MouseRing: Always-available Touchpad Interaction with IMU Rings

Title of the Paper

MouseRing: Always-available Touchpad Interaction with IMU Rings

Paper Information

• Field: Human-Computer Interaction, Wearable Interaction Devices
• Keywords: Finger tracking, IMU, touch interface, input device, wearable device, pointing technology, smart ring, physical modeling, machine learning, user experience

Research Background and Problem Statement

• What problems or challenges did the authors identify?
  • Traditional mice and touchpads are limited in mobile environments and cannot meet the demand for "always-available" input solutions in scenarios such as AR/VR and large-screen interactions.
  • Finger tracking technology based on IMU (Inertial Measurement Unit) suffers from insufficient information and noisy signals, leading to accuracy issues, particularly in achieving high-precision 2D cursor control and target selection tasks.
• Why is this problem important?
  • Enhances user input efficiency and comfort across various scenarios, such as AR/VR, large displays, and smart homes.
  • Provides a portable and low-power interaction solution, reducing the reliance on handheld devices or environment-deployed sensors.
• Motivation and related work:
  • Existing mouse devices and gesture recognition technologies, such as cameras and electromagnetic sensors, require complex setups and high computational power, limiting their universality.
  • IMU-based interaction devices, such as LightRing and AnywhereTouch, have demonstrated the potential for finger sliding and direction prediction, but their accuracy and stability need further improvement.

Solution

• What methods or solutions did the authors propose?
  • Proposed a novel device called "MouseRing," a ring-shaped device equipped with an IMU that enables continuous finger sliding tracking on unmodified physical surfaces.
  • Developed a finger velocity prediction and correction algorithm based on physical constraints and machine learning models, integrating touch state detection.
• What is innovative about this solution?
  • Utilizes physical constraints of finger joints (e.g., coplanarity, velocity consistency, rigidity constraints) combined with machine learning to improve motion sensing accuracy.
  • Designed a high-precision pose estimation algorithm and an RNN-based finger velocity prediction model, with velocity corrections based on physical constraints to enhance stability.
• Implementation steps:
  1. Data Collection: Collected finger motion data using IMU sensors, optical capture systems, and pressure touchpads to build a multimodal finger sliding dataset.
  2. Physical Relationship Analysis: Analyzed the data from the finger sliding process to validate physical principles such as joint coplanarity and velocity correlation.
  3. Algorithm Development and Optimization: Used RNN to predict finger motion velocity and corrected the velocity through physical constraints for real-time trajectory prediction.
  4. Experiments and Evaluation: Evaluated algorithm performance and user interaction experience in both ideal laboratory environments and real-world scenarios.

Research Outcomes

• What specific results were achieved?
  • MouseRing achieved accurate finger sliding tracking through a machine learning method integrated with physical constraints, with an average angular error (𝜃𝑙𝑒𝑟𝑟𝑜𝑟) of 6.61°.
  • Experiments demonstrated that MouseRing's input efficiency was comparable to a touchpad (average task completion times of 658.1ms and 629.1ms, respectively), outperforming portable air mice (AirMouse).
  • The wireless version of MouseRing exhibited good input stability and response speed across different surfaces (e.g., desktop, sofa, wall, leg) and various postures (sitting, standing).
• What advantages does it have compared to existing solutions?
  • The MouseRing device is lightweight, wearable at all times, and requires no calibration, making it suitable for various surfaces and scenarios.
  • Improved the utilization efficiency of IMU sensor data, supporting high-precision and stable 2D cursor control.
  • Offers a more natural and comfortable interaction experience, reducing fatigue from mid-air hand interactions.
• What were the experimental or evaluation results?
  • In Fitts’ Law experiments, the dual-ring configuration of MouseRing achieved input speeds nearly equivalent to a touchpad; the single-ring configuration was about 20% slower in task completion time.
  • In large-screen operation scenarios, MouseRing demonstrated stable input efficiency on both hard surfaces (e.g., desktops, walls) and soft surfaces (e.g., sofas, legs).
  • Subjective user evaluations indicated that MouseRing was more comfortable than a mouse or AirMouse, with the single-ring configuration offering better long-term wearability.
• Limitations and future directions
  • The single-ring configuration's input precision is suitable for non-fine tasks but shows limitations in precise target selection.
  • Indoor magnetic field interference restricts the use of 9-axis IMUs; future improvements could involve predicting spatial magnetic field distributions or testing in outdoor environments.
  • Long-term wear may cause comfort issues, which could be addressed by optimizing the shape or wearing position of the ring device.
  • Personalized online calibration algorithms for individual users could further enhance the accuracy and adaptability of MouseRing.

The analysis above comprehensively summarizes the design background, technical innovations, and evaluation results of MouseRing, demonstrating its strong potential in multi-scenario interactions.