Software Campus: MARC² – Improving Activity Recognition on Mobile Devices Using Crowdsourcing and Cloud Information

MARC² (Mobile Activity Recognition with Crowdsourcing and Cloud) focuses on improving human activity recognition through distributed data collection and cloud-based learning.

Enhance mobile activity recognition using crowdsourced labeled data. Utilize cloud computation for continuous model improvement. Reduce on-device computation and energy use.

Combines mobile sensing, crowdsourcing, and cloud analytics. Learns from diverse user data for robust recognition. Conducted under the Software Campus program.

Improves accuracy and adaptability of mobile activity recognition. Enables scalable user-specific learning while maintaining device efficiency. Advances context-aware computing research at TECO.