A scalable event-driven spatiotemporal feature extraction circuit

2025-01-17

AI Summary

Existing analog time-difference encoding (TDE) circuits for asynchronous sensors like event cameras suffer from process variation sensitivity, low integration density, and insufficient real-time performance when implemented on CMOS hardware. To address these challenges, this work proposes a scalable analog TDE circuit architecture that supports linear temporal integration of input events and exhibits robust tolerance to process variations, enabling high-density monolithic TDE array integration. Fabricated in standard CMOS technology, the circuit performs parallel spatiotemporal feature extraction on large-scale event streams with nanosecond-level latency. It achieves a two-order-of-magnitude improvement in event throughput and significantly outperforms conventional approaches in energy efficiency—reducing power consumption by over 90%. This work presents the first monolithic, fully analog TDE implementation capable of real-time, high-event-rate, low-latency, and ultra-low-power processing, establishing a scalable hardware foundation for edge-based neuromorphic vision systems.

Technology Category

• Machine Learning: Hardware-aware ML
• Computer Vision: Multi-modal Vision
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Security and Privacy: Large-scale security measurements
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments

Abstract

Event-driven sensors, which produce data only when there is a change in the input signal, are increasingly used in applications that require low-latency and low-power real-time sensing, such as robotics and edge devices. To fully achieve the latency and power advantages on offer however, similarly event-driven data processing methods are required. A promising solution is the TDE: an event-based processing element which encodes the time difference between events on different channels into an output event stream. In this work we introduce a novel TDE implementation on CMOS. The circuit is robust to device mismatch and allows the linear integration of input events. This is crucial for enabling a high-density implementation of many TDEs on the same die, and for realising real-time parallel processing of the high-event-rate data produced by event-driven sensors.

Problem

• Time Difference Encoder (TDE)
• efficiency
• stability

Innovation

• Time Difference Encoder (TDE)
• Robustness and Parallel Processing
• CMOS Integration