Researchers designed an automated acoustic monitoring tool that flags indoor coughing without identifying individual patients.
RT’s Three Key Takeaways:
- Acoustic Surveillance: The CoughNet prototype uses networked sensors and artificial intelligence to count room coughs and differentiate between individual coughers without collecting personal identifiable information.
- Energy and Cost Efficiency: Built with low-cost hardware and low-power, long-range wireless technology, the system operates on minimal battery power and local processing to reduce operational burdens.
- Outbreak Detection: Future iterations aim to identify disease-specific cough signatures to help healthcare facilities track potential spikes in respiratory conditions like influenza and bronchitis.
A team of researchers at Binghamton University has developed an acoustic sensor system designed to help healthcare professionals monitor cough frequency and detect potential respiratory disease outbreaks in indoor spaces, according to the university.
The prototype, known as CoughNet, was developed by Dali Ismail, assistant professor at Binghamton University’s Thomas J Watson College of Engineering and Applied Science, along with PhD students Amir Esmaeili and Maryam Fazli. The team presented the system at the 2026 Institute of Electrical and Electronics Engineers/Association for Computing Machinery (IEEE/ACM) Conference on Connected Health: Applications, Systems, and Engineering Technologies (CHASE).
CoughNet functions as a localized network of listening stations that capture cough noises, confirm whether the audio is an authentic cough, and determine if the event originated from an existing or new individual in the space.
Triangulating Coughs While Preserving Privacy
To track coughing events, the setup uses three Raspberry Pi single-board computers placed throughout a room. Each unit includes a built-in microphone. When a cough occurs, all three devices record the event, with the unit positioned closest to the individual capturing the highest amplitude.
The system runs the short audio file through an artificial intelligence (Ai) model to verify the sound. To determine whether the cough originated from someone already logged by the system, CoughNet assesses sound strength and arrival time across all three microphones.
“The microphone closer to the source acts as a reference microphone, and we can do correlation to determine if this cough is from the same exact person, because the same cough will be heard by the two other microphones, which are a little bit far away,” said Ismail, assistant professor of computing at Binghamton University, in a news release.
By establishing consistent arrival times and sound levels across repeated events, the platform can distinguish between separate patients in the same room. The system is designed specifically for healthcare environments where patient privacy is vital. Unlike surveillance platforms that incorporate thermal imaging, cameras, or voice-to-text transcription, CoughNet detects only acoustic patterns associated with coughing and deletes raw audio files immediately after analysis, according to the researchers.
“If this can achieve the same level of accuracy without relying on a camera or additional sensors, then I think we did the job,” said Ismail, assistant professor of computing at Binghamton University. “Especially if you think about indoor environments like a hospital.”
Energy Efficiency and Clinical Utility
The hardware uses low-power, long-range (LoRa) wireless networking, which provides indoor signal coverage that exceeds standard Wi-Fi setups while drawing minimal power. The sensors process clean audio files locally and only transmit complex, noisy recordings to a secondary computer when necessary. According to the development team, this communication model allows sensor nodes to operate on AA batteries for extended deployments.
“Most of the work on public health and disease monitoring has been very machine learning‑heavy and processing‑intensive,” said Ismail, assistant professor of computing at Binghamton University, in a news release. “We wanted to look at cough detection and identifying who is coughing as a lightweight application that could eventually run on practical devices.”
Looking ahead, the researchers are working to train the model to analyze specific acoustic signatures, aiming to differentiate cough sounds tied to specific conditions such as bronchitis, COVID-19, and influenza. If deployed in clinical waiting rooms, emergency departments, and inpatient units, the technology could alert clinical staff to sudden increases in respiratory activity, assisting with infection control protocols and resource allocation.
