Litcius/Paper detail

12.1 A 148nW General-Purpose Event-Driven Intelligent Wake-Up Chip for AIoT Devices Using Asynchronous Spike-Based Feature Extractor and Convolutional Neural Network

Zhixuan Wang, Le Ye, Ying Liu, Peng Zhou, Zhichao Tan, Haitao Fan, Yihan Zhang, Jiayoon Ru, Yangyuan Wang, Ru Huang

202133 citationsDOI

Abstract

Power is a major bottleneck in AIoT devices, which usually operate in random-sparse-event (RSE) scenarios [1] (Fig. 12.1.1, bottom). To process RSEs energy-efficiently, one can integrate an always-on wake-up chip [1-5] that only turns on the power-hungry high-performance system (HPS) once valid events are detected. Previous dedicated voice-activity-detection (VAD) wake-up chips using analog front-ends (AFE) followed by neural-network (NN) circuits consume 1μW [2] and 142nW [3]. Keyword spotting (KWS) wake-up, which is more complicated than VAD, can be achieved by an NN engine at 510nW [4], but it needs an off-chip power-hungry 16b ADC to drive its digital input. For general purpose, an event-driven wake-up chip [1] consumes 57nW. But without an intelligent inference engine (IIE), false detection due to invalid events will consume excessive power. The IIE can be realized by a finite-state-machine (FSM) pattern recognition circuit but with massive 2.2μW [5]. This work, to our best knowledge, presents a first reported event-driven intelligent wake-up chip that achieves ultra-low power and intelligent event detection for various AIoT RSEs.

Topics & Concepts

Computer scienceChipEvent (particle physics)Convolutional neural networkBottleneckAsynchronous communicationFeature (linguistics)Power (physics)WakeSystem on a chipReal-time computingComputer hardwareArtificial intelligenceEmbedded systemEngineeringTelecommunicationsAerospace engineeringLinguisticsPhilosophyPhysicsQuantum mechanicsAdvanced Memory and Neural ComputingFerroelectric and Negative Capacitance DevicesSemiconductor materials and devices