A Two-Axis Goniometric Sensor for Tracking Finger Motion

Lefan Wang, Turgut Meydan, Paul Ieuan Williams2017 · DOI

TL;DR

Researchers have developed a novel two-axis optical sensor for tracking complex finger movements, overcoming limitations of existing systems that often only measure single degrees of freedom. This 3D-printed sensor, based on Malus' Law, offers high accuracy and a wide measurement range, paving the way for more sophisticated hand motion analysis in fields like biomechanics and medicine.

What Problem Does This Paper Address?

Tracking precise finger movements is crucial for understanding human dexterity, with applications ranging from rehabilitation to virtual reality. However, existing hand tracking systems often struggle to capture the full spectrum of finger motion. While flexion/extension (bending and straightening) is relatively easy to measure, more complex movements like abduction/adduction (spreading and bringing fingers together) and circumduction (circular motion) are significantly harder to quantify accurately. Many current systems are limited to single degrees of freedom (DOF), and even advanced glove-based systems can suffer from inaccuracies due to sensor deformation or interference. This paper aims to address this gap by proposing a compact, two-axis optical sensor capable of simultaneously measuring two DOF of finger motion.

Methods

The core of the proposed system is a two-axis optical sensor fabricated using 3D printing. Each axis of the sensor utilizes two optical pathways, each containing a linear polarizer, a polarizing film analyzer, and a photodiode. The principle of operation is based on Malus' Law, which describes the attenuation of light transmitted through polarizing filters. As the finger joint moves, the sensor's orientation changes, altering the angle between the polarizers and analyzers. This change in angle affects the transmitted light intensity, which is then converted into an electrical signal by the photodiode.

To achieve a wide measurement range and improve sensitivity, a data fusion approach is employed. For each axis, two channels are used, with their analyzers oriented at an angle to each other. This ensures that at least one channel operates in a region of maximum sensitivity, regardless of the finger's position. A weighted average method is then used to combine the readings from these two channels, allowing for a full 180° measurement range and improved accuracy. The sensor itself is constructed from a nanoparticle-filled material (RCP30) for stiffness and opacity, minimizing interference from ambient light.

Datasets

Not discussed in the paper. The paper focuses on evaluating the sensor's performance using an automated measurement apparatus and real-world finger motion tests.

Key Insights and Ideas

  • Malus' Law for Dexterity: Applying the fundamental physics of light polarization (Malus' Law) to a compact, 3D-printed sensor provides a surprisingly effective and accurate way to measure angular motion.
  • Two-Channel Fusion for Range: Using two overlapping channels per axis, rather than a single sensor, is key to achieving a wide 180° measurement range and maintaining sensitivity across the entire spectrum of motion.
  • 3D Printing for Customization: The use of 3D printing allows for the creation of a compact and integrated sensor design, which can be further customized for different applications or finger joints.
  • Beyond Flexion/Extension: The sensor's ability to simultaneously track abduction/adduction and even circumduction opens up new possibilities for detailed hand biomechanics research and clinical assessment.

Key Results

The developed two-axis optical sensor demonstrated:

  • High Accuracy: An accuracy of ±0.3° was achieved, which is significantly better than many existing technologies like linear potentiometers (0.7°), optical fiber sensors (0.9°), and resistive bend sensors (1.5°).
  • Wide Measurement Range: The sensor can measure up to 180° for each axis, sufficient for the full range of finger motion.
  • High Repeatability: A repeatability of 0.5% was reported.
  • Low Hysteresis Error: The overall hysteresis error was less than 1.7%.
  • Successful Motion Tracking: The sensor successfully tracked real-time finger movements, including flexion/extension, abduction/adduction, and circumduction, on the metacarpophalangeal (MCP) joint of the index finger.
  • Comparison to Photogrammetry: Angular measurements were within 2.7° of reference values obtained using photogrammetry.

Limitations and Future Work

The authors acknowledge several limitations and suggest future directions:

  • Attachment Method: The sensor was attached using adhesive plasters, which improves accuracy but may not be suitable for long-term monitoring or all applications. Future work could explore glove-based integration.
  • Single Finger and Subject: The study was conducted on a single finger (left index) of one test subject. Further studies with multiple participants and finger joints are needed to validate the sensor's reliability and reproducibility.
  • Size and Collision: The current sensor size may lead to collisions when multiple sensors are placed close together on the hand. Future work will focus on reducing the sensor's dimensions.
  • Circumduction Monitoring: While the sensor shows potential for circumduction tracking, the current single-camera setup was not sufficient for dynamic circumduction monitoring. A multi-camera system is suggested for this purpose.
  • Calibration: The authors note that time-consuming calibration is unnecessary for their sensor, which is a significant advantage.
  • Broader Applications: The sensor's design is adaptable and could be fabricated with different dimensions for monitoring other joints in the body, such as the ankle, wrist, and knee.

Summary updated

BibTeX
@misc{wang2017two,
  author = {Lefan Wang and Turgut Meydan and Paul Ieuan Williams},
  title = {{A Two-Axis Goniometric Sensor for Tracking Finger Motion}},
  year = {2017},
  doi = {10.3390/s17040770},
}

Loading related content...

Figure