Near-sensor and in-sensor computing

Zhou, Feichi; Chai, Yang

doi:10.1038/s41928-020-00501-9

Cited by 511 publications

(471 citation statements)

References 45 publications

Supporting

Mentioning

418

Contrasting

Order By: Relevance

“…For the AND gate, the widths of both optical and electrical pulses are decreased to 1 s and 0.8 s respectively. Such definitions are consistent with the published work [34,35]. When there exist either an optical or electrical signal, the corresponding variable can be considered "1"; otherwise, p and q are assigned to "0".…”

Section: Introductionsupporting

confidence: 83%

Optically Controlled Ferroelectric Nanodomains for Logic-in-Memory Photonic Devices With Simplified Structures

Xue

Periyanagounder

et al. 2021

IEEE Trans. Electron Devices

View full text Add to dashboard Cite

We report visible-light-induced ferroelectric nanodomain reversal in conductive van der Waals (vdW) ferroelectric -In2Se3. We show its promising application in two-terminal optoelectronic memory devices. Compared to other vdW materials based optoelectronic memories, such a novel working prototype allows for the device operation confined within a single material channel bridging its two electrodes, which greatly reduces the complexities of device construction. In addition, using -In2Se3 memory devices we also demonstrate the universal OR and AND optical logic gates for logic-in-memory application. Our results provide a new avenue to design simplified structures of vdW materials based optoelectronic memories for dense device integration and next-generation computation.

show abstract

Section: Introductionsupporting

confidence: 83%

Optically Controlled Ferroelectric Nanodomains for Logic-in-Memory Photonic Devices With Simplified Structures

Xue

Periyanagounder

et al. 2021

IEEE Trans. Electron Devices

View full text Add to dashboard Cite

show abstract

“…[54,55] However, memristors suffer from non-linear weight update, limited conductance states, as well as coupled write and read operations. For optoelectronic synapses, optical signal sensing and synaptic functions are integrated into one device, [8,23,56] exhibiting great advantages in application of neuromorphic visual systems with simplified circuitry, high efficiency, and low energy consumption. Table S1, Supporting Information, summarizes the parameters of the GDY/ MoS 2 -based EGT in comparison with those of the reported EGTs.…”

Section: (5 Of 12)mentioning

confidence: 99%

Non‐Volatile Electrolyte‐Gated Transistors Based on Graphdiyne/MoS₂ with Robust Stability for Low‐Power Neuromorphic Computing and Logic‐In‐Memory

Yao

Chen

et al. 2021

Adv Funct Materials

View full text Add to dashboard Cite

Artificial synapses are the key building blocks for low-power neuromorphic computing that can go beyond the constraints of von Neumann architecture. In comparison with two-terminal memristors and three-terminal transistors with filament-formation and charge-trapping mechanisms, emerging electrolyte-gated transistors (EGTs) have been demonstrated as a promising candidate for neuromorphic applications due to their prominent analog switching performance. Here, a novel graphdiyne (GDY)/MoS 2 -based EGT is proposed, where an ion-storage layer (GDY) is adopted to EGTs for the first time. Benefitting from this Li-ion-storage layer, the GDY/MoS 2 -based EGT features a robust stability (variation < 1% for over 2000 cycles), an ultralow energy consumption (50 aJ µm −2 ), and long retention characteristics (>10 4 s). In addition, a quasi-linear conductance update with low noise (1.3%), an ultrahigh G max /G min ratio (10 3 ), and an ultralow readout conductance (<10 nS) have been demonstrated by this device, enabling the implementation of the neuromorphic computing with near-ideal accuracies. Moreover, the nonvolatile characteristics of the GDY/MoS 2 -based EGT enable it to demonstrate logic-in-memory functions, which can execute logic processing and store logic results in a single device. These results highlight the potential of the GDY/MoS 2 -based EGT for next-generation low-power electronics beyond von Neumann architecture.

show abstract

“…This work trained a binarized CNN which introduces batch norm and use both binary weights and activations to reduce the error caused by continuous analogue electronic current computing [11]. A car localisation and tracking task are exploited based on the proposed neural network structure.…”

Section: Experiments On Platform a Binarized Cnn For Car Localisation From A Dronementioning

confidence: 99%

Agile reactive navigation for a non‐holonomic mobile robot using a pixel processor array

Liu

Bose

Greatwood

et al. 2021

IET Image Processing

View full text Add to dashboard Cite

This paper presents an agile reactive navigation strategy for driving a non-holonomic ground vehicle around a pre-set course of gates in a cluttered environment using a lowcost processor array sensor. This enables machine vision tasks to be performed directly upon the sensor's image plane, rather than using a separate general-purpose computer. The authors demonstrate a small ground vehicle running through or avoiding multiple gates at high speed using minimal computational resources. To achieve this, target tracking algorithms are developed for the Pixel Processing Array and captured images are then processed directly on the vision sensor acquiring target information for controlling the ground vehicle. The algorithm can run at up to 2000 fps outdoors and 200 fps at indoor illumination levels. Conducting image processing at the sensor level avoids the bottleneck of image transfer encountered in conventional sensors. The real-time performance of onboard image processing and robustness is validated through experiments. Experimental results demonstrate the algorithm's ability to enable a ground vehicle to navigate at an average speed of 2.20 m/s for passing through multiple gates and 3.88 m/s for a 'slalom' task in an environment featuring significant visual clutter. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

Near-sensor and in-sensor computing

Cited by 511 publications

References 45 publications

Optically Controlled Ferroelectric Nanodomains for Logic-in-Memory Photonic Devices With Simplified Structures

Optically Controlled Ferroelectric Nanodomains for Logic-in-Memory Photonic Devices With Simplified Structures

Non‐Volatile Electrolyte‐Gated Transistors Based on Graphdiyne/MoS₂ with Robust Stability for Low‐Power Neuromorphic Computing and Logic‐In‐Memory

Agile reactive navigation for a non‐holonomic mobile robot using a pixel processor array

Contact Info

Product

Resources

About

Near-sensor and in-sensor computing

Cited by 511 publications

References 45 publications

Optically Controlled Ferroelectric Nanodomains for Logic-in-Memory Photonic Devices With Simplified Structures

Optically Controlled Ferroelectric Nanodomains for Logic-in-Memory Photonic Devices With Simplified Structures

Non‐Volatile Electrolyte‐Gated Transistors Based on Graphdiyne/MoS2 with Robust Stability for Low‐Power Neuromorphic Computing and Logic‐In‐Memory

Agile reactive navigation for a non‐holonomic mobile robot using a pixel processor array

Contact Info

Product

Resources

About

Non‐Volatile Electrolyte‐Gated Transistors Based on Graphdiyne/MoS₂ with Robust Stability for Low‐Power Neuromorphic Computing and Logic‐In‐Memory