skip to main content

Title: Hardware Implementation of Single-carrier Time-domain Turbo Equalization for Severe Multipath Acoustic Communication Channels
This paper proposes a low-latency FPGA implemen-tation for Turbo equalization to combat very long multipathfading channels where the Intersymbol-interference (ISI) channellength is on the order of 100 taps. Turbo equalization is essentialfor such severe multipath channels, but exhibits very large latencyand high computational complexity due to its sequential anditerative data processing on large-scale matrix arithmetic. Thispaper proposes an FPGA acceleration architecture to exploitthe Hermitian symmetric property of the channel Gram matrixand convolutional nature of Sequential Interference Cancellation(SIC), and successfully implements a linear Turbo equalizerof 100 taps on a Xilinx Zynq UltraScale+ MPSoC ZCU102Evaluation Kit. The architecture is able to support two turboiterations for a 1024-symbol block size and achieve 200 kilo-symbols-per-second (ksps) transmission rate.
Authors:
;
Award ID(s):
1853258
Publication Date:
NSF-PAR ID:
10208375
Journal Name:
Ocean
ISSN:
0197-7385
Sponsoring Org:
National Science Foundation
More Like this
  1. This article investigates a robust receiver scheme for a single carrier, multiple-input–multiple-output (MIMO) underwater acoustic (UWA) communications, which uses the sparse Bayesian learning algorithm for iterative channel estimation embedded in Turbo equalization (TEQ). We derive a block-wise sparse Bayesian learning framework modeling the spatial correlation of the MIMO UWA channels, where a more robust expectation–maximization algorithm is proposed for updating the joint estimates of channel impulse response, residual noise, and channel covariance matrix. By exploiting the spatially correlated sparsity of MIMO UWA channels and the second-order a priori channel statistics from the training sequence, the proposed Bayesian channel estimator enjoysmore »not only relatively low complexity but also more stable control of the hyperparameters that determine the channel sparsity and recovery accuracy. Moreover, this article proposes a low complexity space-time soft decision feedback equalizer (ST-SDFE) with successive soft interference cancellation. Evaluated by the undersea 2008 Surface Processes and Acoustic Communications Experiment, the improved sparse Bayesian learning channel estimation algorithm outperforms the conventional Bayesian algorithms in terms of the robustness and complexity, while enjoying better estimation accuracy than the orthogonal matching pursuit and the improved proportionate normalized least mean squares algorithms. We have also verified that the proposed ST-SDFE TEQ significantly outperforms the low-complexity minimum mean square error TEQ in terms of the bit error rate and error propagation.« less
  2. Receiver nonlinearity gives rise to intermodulation products that are caused by two strong adjacent channel signals called blockers. The nonlinear distortion effects are significantly higher for multiple antenna wideband systems in dispersive environments because third order intermodulation products decreases the signal-to-noise ratio (SNR) at the output of the equalization process. This complicates the demodulation process and increases the bit error rate. This paper considers such nonlinear distortion in the context of space-time shift keying (STSK)-enabled wideband single-carrier systems and proposes an iterative space-time block equalization (ISTBE) framework for frequency domain equalization. We present our design of a practical ISTBE receivermore »based on the turbo principle and numerically demonstrate that it effectively removes the residual inter-symbol interference while suppressing high-power blockers and the in-band intermodulation distortion that they cause. The proposed system is thus suitable for simple wideband radio frequency front ends operating in the weak nonlinear region and enables adjacent channel spectrum coexistence with heterogeneous transmitters and receivers of different qualities.« less
  3. This paper proposes a post-experimental field data reuse method to test the single carrier modulation (SCM) and orthogonal frequency division multiplexing (OFDM) signals interchangeably for multiple access underwater acoustic (UWA) communications. We call this approach the cross evaluation that transforms a set of SCM or OFDM post-experimental field data to their corresponding OFDM or SCM scheme under test (SUT) via linear matrix operation such as fast Fourier transform (FFT) and its inverse (IFFT). At the receiver side, we derived a general framework of turbo equalization (TEQ) that alters the two physical layer schemes but keeps the passband transmitted and receivedmore »data unchanged. Inherently, some efficient techniques such as pre-cursor and post-cursor interference cancellation (IC), and overlap adding (OLA) operations enhance the equivalence of input and output (I/O) system model between the SCM and OFDM. The proposed approach will bring the gap between the SCM and OFDM, and evaluate the two physical layer schemes under similar or tougher test conditions. The experimental results of the undersea 2008 Surface Processes and Acoustic Communications Experiment (SPACE08) have verified the feasibility of the cross evaluation approach in terms of the BER benchmark.« less
  4. Large-scale deep neural networks (DNNs) are both compute and memory intensive. As the size of DNNs continues to grow, it is critical to improve the energy efficiency and performance while maintaining accuracy. For DNNs, the model size is an important factor affecting performance, scalability and energy efficiency. Weight pruning achieves good compression ratios but suffers from three drawbacks: 1) the irregular network structure after pruning, which affects performance and throughput; 2) the increased training complexity; and 3) the lack of rigirous guarantee of compression ratio and inference accuracy. To overcome these limitations, this paper proposes CirCNN, a principled approach tomore »represent weights and process neural networks using block-circulant matrices. CirCNN utilizes the Fast Fourier Transform (FFT)-based fast multiplication, simultaneously reducing the computational complexity (both in inference and training) from O(n2) to O(n log n) and the storage complexity from O(n2) to O(n), with negligible accuracy loss. Compared to other approaches, CirCNN is distinct due to its mathematical rigor: the DNNs based on CirCNN can converge to the same "effectiveness" as DNNs without compression. We propose the CirCNN architecture, a universal DNN inference engine that can be implemented in various hardware/software platforms with configurable network architecture (e.g., layer type, size, scales, etc.). In CirCNN architecture: 1) Due to the recursive property, FFT can be used as the key computing kernel, which ensures universal and small-footprint implementations. 2) The compressed but regular network structure avoids the pitfalls of the network pruning and facilitates high performance and throughput with highly pipelined and parallel design. To demonstrate the performance and energy efficiency, we test CirCNN in FPGA, ASIC and embedded processors. Our results show that CirCNN architecture achieves very high energy efficiency and performance with a small hardware footprint. Based on the FPGA implementation and ASIC synthesis results, CirCNN achieves 6 - 102X energy efficiency improvements compared with the best state-of-the-art results.« less
  5. Orthogonal signal-division multiplexing (OSDM) is one of the generalized modulation schemes that bring the gap between orthogonal frequency division multiplexing (OFDM) and single carrier frequency domain equalization (SC-FDE). By performing encoding upon subvectors of each interleaved block, it enjoys a flexible resource management with low peak-to-average power ratio (PAPR). Meanwhile, the OSDM induces the intervector interference (IVI) inherently, which requires a more powerful equalizer. By deriving the input and output system model, this paper proposes a time domain soft decision feedback equalizer (SDFE) on per vector equalization with successful soft interference cancellation (SSIC). In addition, this paper takes the wholemore »OSDM block to perform the channel encoding rather than on each vector of the OSDM. Simulation and experimental results demonstrate that the proposed SDFE with SSIC structure outperforms the conventional minimum mean square error (MMSE) equalizer and the block encoding (BE) scheme outperforms the vector encoding (VE) scheme, because theoretically the longer the encoded bit stream is, the more stable and more confident the maximum a posteriori probability (MAP) decoder will be.« less