\) then (50) becomes

$$\begin{aligned} \sum _{j=1}^{i+1}{P_j}^{bc}=\frac{N_{0}{P_{i+1}}^{mac}+A_i \sum _{j=1}^{i}{P_j}^{bc}}{A_{i+1}} \end{aligned}$$
(51)

We can replace \({A_i} \sum _{j=1}^{i}{P_j}^{bc}\) from (42) as

$$\begin{aligned} {A_i} \sum _{j=1}^{i}{P_j}^{bc}= {\tilde{\Gamma }_{1,bc}}N_{0}\sum _{j=1}^{i}{P_j}^{mac}, \end{aligned}$$
(52)

Substituting (52) in (51) we get

$$\begin{aligned} \sum _{j=1}^{i+1}{P_j}^{bc}&=\frac{N_{0}{P_{i+1}}^{mac}+ {{\tilde{\Gamma }_{1,bc}} N_{0}}\sum _{j=1}^{i}{P_j}^{mac}}{A_{i+1}} \end{aligned}$$
(53)
$$\begin{aligned} \sum _{j=1}^{i+1}{P_j}^{bc}&=\frac{\tilde{\Gamma }_{1,bc} N_{0}\sum _{j=1}^{i+1}{P_j}^{mac}}{A_{i+1}} \end{aligned}$$
(54)
$$\begin{aligned} \sum _{j=1}^{i+1}{P_j}^{bc}&=\frac{\tilde{\Gamma }_{1,bc} N_{0}\sum _{j=1}^{i+1}{P_j}^{mac}}{A_{i+1}} \end{aligned}$$
(55)

By using (42) for \(i=K\)

$$\begin{aligned} \sum _{j=1}^{K}{P_j}^{bc}= \frac{{{\tilde{\Gamma }_{1,bc}} N_{0}}\sum _{j=1}^{K}{P_j}^{mac}}{A_K}, \end{aligned}$$
(56)

from (40)

$$\begin{aligned} A_K=\tilde{\Gamma }_{K,mac}N_{0} \end{aligned}$$
(57)

Substituting (57) into (56)

$$\begin{aligned} \sum _{j=1}^{K}{P_j}^{bc}= \frac{{{\tilde{\Gamma }_{1,bc}} N_{0}}\sum _{j=1}^{K}{P_j}^{mac}}{\tilde{\Gamma }_{K,mac}N_{0}}, \end{aligned}$$
(58)

The single user gap values for a given probability of error vary slightly with modulation scheme [3]. So, \(\tilde{\Gamma }_{1,bc}\sim \tilde{\Gamma }_{K,mac}\).

$$\begin{aligned} \sum _{j=1}^{K}{P_j}^{bc}= \sum _{j=1}^{K}{P_j}^{mac} \end{aligned}$$
(59)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Rehman, I.A., Ghaffar, R. & Qaisar, S. Duality Between the Improved Rate Regions of Gaussian Multiple Access and Broadcast Channel for Discrete Constellations. Wireless Pers Commun 84, 2397–2415 (2015). https://doi.org/10.1007/s11277-015-2711-9

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11277-015-2711-9

Keywords

Navigation