Energy Efficient Artificial Olfactory System with Integrated Sensing and Computing Capabilities for Food Spoilage...

Area of Science:

• Materials Science
• Electrical Engineering
• Sensor Technology

Background:

• Artificial olfactory systems (AOSs) aim to replicate biological smell but often face challenges with high energy consumption and latency.
• Existing AOS designs struggle with data conversion and transmission, limiting their practical application.
• Developing efficient and compact AOSs is crucial for advancing real-world sensory technologies.

Purpose of the Study:

• To propose an energy- and area-efficient artificial olfactory system (AOS) leveraging near-sensor computing.
• To integrate sensing units and nonvolatile memory (NVM) for in-memory computation, reducing data processing overhead.
• To demonstrate the AOS's capability in detecting food spoilage markers and monitoring food freshness.

Main Methods:

• Designed an AOS integrating merged field effect transistor (FET)-type gas sensors and amplifier circuits with an AND-type nonvolatile memory (NVM) array.
• Implemented near-sensor computing where sensing signals are processed directly within the NVM array.
• Utilized thin zinc oxide films as gas-sensing materials for detecting hydrogen sulfide (H₂S) and ammonia (NH₃).

Main Results:

• Achieved low detection limits of 0.01 ppm for H₂S and NH₃, key indicators of high-protein food spoilage.
• Successfully demonstrated continuous monitoring of the entire spoilage process for chicken tenderloin.
• The system provided real-time freshness scores and tracked food conditions throughout spoilage.

Conclusions:

• The proposed near-sensor computing AOS offers significant improvements in energy and area efficiency compared to traditional systems.
• The developed AOS effectively detects critical food spoilage gases and can monitor food quality over time.
• The adaptable platform, with adjustable sensing temperature and programmable NVM cells, is suitable for diverse applications beyond food monitoring.