TL;DR
This paper presents a novel hybrid system for a "padless" mouse that uses MEMS accelerometers to track 2D hand movements and convert them into cursor positions. The system achieves a minimum detectable acceleration of 5.3 mg, enabling precise cursor control without requiring a special surface.
What Problem Does This Paper Address?
Traditional mouse technologies, like rolling-ball or optical sensors, rely on physical contact with a surface or a dedicated pad. This limits user freedom and can be problematic on uneven or obstructed surfaces. The authors aim to create a more versatile mouse that can track position in 2D space without any external surface dependency, offering greater flexibility and a more desirable user experience.
Methods
The proposed system is a hybrid design, meaning its components are not integrated onto a single chip but rather interfaced externally. It comprises four key modules:
- MEMS Accelerometers: Two lateral capacitive accelerometers, fabricated using Silicon-On-Glass (SOG) technology, are used to measure X and Y-axis accelerations. These devices leverage the change in capacitance due to the displacement of a proof mass.
- CMOS Analog Read-Out Circuitry: A switched-capacitor sigma-delta modulator is employed to convert the differential capacitance variations from the MEMS sensors into a digital pulse-width modulated (PWM) signal. This circuit is designed for high sensitivity and low drift.
- Acceleration Magnitude Extractor: This module processes the PWM signal to extract both the direction (polarity) and magnitude of the acceleration. The pulse width directly corresponds to the acceleration magnitude.
- 16-bit RISC Microprocessor: This custom-designed processor receives the processed acceleration data, filters out noise, and performs double integration to calculate the X and Y coordinates for cursor movement.
The design methodology follows a top-down approach for partitioning and a bottom-up approach for verification, with individual modules being full-custom designs.
Datasets
Not discussed in the paper.
Key Insights and Ideas
- Padless Operation: The core innovation is the ability to track mouse movement in 2D space without relying on a physical surface or pad, significantly enhancing user freedom.
- Hybrid System Design: The paper demonstrates a practical approach to integrating disparate technologies (MEMS, analog CMOS, digital RISC) through careful interfacing, though it acknowledges the limitations of this approach.
- Capacitive MEMS for Sensitivity: The choice of capacitive MEMS accelerometers is highlighted for their high sensitivity, good DC response, and low drift, making them suitable for detecting subtle hand movements.
- Sigma-Delta Modulation for Analog-to-Digital Conversion: The use of a switched-capacitor sigma-delta modulator provides an efficient way to convert analog capacitance changes into a digital PWM signal, simplifying the interface to the digital processor.
- Noise Dominance: The analysis reveals that mass residual motion noise is the dominant factor limiting the minimum detectable acceleration, suggesting that increasing sampling frequency could further improve performance at the cost of power and complexity.
Key Results
- Minimum Detectable Acceleration: The system is designed to detect accelerations as small as 5.3 mg, which is stated to be below the smallest acceleration a human hand can generate or perceive.
- Operating Frequency: The system can operate up to 18 MHz, with the microprocessor achieving a maximum clock frequency of approximately 18 MHz, well above the target of 100 kHz for coordinate calculation.
- Analog Circuit Power Consumption: The analog read-out circuit consumes an estimated 3 mW.
- Microprocessor Power Consumption: The microprocessor consumes an estimated 10 µW at 100 kHz operation frequency.
- MEMS Accelerometer Sensitivity: The MEMS accelerometer is designed with a sensitivity of 0.78 pF/g.
Limitations and Future Work
- Hybrid System Drawbacks: The authors acknowledge that the hybrid, multi-chip nature of the system limits its size and can introduce capacitive and resistive parasitics that might degrade performance.
- Monolithic Implementation: As a future direction, the paper proposes a System-on-Chip (SoC) design and fabrication approach where all modules (MEMS, analog, digital) are integrated onto a single wafer. This would reduce costs, minimize parasitic effects, and potentially improve performance.
- 3D Tracking: The system can be extended to three-dimensional position detection by adding a third MEMS accelerometer to measure Z-axis movement.
- Noise Reduction: While mass residual motion noise dominates, the paper suggests that increasing the sampling frequency could further reduce this noise, though it would increase power consumption and complexity.
Summary updated