MUSIC-lite: Efficient MUSIC using Approximate Computing: An OFDM Radar Case Study

TL;DR

MUSIC-lite introduces an approximate computing approach to optimize the computationally intensive MUSIC algorithm, specifically targeting SVD in OFDM radar, achieving significant area and power savings with minimal accuracy loss.

Contribution

This work presents MUSIC-lite, a novel approximate computing framework that reduces hardware complexity and power consumption of MUSIC's SVD component in OFDM radar applications.

Findings

Average 17.25% area reduction

Average 19.4% power savings

Minimal 0.14% accuracy error

Abstract

Multiple Signal Classification (MUSIC) is a widely used Direction of Arrival (DoA)/Angle of Arrival (AoA) estimation algorithm applied to various application domains such as autonomous driving, medical imaging, and astronomy. However, MUSIC is computationally expensive and challenging to implement in low-power hardware, requiring exploration of trade-offs between accuracy, cost, and power. We present MUSIC-lite, which exploits approximate computing to generate a design space exploring accuracy-area-power trade-offs. This is specifically applied to the computationally intensive singular value decomposition (SVD) component of the MUSIC algorithm in an orthogonal frequency-division multiplexing (OFDM) radar use case. MUSIC-lite incorporates approximate adders into the iterative CORDIC algorithm that is used for hardware implementation of MUSIC, generating interesting accuracy-area-power trade-offs. Our experiments demonstrate MUSIC-lite's ability to save an average of 17.25% on-chip area and 19.4% power with a minimal 0.14% error for efficient MUSIC implementations.