Algorithm for the Comparison of Human Periodic Movements Using Wearable Devices.

Area of Science:

• Motor Skill Acquisition
• Virtual Learning Environments
• Biomechanical Analysis

Background:

• Traditional motor skill learning often relies on video demonstrations with limited feedback.
• Virtual environments offer immersive platforms for skill acquisition but often lack detailed performance analysis.
• Effective feedback mechanisms are crucial for optimizing the learning process in motor skills.

Purpose of the Study:

• To propose and evaluate an algorithm for comparing expert and learner movements in a virtual environment.
• To provide quantitative feedback on motor skill execution based on movement similarity.
• To enhance the effectiveness of motor skill acquisition through objective performance analysis.

Main Methods:

• Data capture using wearable accelerometers for expert and learner movements (salsa dance steps).
• Data preprocessing including Pearson iterations, synchronization, filtering, and normalization.
• Comparative analysis using Dynamic Time Warping (DTW), linear regression, and error analysis.

Main Results:

• The algorithm successfully compares periodic movements captured via wearable sensors.
• Quantitative assessment of movement similarity between expert and learner is achieved.
• Identification of discrepancies in movement cycles and execution is possible.

Conclusions:

• The developed algorithm offers a novel approach to providing feedback in virtual motor skill learning.
• Objective comparison of movements facilitates targeted improvement for learners.
• This method has potential applications in various domains requiring precise motor skill development.