Distributed power control in asymmetric interference-limited networks

Aakanksha Chowdhery

Abstract:Power control in wireless communication networks reduces i nterference by regulating access to the shared wireless med iu and enhances the overall network capacity as well as battery lif of nodes. Beginning with the Foschini-Miljanic (FM) al gorithm [1], distributed power control for wireless networks has be en an active area of research. The problem of distributed pow er control for packetized traffic was first addressed by Kanduku ri and Bambos [2] by proposing “Power-controlled multiple a ccess” (PCMA) family of algorithms, and subsequently investigate d in [3], [5], [6]. PCMA algorithm uses dynamic programming approach to balance the power versus delay cost of the transm itter. It introduces three separate regimes: aggressive, s oft backoff, hard backoff. In the low-interference regime, transmitter power tries to aggressively match and overcome the interfer ence – up to a certain point, while in the high interference regime, th e power is set to zero (hard backoff) and the transmitter ceas es to transmit waiting for the low interference periods. Thus, PC MA algorithms usually induce a soft time-division multiple access (iTDMA) allowing links facing high interference to tempora rily reduce their power and incur a delay cost instead. On the other hand, in Foschini-Miljanic (FM) algorithm, each link attempts to maintain a target SIR by overcoming the interfer nce presented by the other links. In this project, distributed power control algorithms are i nvestigated in a two-link interference-limited network wi th asymmetric interference. Such asymmetric interference can occ ur in any cellular as well as adhoc wireless networks, when th first link receives strong interference from the second link, and the second link receives weak interference from the first lin k. For example, a downlink user connected to a distant macro-cell c an be jammed due to the presence of a closer downlink user connected to a femtocell using the same frequency/time slot ( ee Fig. 1(a)). In adhoc wireless networks, the random plac ement of links may also result in asymmetric interference. Since d istributed power control algorithms control transmitter p ower based on the interference seen by the transmitter, they can p erform poorly in asymmetric interference-limited network s. If FM algorithm is used in asymmetric interference-limited netw orks, the weak link uses high power proportional to the inter fer nce from the strong link to maintain its target SIR, and its queue starts exploding at very low loads. PCMA algorithm performs better than FM algorithm in asymmetric interference-limit ed network, but it is unable to induce a iTDMA effect because t he strong link can get stuck in aggressive or soft backoff regim e when the average load is high and interference seen is low. Distributed power control algorithms have not been investi gated in asymmetric interference-limited networks to the b est of my knowledge. In this project, distributed power control al gorithms [1], [3] are simulated, and their performance is an alyzed for two-link asymmetric interference-limited network. Ba sed on the results, some variations are proposed to improve t he performance of PCMA algorithms, and these variations are st ill under investigation.