In-Memory-Computing Realization with a Photodiode/Memristor Based Vision Sensor

Abstract

State-of-the-art IoT technologies request novel design solutions in edge computing, resulting in even more portable and energy-efficient hardware for in-the-field processing tasks. Vision sensors, processors, and hardware accelerators are among the most demanding IoT applications. Resistance switching (RS) two-terminal devices are suitable for resistive RAMs (RRAM), a promising technology to realize storage class memories. Furthermore, due to their memristive nature, RRAMs are appropriate candidates for in-memory computing architectures. Recently, we demonstrated a CMOS compatible silicon nitride (SiNx) MIS RS device with memristive properties. In this paper, a report on a new photodiode-based vision sensor architecture with in-memory computing capability, relying on memristive device, is disclosed. In this context, the resistance switching dynamics of our memristive device were measured and a data-fitted behavioral model was extracted. SPICE simulations were made highlighting the in-memory computing capabilities of the proposed photodiode-one memristor pixel vision sensor. Finally, an integration and manufacturing perspective was discussed.

Energy-adaptive resistive switching with controllable thresholds in insulator–metal transition

“…Advanced information devices are fundamental and required in intensive information data and modern electronic applications. [1][2][3][4][5] Resistive switching (RS) has exhibited excellent prospects in low operation power and exible architecture towards neuromorphic techniques and articial intelligence. [6][7][8][9][10][11][12] These neuromorphic RS have shown promise in deep learning, 13,14 autopilot systems, 15,16 and man-machine interaction systems.…”

Neuromorphic Computing Based on CMOS-Integrated Memristive Arrays: Current State and Perspectives

“…Of course, this growth cannot be indefinitely large, and potential high performance and energy efficiency will be more influenced by design solutions at the processor and computing system architecture levels, especially the growing overhead costs of routing and input/output data in digital form (see also Section 3). For example, when it comes to processing signals of different nature, performance will be limited by the characteristics of sensors and information transmission interfaces, so devices for computing in sensors with direct transmission of information in analog form to a memristor-based computing device are currently being developed for such tasks [31,49].…”

Toward Reflective Spiking Neural Networks Exploiting Memristive Devices

“…The first steps have already been taken to combine memristive devices with photosensors. The described architecture of a 1D1R sensor for machine vision is a 20 × 20 or 32 × 32 matrix of SiN x memristive devices coupled to a photodiode or a phototransistor ( Vasileiadis et al, 2021a , b ). The coupling of memristors with photosensors shows that this approach can simulate some retinal functions ( Chen et al, 2018 ; Eshraghian et al, 2018 ).…”

“…Figure 8 illustrates the concept of analog memristive vision exploiting coupled memristors and photodiodes ( Vasileiadis et al, 2021a , b ). The 1D1R memristive sensor receives visual information ( Figure 8A ).…”