Exploiting Pull-In/Pull-Out Hysteresis in Electrostatic MEMS Sensor Networks to Realize a Novel Sensing...

Area of Science:

• Electronics
• Materials Science
• Computer Science

Background:

• Wearable electronics often require significant power, size, and cost due to reliance on cloud computing and separate processing units.
• Microelectromechanical systems (MEMS) are typically used solely for sensing, with computation handled externally.
• Existing MEMS networks face challenges in device coupling, limiting their computational capabilities.

Purpose of the Study:

• To present a novel computing approach for reducing power consumption, size, and cost in wearable electronics.
• To demonstrate the feasibility of using microelectromechanical systems (MEMS) sensors for simultaneous sensing and computing.
• To overcome coupling challenges in MEMS networks for advanced applications.

Main Methods:

• Utilizing the inherent pull-in/pull-out hysteresis of MEMS sensors for local data processing.
• Developing and fabricating a network of MEMS devices with specialized actuating and biasing assemblies for inter-device coupling.
• Simulating a network of MEMS accelerometers to classify acceleration signals based on hysteresis.

Main Results:

• A simulation model successfully classified acceleration signals using MEMS accelerometer hysteresis, proving the concept.
• A fabricated MEMS network demonstrated effective coupling between devices using novel assemblies.
• The proposed method eliminates the need for cloud computing, reducing reliance on analog-to-digital converters and digital processors.

Conclusions:

• Local sensing and computing at the MEMS sensor node is a viable strategy for low-power, low-cost wearable electronics.
• The inherent hysteresis of MEMS devices can be leveraged for computational tasks, simplifying system architecture.
• The developed coupling mechanisms address a key challenge in MEMS neural networks, paving the way for more complex integrated systems.