An Analogmemoryless Near Sensor Computing Architecture For Alwayson Intelligent Perception Applications

• doi: 10.1109/ICTA48799.2019.9012906
• title: An Analog-Memoryless Near Sensor Computing Architecture for Always-On Intelligent Perception Applications
• publisher: IEEE
• isbn: 978-1-7281-5041-3
• issn:
• rank: 1473
• access_type: LOCKED
• content_type: Conferences
• abstract: In always-on intelligent visual perception applications, Convolution Neural Network algorithm has been widely applied. However, the implementation of it's processing is limited by hardware resources and energy budgets, which come from data access, interface overhead and inflexible configuration. In this paper, we present a novel analog architecture that implements low-precision Binary Neural Network and fuses with CMOS image sensor for always-on applications. By enabling current-mode computation in analog domain, the overhead of analog-to-digital converter as well as analog buffer is eliminated. And in order to relax the overhead of CIS, conventional correlating-double-sampling is moved to Computation Module. Moreover, the architecture is also dynamic-configurable, which is optimized specifically for CIS data flow. Simulations with HSPICE show that the peak energy efficiency of proposed architecture can achieve 13.91 TOPS/W at the 8-bit percision of multiplication-and-accumulation analog computation, while accuracy loss is only about 5.66%.
• article_number: 9012906
• pdf_url: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9012906
• html_url: https://ieeexplore.ieee.org/document/9012906/
• abstract_url: https://ieeexplore.ieee.org/document/9012906/
• publication_title: 2019 IEEE International Conference on Integrated Circuits, Technologies and Applications (ICTA)
• conference_location: Chengdu, China
• conference_dates: 13-15 Nov. 2019
• publication_number: 8989856
• is_number: 9012808
• publication_year: 2019
• publication_date: 13-15 Nov. 2019
• start_page: 150
• end_page: 155
• citing_paper_count: 3
• citing_patent_count: 1
• download_count: 570
• insert_date: 20200228
• index_terms:
  • ieee_terms:
    • Arrays
    • Convolution
    • Neural networks
    • Sensors
    • Random access memory
    • Voltage control
  • author_terms:
    • Binary Neural Network
    • Low energy
    • CMOS Image Sensor
    • Current Mode
    • Dynamic Configuration
  • dynamic_index_terms:
    • Application Of Intelligence
    • Perceptions Of Intelligence
    • Neural Network
    • Energy Efficiency
    • Flow Data
    • Analog-to-digital Converter
    • Analog-to-digital
    • Digital-to-analog Converter
    • Image Sensor
    • Camera Sensor
    • Accuracy Loss
    • Energy Budget
    • Binary Network
    • Convolutional Neural Network Algorithm
    • Calculation Module
    • Computation Module
    • Analog Computing
    • Analogous Calculations
    • Analogous Computation
    • Analog Domain
    • Convolutional Layers
    • Input Image
    • Feature Maps
    • Processing Unit
    • Batch Normalization
    • Voltage Regulation
    • Voltage Stability
    • MNIST Dataset
    • MNIST Data Set
    • Pixel Array
    • System Energy Efficiency
    • Example Of Network
    • Output Buffer
    • Peripheral Circuits
    • Perceptual System
    • Technology Node
    • Computing Units
    • Calculation Unit
• isbn_formats:
  • format: Print on Demand(PoD) ISBN, value: 978-1-7281-5041-3, isbnType: New-2005
  • format: USB ISBN, value: 978-1-7281-5039-0, isbnType: New-2005
  • format: Print ISBN, value: 978-1-7281-5167-0, isbnType: New-2005
  • format: Electronic ISBN, value: 978-1-7281-5040-6, isbnType: New-2005
• authors:
  • Author Name: Tianrui Ma
    Affiliation: Department of Electronic Engineering, BNRist, Tsinghua University, Beijing, China
    Author URL: https://ieeexplore.ieee.org/author/37088214308
    ID: 37088214308
    Order: 1
    Author Affiliations:
    • Department of Electronic Engineering, BNRist, Tsinghua University, Beijing, China
  • Author Name: Kaige Jia
    Affiliation: Department of Electronic Engineering, BNRist, Tsinghua University, Beijing, China
    Author URL: https://ieeexplore.ieee.org/author/37085990712
    ID: 37085990712
    Order: 2
    Author Affiliations:
    • Department of Electronic Engineering, BNRist, Tsinghua University, Beijing, China
  • Author Name: Xin Zhu
    Affiliation: Department of Electronic Engineering, BNRist, Tsinghua University, Beijing, China
    Author URL: https://ieeexplore.ieee.org/author/37088216652
    ID: 37088216652
    Order: 3
    Author Affiliations:
    • Department of Electronic Engineering, BNRist, Tsinghua University, Beijing, China
  • Author Name: Fei QiaO
    Affiliation: Department of Electronic Engineering, BNRist, Tsinghua University, Beijing, China
    Author URL: https://ieeexplore.ieee.org/author/37272195200
    ID: 37272195200
    Order: 4
    Author Affiliations:
    • Department of Electronic Engineering, BNRist, Tsinghua University, Beijing, China
  • Author Name: Qi Wei
    Affiliation: Department of Precision Instrument, Tsinghua University, Beijing, China
    Author URL: https://ieeexplore.ieee.org/author/37582779500
    ID: 37582779500
    Order: 5
    Author Affiliations:
    • Department of Precision Instrument, Tsinghua University, Beijing, China
  • Author Name: Huichan Zhao
    Affiliation: Department of Mechanical Engineering, Tsinghua University, Beijing, China
    Author URL: https://ieeexplore.ieee.org/author/37088219492
    ID: 37088219492
    Order: 6
    Author Affiliations:
    • Department of Mechanical Engineering, Tsinghua University, Beijing, China
  • Author Name: Xinjun Liu
    Affiliation: Department of Mechanical Engineering, Tsinghua University, Beijing, China
    Author URL: https://ieeexplore.ieee.org/author/37085996613
    ID: 37085996613
    Order: 7
    Author Affiliations:
    • Department of Mechanical Engineering, Tsinghua University, Beijing, China
  • Author Name: Huazhong Yang
    Affiliation: Department of Electronic Engineering, BNRist, Tsinghua University, Beijing, China
    Author URL: https://ieeexplore.ieee.org/author/37291236500
    ID: 37291236500
    Order: 8
    Author Affiliations:
    • Department of Electronic Engineering, BNRist, Tsinghua University, Beijing, China

Image Sensor

• sensor_type: CMOS
• resolution: 152x152x3
• dynamic_range: Not specified
• pixel_size: Not specified
• dark_current: Not specified

Optical Data

• focal_length: Not specified
• aperture: Not specified
• field_of_view: Not specified
• distortion: Not specified

Performance Metrics

• power_consumption: 27 μJ/frame
• noise: 70%-85% energy consumption from ADCs and digital readout circuits

Applications & Benefits

• cell_imaging: Used in emerging smart-sensing systems like wearable devices, AR/VR, wireless sensing nodes.
• benefits: Continuous imaging and instantaneous processing, with low energy consumption requirements.

Supporting Organizations

• supported_by: Tsinghua University, BNRist, Beihang University

Manuscript Details

• publication_date: 13-15 Nov. 2019

Relevancy Score

• score: 9
• missing_fields:
  • fill factor
  • quantum efficiency
  • readout speed
  • noise sources
  • analog-to-digital conversion techniques
  • microlenses
  • on-chip colour filters
  • global or rolling shutters
  • backside illumination

Processed JSON Filename

request_3143a095-1426-49c9-b0c3-fafd7660d392-an_analogmemoryless_near_sensor_computing_architecture_for_alwayson_intelligent_perception_applications.json