EarRumble: Discreet Hands- and Eyes-Free Input by Voluntary Tensor Tympani Muscle Contraction

Title of the Paper

EarRumble: Discreet Hands- and Eyes-Free Input by Voluntary Tensor Tympani Muscle Contraction

Paper Information

• Research Domain: Non-traditional input technologies in human-computer interaction
• Keywords: tensor tympani muscle, discreet interaction, subtle gestures, earables, hearables, in-ear barometry

Research Background and Problem Statement

• Problem or Challenge:

  • In mobile scenarios, users require a way to interact without using their hands or eyes. This method should avoid large gestures and ensure discretion and social acceptability.
  • Current earphone-related interaction technologies often rely on noticeable movements, which may raise privacy concerns or attract external attention.
  • Technologies leveraging voluntary control of ear muscles—such as the tensor tympani—have not yet been developed for device interaction.

• Research Significance:

  • Providing a discreet, hands- and eyes-free interaction technology can enhance the mobile interaction experience, especially in noisy public or crowded environments.
  • Exploring the feasibility of using the tensor tympani muscle as an input source opens new avenues for future earphone interaction technologies.

• Research Motivation and Related Work:

  • Related studies have explored areas such as hand micro-movements, oral interfaces, and motion detection in earphones, but have not sufficiently investigated ear muscle control-based interaction.
  • The tensor tympani is a small muscle in the human middle ear. Its voluntary contraction causes eardrum displacement and pressure changes, which can be detected to identify user actions.

Solution

• Method or Solution:

  • Introducing a technology called "EarRumble," which utilizes voluntary tensor tympani muscle contractions for interaction.
  • Measuring pressure changes caused by muscle contractions using sealed in-ear pressure sensors.

• Innovative Contributions:

  • The first proposal to use voluntary tensor tympani muscle contractions for interaction.
  • Offers a highly discreet, low-cost interaction method that requires no significant physical movements.
  • Expands the interaction design space by introducing three simple ear actions (single contraction, double contraction, sustained contraction).

• Implementation Steps and Key Technologies:

  1. Hardware Design: Custom earphones equipped with pressure sensors and speaker components to measure pressure changes in the sealed ear canal.
  2. Action Recognition Algorithm: Employing machine learning classifiers (e.g., XGBoost) to process pressure data and recognize ear actions (single contraction, double contraction, sustained contraction).
  3. User Study: Collecting data on tensor tympani muscle control ability through online surveys and validating feasibility in a laboratory setting.

Research Outcomes

• Specific Results:

  • Approximately 43% of survey participants reported being able to voluntarily control their tensor tympani muscle, demonstrating a broad potential user base.
  • The classifiers for the three ear actions achieved up to 95% accuracy (using the XGBoost model).
  • Experiments showed that EarRumble could be applied to practical scenarios such as answering calls and controlling audio playback.

• Advantages and Contributions:

  • Compared to traditional earphone interaction technologies, EarRumble does not require hands or eyes, offering greater discretion.
  • Actions can be performed without noticeable physical movement, avoiding inconvenience in social environments.
  • Pioneering a novel interaction method based on human biological characteristics.

• Experimental or Evaluation Results:

  • Analysis showed users could quickly initiate actions (average start time of 308ms).
  • Users could easily perform single and double ear contractions using EarRumble, but detecting sustained contractions remains a technical challenge.
  • User experiments revealed that despite technical limitations, participants generally found the interaction method "magical" and "practical."

• Limitations and Future Directions:

  • Limitations:

    • The current technology requires sealing the ear canal, which may cause discomfort or raise safety concerns for prolonged use.
    • Detection of sustained contractions is not sufficiently accurate, requiring further optimization of sensing methods.
    • The social acceptability and performance of this technology in dynamic environments have not yet been tested.

  • Future Directions:

    • Exploring alternative sensing methods, such as incorporating cameras or acoustic impedance measurements, to improve the accuracy of sustained contraction detection.
    • Investigating the health impacts of long-term voluntary tensor tympani muscle contractions.
    • Studying the potential of EarRumble technology in broader mobile interaction and complex application scenarios.