IIIT-Hyderabad students unveil innovative ML model-switching approach for real-time traffic monitoring on smartphones

The concept of model switching refers to using ML models on edge devices such as a smartphone, where the model balancer intelligently switches between models based on the incoming inputs

Hyderabad: The International Institute of Information Technology Hyderabad (IIIT-H) students have come up with a dynamic machine learning (ML) model-switching technique on smartphones, enabling real-time traffic monitoring that adapts to changing conditions depending on the traffic flow.

The team comprising undergraduate second-year CSE students — Kriti Gupta, Ananya Halgatti, Priyanshi Gupta, and Larissa Lavanya — under PhD student Akhila Matathammal’s mentorship and guidance of Prof. Vaidyanathan, worked on a dynamic model switching approach titled EdgeML Balancer. Prof. Vaidyanathan is part of software Architecture 4 Sustainability group at Software Engineering Research Centre, and EdgeML Balancer is for object detection on edge devices such as smartphones.

The concept of model switching refers to using ML models on edge devices such as a smartphone, where the model balancer intelligently switches between models based on the incoming inputs.

In this approach, students not only prototyped on the Qualcomm Innovators Development Kit (QIDK) platform to simulate different scenarios but also evaluated the approach on real-time traffic data using the smartphone.

Students came up with this new approach after they walked into the Embedded Systems workshop — hands-on course for second-year students that emphasises learning by doing — searching for a novel project to work on.