IIT Delhi Study Shows Hidden GPS Clues Can Reveal a User’s Surroundings

Researchers build a system that reads fine-grained GPS signals to sense rooms, activities and building layout

A new study from IIT Delhi shows that even when GPS signals are weak, smartphones still carry enough hidden information to reveal where a user is and what they are doing. The research suggests that subtle variations in GPS data—normally ignored by apps—can quietly expose a person’s environment, activity and even the structure of the building around them.

More than maps and navigation
Most of us think of GPS as a tool for directions or tracking. But every smartphone receives many tiny GPS details such as frequency shifts, signal noise and power changes. Users never notice these signals, and navigation apps do not need them. The study finds that these ignored signals carry rich clues about the user’s surroundings.

Breakthrough by IIT Delhi researchers
The work was done by Soham Nag, M.Tech student at IIT Delhi’s Center of Excellence in Cyber Systems and Information Assurance, and Dr. Smruti R. Sarangi, Professor in the Computer Science and Engineering Department. Their study, “AndroCon: An Android Phone-based Sensor for Ambient, Human Activity and Layout Sensing using Fine-Grained GPS Information”, is published in ACM Transactions on Sensor Networks, a top journal in the field.

AndroCon: a covert sensor hidden inside GPS
The researchers built AndroCon, the first system to show that fine-grained GPS data—available to any Android app with precise location permissions—can act as a silent sensor. Without using the camera, microphone or motion sensors, AndroCon reads nine low-level GPS parameters like Doppler shift, signal strength and multipath interference.

Using this data, it can tell whether a person is sitting, standing, lying down, inside a metro, on a flight, outdoors in a park or in a crowded area. It can also sense whether a room is empty or filled with people.

How the system reads the noise
To convert noisy raw signals into clear results, the team used a mix of classical signal processing and machine learning. According to Prof. Sarangi, a year-long study covering 40,000 square kilometres and many phone models showed AndroCon can detect surroundings with 99% accuracy and human activities with over 87% accuracy—even subtle movements like hand-waving near the phone.

Mapping the inside of buildings with GPS
The same approach can sketch basic indoor layouts. Using only GPS patterns and how the user moves, AndroCon can identify rooms, staircases and elevators with an error of less than four metres.

A new tool—and a warning
While the system opens doors for new context-aware services, it also raises serious privacy concerns. Any Android app with precise location access could potentially infer sensitive details about a user’s life without permission.

Prof. Sarangi said the study shows GPS has a hidden side. It is a powerful channel that can sense the world around us, but also a reminder that familiar technology can have unseen risks if misused.

Contact: Prof. Smruti R. Sarangi – srsarangi@cse.iitd.ac.in