Insight Research Ireland Centre brings ‘breakthrough’ methodology to wearable sensor tech
A team of researchers at the Insight Research Ireland Centre for Data Analytics in University of Galway has developed a new methodology to help machines to recognise human activity from sensors more accurately.
Human Activity Recognition is a technology that decodes our movements from wearable sensors and IoT devices. From tracking rehabilitation progress to enabling safer workplaces, Human Activity Recognition technology has the power to transform healthcare and everyday life. However, these systems are notoriously complex and fragile, demanding endless manual tuning and human oversight.
In a significant breakthrough in the field, Dr Nitesh Bharot (pictured), Dr Priyanka Verma and Prof John Breslin at the Insight Research Ireland Centre have developed the Automated Decision Maker; a system that automates the entire Human Activity Recognition pipeline, removing the bottlenecks of manual intervention.
The Automated Decision Maker automatically selects, tunes and optimises models, making Human Activity Recognition systems faster, more accurate and more reliable.
Published in the IEEE Journal of Biomedical and Health Informatics, this pioneering research has demonstrated accuracy rates of up to 99.78%.
Lead researcher Dr Nitesh Bharot said: ‘Healthcare providers can trust this data. Patients can be monitored remotely with fewer errors. Sports professionals can push their limits safely. Factories and smart homes can detect anomalies before accidents happen. This work is a step toward a future where machines don’t just collect data – they make sense of it reliably and at scale for us.’
Supported by Insight Research Ireland Centre for Data Analytics (Grant 12/RC/2289_P2).
