Interference-Aware Dynamic Spectrum Access in Cognitive Radio Network Rajarshi Mahapatra and Emilio Calvanese Strinati CEA, LETI, MINATEC, 17, rue des Martyrs - 38054 Grenoble, France Email:rajarshim@ieee.org, emilio.calvanese-strinati@cea.fr Abstract—Dynamic spectrum access (DSA) is a challenging issue among the cognitive radio (CR) users. In this paper, we investigate an interference-aware local cartography-based DSA (LC-DSA) technique for OFDM-based CR system. In particular, a joint power and frequency resource block (RB) allocation technique has been proposed for resource allocation among sec- ondary users (SUE) exploiting radio environmental map (REM) of the local surrounding environment. With REM, we produce an interference database for each RB at each SUE’s location. Having knowledge of these measured interference values, LC- DSA technique elaborates an interference cartography (IC) diagram of CR coverage area with the help of spatial inter- polation algorithm. Using IC diagram, CR applies interference classification algorithm to classify RBs at target location and uses them for transmission with appropriate transmit power without producing interference above the threshold value to the primary user (PUE). Our simulation results show how the proposed LC- DSA technique significantly improves the system throughput. Simultaneously, the interference introduced to the PUE remains within a tolerable limit. I. I NTRODUCTION NeXt Generation (xG) wireless communication network aims to minimize the inefficient usage of spectrum by dynam- ically access unused spectrum with the help of cognitive radio (CR) [1]. Dynamic spectrum access (DSA) is a promising approach to alleviate the spectrum scarcity and aims at reusing sparsely occupied frequency bands while causing no (or in- significant) interference to the actual user, i.e. primary user (PUE). The key characteristics of DSA system empower their ability to exploit knowledge of the surrounding environment to adapt their transmission parameters and access the spectrum dynamically. The CR [2], built on software defined radio, is an intelligent, adaptive and frequency agile wireless device that will underlie most form of future DSA system. In CR networks, unused spectrum bands were spread over wide frequency range including both unlicensed and licensed bands. These unused spectrum bands have different charac- teristics, in terms of operating frequency and bandwidth, and these bands are detected through spectrum sensing [1]. Due to unplanned deployment, CR users (SUE) can suffer interference from primary base stations (PBS). On the contrary, CR can also produce interference at the nearby PUE. Therefore, it is absolutely necessary for CR to reduce interference through appropriate DSA technique [3]. Several techniques are available in literature for efficient DSA mechanism of CR network. In [1], authors have consid- ered a centralized control entity, named spectrum broker for xG network. In [4], the authors propose opportunistic access by the cognitive node to increase flexibility with enhancing spectrum efficiency. The interference due to the secondary activity does not exceed a predefined threshold at primary terminal. The interference can be estimated by transmitter detection-based spectrum sensing. Due to its great flexibility in dynamically allocating unused spectrum among the users, [5], orthogonal frequency division multiplexing (OFDM) has already been recognized in the literature as a potential trans- mission technology for the secondary systems. Thus, OFDM- based sensing can give more flexibility and allow accurate allocation of the available resource blocks (RB) to secondary users [6]. In [7], authors introduce a resource allocation al- gorithm based on location information for OFDMA-based CR systems. In [8], a resource allocation among several SUEs in an OFDM-based network has been proposed with the intention of increasing the system throughput while limiting interference to the PUEs. Motivated by the above works, this paper proposes a radio resource allocation technique for among CR users exploiting radio environment maps (REMs). REM is envisioned as an integrated space-time-frequency database consisting of multi- domain information, such as geographical features, available services, spectral regulations, locations and activities of radios, relevant policies, and experiences [9]. In this context, this paper proposes an interference-aware local cartography-based DSA (LC-DSA) technique for CR system, which consists of one CR-transmitter (CR-Tx) and several SUEs. The CR system uses OFDMA-based access technique to measures interference values at user location for each RB by sensing. In particular, a joint power and frequency RB allocation scheme has been pro- posed for CR using LC-DSA techniques to maximize the SUE capacity. In order to achieve this, we introduce the concept of interference cartography (IC) for better resource allocation. The interference cartography (discussed in Sec. III) combines radio measurement data with user’s location information and provides a complete view of the environment for autonomous decision making [10]. In this work, we use spatial interpolation algorithm, called Kriging interpolation to estimate interference values at the unobserved location and makes IC diagram for each RB of CR coverage area. Thereafter, these RBs have been classified based on their interference values using classification technique at desired location. Then, appropriate transmit power have been used on these classified RBs for transmission to maximize downlink transmission capacity of SUEs, while the 22nd IEEE Personal Indoor Mobile Radio Communications 978-1-4577-1348-4/11/$26.00 ©2011 IEEE 398