An official website of the United States government
With the expansion of sensor nodes to newer avenues of technologies, such as the Internet of things (IoT), internet of bodies (IoB), augmented reality (AR), and mixed reality, the demand to support high-speed operations, such as audio and video, with a minimal increase in power consumption is gaining much traction. In this work, we focus on these nodes operating in audio-based AR (AAR) and explore the opportunity of supporting audio at a low power budget. For sensor nodes, communicating one bit of data usually consumes significantly higher power than the power associated with sensing and processing/computing one data bit. Compressing the number of communication bits at the expense of a few computation cycles considerably reduces the overall power consumption of the nodes. Audio codecs such as AAC and LDAC that currently perform compression and decompression of audio streams burn significant power and create a floor to the minimum power possible in these applications. Compressive sensing (CS), a powerful mathematical tool for compression, is often used in physiological signal sensing, such as EEG and ECG, and it can offer a promising low-power alternative to audio codecs. We introduce a new paradigm of using the CS-based approach to realize audio compression that can function as a new independent technique or augment the existing codecs for a higher level of compression. This work, CS-Audio, fabricated in TSMC 65-nm CMOS technology, presents the first CS-based compression, equipped with an ON-chip DWT sparsifier for non-sparse audio signals. The CS design, realized in a pipelined architecture, achieves high data rates and enables a wake-up implementation to bypass computation for insignificant input samples, reducing the power consumption of the hardware. The measurement results demonstrate a 3X-15X reduction in transmitted audio data without a perceivable degradation of audio quality, as indicated by the perceptual evaluation of audio quality mean opinion score (PEAQ MOS) >1.5. The hardware consumes 238 μW power at 0.65 V and 15 Mbps, which is (~20X-40X) lower than audio codecs.
more »
« less
A 16 pJ/bit 0.1-15Mbps Compressive Sensing IC with on-chip DWT Sparsifier for Audio Signals
The emergence of Audio-based Augmented Reality has been calling for increasing data-rates for audio signals, with significant reduction in power to enable extremely energy-constrained sensor nodes. Typically, the communication power dominates sensing and computing power in a node [1]. For highly energy constrained scenarios, compressive sensing (CS) have been demonstrated (Fig. 1), where samples are first compressed at the sensor to contain the same information in a smaller number of samples, before transmitting to a receiver, where the signal is reconstructed. Previous CS works [2]-[5] have focused entirely on “sparse” physiological signals, operating in low speed regime. This work illustrates the first CS design, enabled with a discrete wavelet transform (DWT) sparsifier for catering to non-sparse signals such as high definition audio. Audio recording and playback are quite sensitive to quality, thereby requiring audio codecs, such as. aac, for efficient compression and decompression of audio streams, which usually consume power in the order of mW [6]. Audio inferencing operated in intelligent assistants are more tolerant to input quality, functioning effectively when the Perceptual Evaluation of Audio Quality Mean Opinion Score (PAEQ MOS) [7], an ITU-R standard objective metric for characterizing perceived audio quality, exceeds 1.5. CS presents an opportunity to achieve >10X reduction in transmitted audio data with orders of magnitude lower power, as compared to codecs. The design is implemented in 65 nm CMOS and consumes 238 uW power at 0.65 V and 15 Mbps.
more »
« less
- Award ID(s):
- 1944602
- PAR ID:
- 10320902
- Date Published:
- Journal Name:
- 2021 IEEE Custom Integrated Circuits Conference (CICC)
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Recent advances to hardware integration and realization of highly-efficient Compressive Sensing (CS) approaches have inspired novel circuit and architectural-level approaches. These embrace the challenge to design more optimal nonuniform CS solutions that consider device-level constraints for IoT applications wherein lifetime energy, device area, and manufacturing costs are highly-constrained, but meanwhile the sensing environment is rapidly changing. In this manuscript, we develop a novel adaptive hardware-based approach for non-uniform compressive sampling of sparse and time-varying signals. The proposed Adaptive Sampling of Sparse IoT signals via STochastic-oscillators (ASSIST) approach intelligently generates the CS measurement matrix by distributing the sensing energy among coefficients by considering the signal characteristics such as sparsity rate and noise level obtained in the previous time step. In our proposed approach, Magnetic Random Access Memory (MRAM)-based stochastic oscillators are utilized to generate the random bitstreams used in the CS measurement matrix. SPICE and MATLAB circuit-algorithm simulation results indicate that ASSIST efficiently achieves the desired non-uniform recovery of the original signals with varying sparsity rates and noise levels.more » « less
-
Feature extraction, such as spectral occupancy, interferer energy and type, or direction-of-arrival, from wideband radio-frequency (RF) signals finds use in a growing number of applications as it enhances RF transceivers with cognitive abilities and enables parameter tuning of traditional RF chains. In power and cost limited applications, e.g., for sensor nodes in the Internet of Things, wideband RF feature extraction with conventional, Nyquist-rate analog-to-digital converters is infeasible. However, the structure of many RF features (such as signal sparsity) enables the use of compressive sensing (CS) techniques that acquire such signals at sub-Nyquist rates; while such CS-based analog-to-information (A2I) converters have the potential to enable low-cost and energy-efficient wideband RF sensing, they suffer from a variety of real-world limitations, such as noise folding, low sensitivity, aliasing, and limited flexibility. This paper proposes a novel CS-based A2I architecture called non-uniform wavelet sampling. Our solution extracts a carefully-selected subset of wavelet coefficients directly in the RF domain, which mitigates the main issues of existing A2I converter architectures. For multi-band RF signals, we propose a specialized variant called non-uniform wavelet bandpass sampling (NUWBS), which further improves sensitivity and reduces hardware complexity by leveraging the multi-band signal structure. We use simulations to demonstrate that NUWBS approaches the theoretical performance limits of ℓ₁-norm-based sparse signal recovery. We investigate hardware-design aspects and show ASIC measurement results for the wavelet generation stage, which highlight the efficacy of NUWBS for a broad range of RF feature extraction tasks in cost- and power-limited applications.more » « less
-
null (Ed.)Spiking neural networks (SNNs) have emerged as a new generation of neural networks, presenting a brain-inspired event-driven model with advantages in spatiotemporal information processing. Due to the need for high power consumption of compute-intensive neural accelerators, adequate power delivery network (PDN) design is a key requirement to ensure power efficiency and integrity. However, PDN design for SNN accelerators has not been extensively studied despite its great potential benefit in energy efficiency. In this paper, we present the first study on dynamic heterogeneous voltage regulation (HVR) for spiking neural accelerators to maximize system energy efficiency while ensuring power integrity. We propose a novel sparse-workload-aware dynamic PDN control policy, which enables high energy efficiency of sparse spiking computation on a systolic array. By exploring sparse inputs and all-or-none nature of spiking computations for PDN control, we explore different types of PDNs to accelerate spiking convolutional neural networks (S-CNNs) trained with the dynamic vision sensor (DVS) gesture dataset. Furthermore, we demonstrate various power gating schemes to further optimize the proposed PDN architecture, which leads to a more than a three-fold reduction in total energy overhead for spiking neural computations on systolic array-based accelerators.more » « less
-
Neural codecs have demonstrated strong performance in high-fidelity compression of audio signals at low bitrates. The token-based representations produced by these codecs have proven particularly useful for generative modeling. While much research has focused on improvements in compression ratio and perceptual transparency, recent works have largely overlooked another desirable codec property -- \textit{idempotence}, the stability of compressed outputs under multiple rounds of encoding. We find that state-of-the-art neural codecs exhibit varied degrees of idempotence, with some degrading audio outputs significantly after as few as three encodings. We investigate possible causes of low idempotence and devise a method for improving idempotence through fine-tuning a codec model. We then examine the effect of idempotence on a simple conditional generative modeling task, and find that increased idempotence can be achieved without negatively impacting downstream modeling performance -- potentially extending the usefulness of neural codecs for practical file compression and iterative generative modeling workflows.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/CICC51472.2021.9431569
