Semi-Blind Multiple-Frequency-Offset Estimation for Decode-and-Forward OFDM Cooperative Networks Yu-Ting Sun and Jia-Chin Lin Department of Communication Engineering, National Central University, Jhongda Road 300, Jhongli, Taoyuan 32001, Taiwan. Email: 965403001@cc.ncu.edu.tw, jiachin@ieee.org Tel: 886-3-422-7151 ext. 35514; Fax: 886-3-422-9187. Abstract—Cooperative networks recently became attractive because they can achieve spatial diversity. Multiple decode-and- forward (DF) relay nodes result in multiple carrier frequency offsets (CFOs) because each relay node owns its local oscillator. This paper presents a novel semi-blind multiple-CFO estimator for DF OFDM cooperative networks. A procedure is designed to effectively derive the semi-blind multiple-CFO estimator for relay nodes. Each relay node occupies its own subchannels, and its CFO falls in a nonoverlapped spectrum. The CFO mapping can simply match the corresponding relay nodes. In addition, the proposed method can reduce hardware and computational complexity. Semi-blind random sequences are derived from the characteristics of the signal matrix in the multiple signal charac- terization (MUSIC) algorithm. Comprehensive simulations show that random sequences can replace periodic training sequences. Furthermore, some information can be conveyed by random sequences to increase spectral efficiency. I. I NTRODUCTION Cooperative network systems using multiple relay nodes can achieve spatial diversity, increase capacity and enhance cov- erage [1] [2]. Cooperative networks create a virtual multiple antenna structure by combining multiple relay nodes. How- ever, the process of synchronization might be very intricate due to the numerous nodes in this type of cooperative networks. Most past research on cooperative networks assumes perfect synchronization to simplify the model. Some research papers [3] [4] verify that the performance is degraded singularly if the synchronization errors are large. The presence of multiple carrier frequency offsets (CFOs) in cooperative networks arises from multiple distributed relay nodes. Each relay node owns its local oscillator and brings mismatch between with the destination respectively. Additionally, the OFDM communi- cation is very sensitive to the CFO. The CFO could introduce intercarrier interference and thus dramatically degrade bit error rate performance. Therefore, handling the multiple frequency synchronization problem is a key issue for the successful deployment of a cooperative network. Conventional CFO estimation methods in OFDM systems can be categorized into two types: cyclic-prefix-based meth- ods [5] [6] and pilot-based methods [7]. These conventional CFO estimations only handle a single CFO. Multiple-CFO estimation has been studied in multiple-input-multiple-output (MIMO) systems [8]- [10]. A maximum-likelihood estimator is proposed in [8]. However, the computational complexity of the estimator is very large, and the accuracy is degraded if the CFO values are close to one another. A correlation-based estimator is proposed by using orthogonal training sequences with different antennas [9]. However, the correlation-based es- timator suffers from an error floor. To eliminate the correlation estimator error floor, two iterative algorithms are proposed in [10]. However, the performance of the correlation-based estimator is degraded when large CFOs occur. These methods only handle point-to-point MIMO systems in [9] [10] and thus could not apply to cooperative systems directly. The use of blind synchronization and channel estimation by N antennas at the destination in decode-and-forward (DF) cooperative communication systems is proposed in [11]. The blind method requires a more complex system structure. Non-blind methods are more popular than blind methods. In [12], iterative esti- mators are proposed with the multiple signal characterization (MUSIC) algorithm for DF and amplify-and-forward (AF) relay protocols. Additionally, the proposed estimators can also extract both large and small CFO values. However, iterative processes with the MUSIC algorithm are too onerous. In this paper, a semi-blind multiple-CFO estimator is pro- posed. It is assumed that the OFDM subcarriers are divided into a set of subchannels at relay nodes. The system is designed such that each relay node occupies one subchannel, and the subchannel cannot be used for another relay node. The design rule requires that each relay occupies a different subchannel and that each relay signal is orthogonal to the others. Thus, the range of effective CFOs with its corre- sponding relay node is derived and does not overlap. The estimated effective CFOs can easily match their corresponding relay by applying the MUSIC algorithm. By analyzing the characteristics of the designed system and the signal matrix in the MUSIC algorithm, a derivation can be carried out that replaces periodic training sequences in the MUSIC algorithm with linear independent random sequences. Furthermore, the transmitted signal from source terminal can carry several im- IEEE WCNC'14 Track 1 (PHY and Fundamentals) U.S. Government work not protected by U.S. copyright 463