Elsevier

Applied Acoustics

Volume 117, Part A, February 2017, Pages 160-166
Applied Acoustics

Technical note
Distributed compressed sensing estimation of underwater acoustic OFDM channel

https://doi.org/10.1016/j.apacoust.2016.10.021Get rights and content

Abstract

Orthogonal frequency division multiplexing (OFDM) is recently drawing more and more attention for its high bandwidth efficiency over underwater acoustic (UWA) channels. However, the classic OFDM channel estimation algorithms, e.g. Least Square (LS), Minimum Mean Square Error (MMSE) are subject to significant performance degradation caused by doubly selective UWA channels. It has been recognized that the sparsity contained in UWA channels offers the possibility to improve the performance by compressed sensing (CS) estimation methods such as Orthogonal Matching Pursuit (OMP). Moreover, it has also been observed that multipath arrivals associated with adjacent OFDM symbols usually exhibit varying magnitude but similar delay, which means that UWA channels of several continuous symbols can be modeled as sparse sets with common support. In this paper, a Distributed Compressed Sensing (DCS) method is proposed to transform the problem of OFDM channel estimation into reconstruction of joint sparse signals. By exploiting this type of joint sparsity among adjacent OFDM symbols, we establish the DCS OFDM channel model, and then utilize the Simultaneous Orthogonal Matching Pursuit algorithm (SOMP) to optimize the model. Finally the experimental performance under field test is provided to illustrate the superiority of the proposed DCS channel estimation method, compared to the classic algorithm as well as CS counterparts.

Introduction

With rapidly increasing requirement on efficient acquisition and transmission of underwater observation information associated with ocean environmental monitoring, underwater project engineering and sea bottom resource exploitation tasks, there is an urgent need for R&D of high data rate underwater acoustic communication systems [1]. Unfortunately, compared with wireless channel, underwater acoustic channel is much more complicated due to the strictly limited bandwidth, extensive multi-path spread, Doppler shift and background noise. Design and implementation of high data rate underwater acoustic communication pose a considerable challenge [2] to the research community.

OFDM has recently emerged as a promising alternative to single-carrier systems for underwater acoustic communication because of its robustness to long delay spreads and frequency selectivity [3], [4], [5]. As well, OFDM provides high data rate for underwater acoustic communication which can be used for underwater acoustic speech or photo communication occasions. However, underwater OFDM systems are sensitive to Doppler shifting and phase noise [6], which will destroy the orthogonality of OFDM subcarriers, and pose difficulties in channel estimation and coherence detection [7].

Various channel estimation methods developed based on the assumption of the rich multi-path channel model, have been summarized in the literature [8]. Qiao et al. [9] pose an iterative Lease Square (LS) channel estimation algorithm for MIMO OFDM systems, the algorithm can greatly improve estimation accuracy, and the low-pass filtering in time domain reduces AWGN and ICI significantly. Jeong and Lee [10] pose a low complexity channel tracking for adaptive MMSE channel Estimation in OFDM system, the experimental result shows that the proposed channel parameter estimator tracks channel condition reliably in various channel conditions without significant increase in computational complexity. Morelli and Mengali [11] pose a maximum likelihood estimator (MLE) for OFDM system. However, the algorithms mentioned above require larger number of pilots or preambles to guarantee the estimated channel accuracy, which, unfortunately, will reduce the bandwidth efficiency and increase a high computation overhead. Meanwhile, there will exist significant estimated noise in non-zeros taps.

The underwater acoustic channels are considered to be sparse both in time and frequency domain, i.e., the delay-Doppler spread function has a limited number of nonzero elements [12]. The compressed sensing (CS) methods have been widely used for channel estimation to exploit the channel sparsity [13], [14], [15], [16]. Wu and Tong [13] propose the Non-uniform norm constraint LMS algorithm for sparse underwater acoustic channel estimation. Singh et al. [14] investigate the compressed sensing algorithms for estimation of OFDM channel. By adopting the compressed sensing methods, the length of training sequence can be significantly cut to improve efficiency, and the estimation noise at non-zero taps can be effectively suppressed [17].

While the UWA channels typically consists of sparse multipath arrivals, the multipath structure of adjacent OFDM symbols appears similar sparse structure because the delay of multipath arrivals tend to exhibit much slower variation than the corresponding magnitude does [18]. It means that UWA channels of several continuous symbols can be modeled as sparse sets with common support. Baron et al. [19] study three joint sparsity models (JSMs): JSM1 has sparse common component, JSM2 has common sparse supports and JSM3 has non-sparse common component. Underwater acoustic OFDM channels among adjacent symbols can be described as a JSM2 type model. Based on the basic concept of CS, Distributed compressed sensing (DCS) is proposed to exploit the joint sparseness contained among different sparse signals so as to achieve further performance enhancement. The temporal, spatial correlation among multiple sparse targets have been employed for DCS sparse recovery in wireless networks [20].

In this paper, a temporal joint sparse recovery approach is proposed to exploit the sparse correlation among adjacent OFDM symbols to improve the performance of OFDM channel estimation. By converting the problem of OFDM channel estimation into reconstruction of joint sparse signals with common support, a joint sparse model under the framework of DCS is adopted to derive a joint sparse recovery OFDM channel estimation algorithm. The performance of the proposed method is quantitatively evaluated with the experimental bit error rate (BER) of an UWA OFDM communication system. Finally, underwater OFDM communication experimental results obtained in Xiamen harbor are provided to demonstrate the effectiveness of the proposed method in improving the performance of underwater acoustic OFDM communication, compared to the classic channel estimation methods.

The rest of this paper is organized as follows. In Section 2, the model of OFDM and distributed compressed sensing are derived. The experimental performance analysis and comparison are provided in Section 3. Some conclusions are made in Section 4.

Section snippets

Transmitting model

We consider an OFDM baseband system with K equally spaced subcarriers at frequenciesfk=kΔf,k=0,K-1where Δf is the K subcarrier separation, therefore, the entire signal bandwidth is B=KΔf, and each OFDM symbol lasts T0=1/Δf. The K subcarriers can be allocated as either data symbols or pilot symbols, depending on the packet structure. Define the K subcarrier symbols of i-th OFDM symbol as bi:bi=[b(i,0),,b(i,K-1)]T

The transmitted time-domain discrete signal OFDM symbol can be expressed as:x=FHbi

Setup of at-sea experiment

The experimental field data was collected from a shallow water acoustic channel with slight wind condition at Wuyuan Bay, Xiamen, China. The depth of the experiment area is about 10 m. The OFDM signal was transmitted from a transducer, at a depth of 2 m. The transmitted signal was received by two receivers submerged at the depth of 2 m and 6 m respectively. The transmitting transducer was suspended under a pier and the receivers are mounted at the rear of one anchored ship, with a distance of 1000 m

Conclusion

While the sparsity contained in UWA channels have been popularly employed to improve the performance of channel estimation by CS methods, for UWA channels of adjacent symbols, similar delay pattern of multipath arrivals provides attractable possibility for further performance enhancement under the framework of DCS. In this paper, we investigated the joint estimation of underwater acoustic OFDM channels by modeling the channels of continuous OFDM systems as sparse set with common support. By

Acknowledgement

The authors are grateful for the funding of the National Nature Science Foundation of China (No. 11274259 and No. 11574258) in support of the present research.

References (26)

  • M. Stojanovic

    Efficient processing of acoustic signals for high-rate information transmission over sparse underwater channels

    Phys Commun

    (2008)
  • D. Rouseff et al.

    Effect of reflected and refracted signals on coherent underwater acoustic communication: results from the Kauai experiment (KauaiEx 2003)

    J Acoust Soc Am

    (2009)
  • M. Chitre et al.

    Recent advances in underwater acoustic communications & networking [C]//OCEANS 2008

    IEEE

    (2008)
  • B. Li et al.

    MIMO-OFDM for high rate underwater acoustic communications

    IEEE J Ocean Eng

    (2009)
  • B. Li et al.

    Multicarrier communications over underwater acoustic channels with nonuniform Doppler shifts

    IEEE J Ocean Eng

    (2008)
  • A. Radosevic et al.

    Adaptive OFDM modulation for underwater acoustic communications: design considerations and experimental results

    Ocean Eng IEEE J

    (2014)
  • Y. Zhang et al.

    MIMO-OFDM systems in the presence of phase noise and doubly selective fading

    Veh Technol IEEE Trans

    (2007)
  • Z. Wang et al.

    Clustered adaptation for estimation of time-varying underwater acoustic channels

    Signal Process IEEE Trans

    (2012)
  • M. Ozdemir et al.

    Channel estimation for wireless OFDM systems

    IEEE Commun Surv Tutorials

    (2007)
  • Y. Qiao et al.

    Research on an iterative algorithm of LS channel estimation in MIMO OFDM systems

    IEEE Trans Broadcast

    (2005)
  • K. Jeong et al.

    Low complexity channel tracking for adaptive MMSE channel estimation in OFDM [C]//Information Sciences and Systems, 2007

  • M. Morelli et al.

    A comparison of pilot-aided channel estimation methods for OFDM systems

    Signal Process IEEE Trans

    (2001)
  • W. Li et al.

    Estimation of rapidly time-varying sparse channels

    IEEE J Ocean Eng

    (2007)
  • Cited by (45)

    • Joint message-passing-based accurate bidirectional channel estimation and generalized-approximation-message-passing-based low-complexity perfect equalization with underwater moving transceivers

      2022, Applied Acoustics
      Citation Excerpt :

      There are three types of channel estimation algorithms: blind, semi-blind, and training-sequence-based (pilot-based) channel estimation algorithms. Compared with blind and semi-blind estimation algorithms, the training-sequence-based channel estimation algorithms have better channel estimation and tracking capabilities [4–15]. The training-sequence-based channel estimation algorithms are divided into two categories: the time division multiplexing training scheme and superimposed training (ST) scheme.

    • An effective framework for underwater acoustic data acquisition

      2021, Applied Acoustics
      Citation Excerpt :

      The collected data of AUV is transmitted to the platforms or the sub-AUVs [8]. On the other hand, orthogonal frequency division multiplexing (OFDM) is recently used for underwater acoustic channel estimation [9]. A distributed compressed sensing method is proposed to transform the problem of OFDM channel estimation into a reconstruction of joint sparse signals in [9].

    • Exploiting dynamic sparsity for time reversal underwater acoustic communication under rapidly time varying channels

      2021, Applied Acoustics
      Citation Excerpt :

      However, UWA channels pose significant challenges for high-rate and reliable UWA communications because of severe multipath spread, time variations and limited bandwidth [1–3]. Especially in shallow water environment, the rapidly time varying arrivals induced by highly dynamic ocean surfaces unfortunately lead to extremely serious difficulties to the estimation and equalization of UWA channel [4,5]. As a potential solution to mitigate these difficulties, it has been recognized by the research community that the sparse structure contained in UWA channels enable the exploitation of sparsity to improve the estimation performance [6,7].

    View all citing articles on Scopus
    View full text