🤖 AI Summary
This study addresses the energy bottleneck imposed by data movement in edge AI and presents the first evidence-graded review of the synergistic potential between event-driven vision sensors (DVS) and in-memory analog computing, particularly memristor-based systems—a combination that has not been systematically evaluated to date. The work introduces a three-paradigm architectural taxonomy and integrates technology readiness level (TRL) assessments with benchmark comparisons against digital neuromorphic approaches. Findings reveal that half of the six key application domains rely solely on theoretical projections, while existing hardware implementations predominantly reside at TRL 2–5. By delineating the gap between prototypes and deployable systems, the study establishes verifiable targets for accuracy and power efficiency, thereby filling a critical void in the research roadmap for DVS–memristor integration.
📝 Abstract
Edge-AI deployment is bottlenecked by data-movement energy; pairing event-driven vision sensors with in-memory analog compute could lift that ceiling by orders of magnitude. Both technologies are individually mature; the framework distinguishing fabricated demonstrations from projected systems is missing. Of six application domains surveyed (robotics, autonomous vehicles, AR/VR, surveillance, medical imaging, IoT), half rest entirely on projection, and existing hardware sits at Technology Readiness Levels 2-5. This evidence-graded review applies a three-paradigm architectural taxonomy and benchmarks the gap against current digital neuromorphic alternatives. It identifies an end-to-end integrated DVS-memristor system as the field's open challenge, with testable accuracy and power targets.