Distributed transmit power allocation for multihop cognitive-radio systems

In this paper, we consider a relay-assisted wideband cognitive-radio (CR) system under the assumption that the frequency band chosen by the CR relay network for unlicensed spectrum usage overlaps with one or more bands dedicated to primary (e.g., licensed) narrowband links. Our objective is to optimize the performance of the CR system while limiting the interference in direction of the primary receivers, without requiring any adaptation of the transmitted signal spectra at the cognitive nodes. To this end, we study appropriate transmit power allocation (TPA) strategies among the cognitive relays. We first investigate the optimal centralized (OC) TPA solution and show that it can be formulated as a linear program. Since the OC-TPA solution requires a considerable amount of information exchange between the cognitive nodes, we develop two distributed TPA schemes, namely (i) a fully decentralized (FD) TPA scheme and (ii) a distributed feedback-assisted (DFA) TPA scheme. The FD-TPA scheme aims at maximizing the output signal-to-interference- plus-noise ratio (SINR) at the destination node of the CR network according to a best-effort strategy. It requires neither feedback information from the destination node nor an exchange of channel state information between the cognitive relays. The DFA-TPA scheme, on the other hand, utilizes feedback information from the destination node, in order to achieve a predefined target output SINR value, while minimizing the overall transmit power spent by the relays. Analytical and simulation-based performance results illustrate that notable performance improvements compared to non-cooperative transmission (i.e., without relay assistance) are achieved by the proposed schemes, especially when more than two hops are considered. In particular, the proposed distributed TPA schemes typically perform close to the OC-TPA solution.

I. Introduction

The Concept of cognitive radio (CR) has recently attracted considerable interest in the wireless communications community [1]–[3]. Traditionally, radio spectrum usage has been organized according to fixed frequency plans defined through government licenses. However, spectrum occupancy measurements have shown that within confined geographical areas significant amounts of licensed spectrum are typically underutilized [4]. As a central feature, CR systems are envisioned to take advantage of unused or only partially occupied bands in an adaptive, dynamic, and unlicensed (‘secondary’) fashion, thus allowing for a more efficient spectrum utilization [5]. To this end, CR systems will require spectrum-sensing capabilities [6], [7], based on which they adjust key transmission parameters such as frequency bands and radiated transmit power. For example, CR capabilities will be relevant for ultra-wideband (UWB) radio systems¹ [9], which have been approved by regulatory bodies around the world for unlicensed spectrum usage in (parts of) the 3.1–10.6 GHz band [10]. In this paper, we focus on wideband (or UWB) CR networks consisting of a possibly large number of low-power transceivers for short-range transmission (on the order of a couple of meters). Such a setup is relevant for wireless sensor networks (WSNs) employed for monitoring and control tasks, as well as for future personal area networks (PANs), e.g., for wireless exchange of multimedia content between laptops/personal computers and peripheric devices. In order to achieve connectivity and to guarantee a certain quality of service for such networks, relaying techniques appear to be an attractive choice. Available relays can either be dedicated cognitive relays, which do not disseminate any data of their own, or temporarily inactive cognitive devices that act as relays to assist the current source-destination link.