Encoding Manual Rotations on a Motionless Knob.

Area of Science:

• Human-Computer Interaction
• Embedded Systems Engineering
• Machine Learning Applications

Background:

• Physical knobs are preferred in professional settings for their cost-effectiveness and efficient rotation encoding.
• Traditional knobs are susceptible to electro-mechanical damage in harsh environments, posing safety and productivity risks.

Purpose of the Study:

• To design and prototype a motionless cylindrical device for encoding manual rotation.
• To overcome the limitations of physical knobs by offering a durable, damage-resistant alternative.

Main Methods:

• Utilized capacitive sensing to track finger contact positions on the device's lateral surface.
• Developed a neural network-based encoding algorithm for real-time manual rotation classification on embedded hardware.

Main Results:

• User tests demonstrated accuracy comparable to previous motionless knob experiments.
• A slight decrease in precision was noted, potentially due to the sensing technology and algorithm.
• Subjective feedback indicated a need for user adaptation to achieve intuitive interaction with the motionless knob.

Conclusions:

• The developed motionless knob provides a viable, robust alternative to traditional physical knobs.
• Achieving natural interaction with this novel device requires adapting deeply ingrained motor skills.
• Further refinement of sensing and algorithms may enhance precision and user experience.