[Paper Review] Demodulating Subsampled Direct Sequence Spread Spectrum Signals using Compressive Signal Processing

Karsten Fyhn, Thomas Arildsen+2 more|arXiv (Cornell University)|Oct 24, 2011

Sparse and Compressive Sensing TechniquesEngineering12 references20 citations

TL;DR

This paper proposes using compressive signal processing to demodulate subsampled Direct Sequence Spread Spectrum (DSSS) signals, enabling a 50% reduction in sampling rate—demonstrated via IEEE 802.15.4 2.4 GHz OQPSK signals—without requiring complex hardware filters, thus lowering power consumption and ADC cost, though performance degrades by ~4–5 dB due to noise folding.

ABSTRACT

This poster is from EUSIPCO 2012 in Bucharest. We show that to lower the sampling rate in a spread spectrum communication system using Direct Sequence Spread Spectrum (DSSS), compressive signal processing (CSP) may be applied to demodulate the received signal. This may lead to a decrease in the power consumption or the manufacturing price of wireless receivers using spread spectrum technology. We propose a novel, simplified measurement scheme for spread spectrum systems, that is simpler and cheaper to implement than other current state-of-the-art aquisition methods. Our theoretical work is exemplified with a numerical experiment using the IEEE 802.15.4 standard’s 2.4 GHz band specification.

Motivation & Objective

• To reduce the sampling rate in DSSS receivers below the Nyquist rate using compressive signal processing.
• To simplify hardware implementation by avoiding complex filter structures typically required in compressive sensing.
• To demonstrate feasibility and performance trade-offs in a real-world standard (IEEE 802.15.4) with practical hardware comparisons.
• To evaluate the impact of noise folding on receiver performance in subsampled DSSS systems.
• To explore the potential for further sampling rate reduction in systems with longer chipping sequences and multi-user scenarios.

Proposed method

• Apply compressive signal processing to DSSS signals by exploiting inherent sparsity in the selection of spreading codes.
• Use a repeated matched filter structure instead of random measurement matrices, simplifying hardware implementation.
• Model the signal using baseband representations of QPSK-modulated DSSS signals with chip sequences from IEEE 802.15.4.
• Implement a compressive sensing-based receiver that performs inference (demodulation) directly from subsampled measurements.
• Use least squares estimation for signal recovery in the compressive framework, comparing performance to classical Nyquist-rate receivers.
• Conduct numerical experiments with AWGN in a 2 MHz bandwidth, simulating 1016-bit packets across varying Eb/N0 levels.

Experimental results

Research questions

• RQ1 Can compressive signal processing be effectively applied to DSSS signals to enable subsampling below the Nyquist rate?
• RQ2 How does the performance of a compressive sensing-based DSSS receiver compare to a classical Nyquist-rate receiver in terms of BER?
• RQ3 What is the impact of noise folding on the performance of subsampled DSSS receivers?
• RQ4 Can the hardware complexity of compressive sensing be reduced in DSSS systems by leveraging existing signal structure?
• RQ5 Under what conditions might compressive signal processing outperform classical receivers in terms of energy efficiency or cost?

Key findings

• A 50% reduction in sampling rate (κ = 0.5) is achievable in DSSS systems using compressive signal processing, demonstrated with IEEE 802.15.4 signals.
• The compressive sensing receiver achieves a BER performance that lags the classical receiver by approximately 4–5 dB due to noise folding.
• The proposed method avoids complex filter structures by reusing the matched filter, enabling simpler and more energy-efficient hardware implementation.
• The AD7819 ADC, operating at 200 kS/s, can replace the higher-cost AD7813 (400 kS/s) when using compressive signal processing, reducing cost from $2.98 to $2.29.
• The performance gap due to noise folding may be mitigated by higher-resolution quantization in compressive systems, favoring them in low-power applications.
• Longer chipping sequences and multi-user scenarios may allow for greater sampling rate reduction while maintaining or improving BER performance.