MARS, a multi-agent system for assessing rowers' coordination via motion-based stigmergy.

Area of Science:

• Sports Science
• Robotics and Intelligent Systems
• Human-Computer Interaction

Background:

• Rowing performance heavily relies on crew synchronization, which is challenging to achieve and monitor.
• Existing methods use wireless sensors but lack direct support for coaching decisions without complex biomechanical analysis.
• A need exists for intuitive systems to measure crew coordination and individual rower asynchrony.

Purpose of the Study:

• To present a multi-agent information-processing system for on-water measurement of rowing crew asynchrony.
• To provide coaches with a tool to assess overall crew coordination and individual rower performance relative to the crew.
• To avoid complex biomechanical analysis while supporting coaching decisions.

Main Methods:

• A multi-agent system processing information on-water.
• Level 1: Marking agents capture rower motion and generate collective marks via stigmergic cooperation.
• Level 2: Similarity agents assess motion against optimal patterns.
• Level 3: Granulation agents extract asynchrony indicators.

Main Results:

• The system effectively measures both overall crew asynchrony and individual rower asynchrony.
• Experimental validation in real-world rowing scenarios demonstrated the system's effectiveness.
• The system provides actionable insights for coaching without requiring deep biomechanical expertise.

Conclusions:

• The proposed multi-agent system offers a practical solution for monitoring rowing synchronization.
• It enhances training by providing objective data on crew and individual asynchrony.
• This approach supports coaches' decision-making with accessible, sensor-based performance metrics.