{"id":711,"date":"2026-09-03T23:58:25","date_gmt":"2026-09-03T23:58:25","guid":{"rendered":"https:\/\/deeptech.egr.msu.edu\/?page_id=711"},"modified":"2026-09-04T17:32:43","modified_gmt":"2026-09-04T17:32:43","slug":"research-2026","status":"publish","type":"page","link":"https:\/\/deeptech.egr.msu.edu\/?page_id=711","title":{"rendered":"Research"},"content":{"rendered":"\n

Research<\/p>

Research that works in the wild<\/h1>

We use acoustics, sensing, and intelligent systems to understand machines, support decisions, and close loops.<\/p><\/div>

\"A
Acoustic diagnostics outside the lab<\/figcaption><\/figure><\/div><\/div>\n\n\n\n

Acoustic intelligence and diagnostics<\/p>

Hear the failure earlier<\/h2>

A motor, fan, tire, tool, or pump changes its sound as its condition changes. We use ordinary microphones and machine learning to find those changes before a failure becomes obvious.<\/p>

Training a monitor usually means recording many failures, which is slow and expensive when the failures are rare. We work in the other direction, building models that hold up on short recordings, small sample counts, and narrow operating ranges.<\/p>

Across devices<\/h3>

A model should not stop working because the microphone changed. Current work corrects differences among phones and other commodity sensors.<\/p><\/div>

Across time and place<\/h3>

Recordings collected in one session are rarely the whole story. We test across locations, operating conditions, and stages of wear.<\/p><\/div>

With honest uncertainty<\/h3>

A useful system should say how certain it is. Conformal methods add defensible intervals to audio quality predictions.<\/p><\/div>

Current team:<\/strong> Kehinde Elelu, Tashfain Ahmed, Kyle Foster, Paremeswar Nair, Tejas Agrawal, Mohammadali Saffary, and Joshua Siegel.<\/p><\/div>\n\n\n\n

Embedded expertise<\/h2>

We build systems that carry technical judgment into the moment it is needed. They may recommend a maintenance action, flag a risky condition, change how a machine operates, or ask a person to review the evidence.<\/p>

The form follows the job: a predictive model, a generative system, a structured workflow, a controller, or something simpler.<\/p><\/div>\n\n\n\n

Robotics, automation, and manufacturing<\/h2>

We put perception and intelligence on real equipment. What can the machine observe? What does that observation mean? What should happen next?<\/p>

\"A
Testing acoustic diagnostics on an air compressor<\/figcaption><\/figure>
\"Two
Autonomous systems in the lab<\/figcaption><\/figure><\/div>

Connected vehicles and mobility<\/h2>

Vehicles are machines, sensor networks, and participants in a larger transportation system. We use their existing signals to study condition, context, safety, and interaction. Earlier projects explored smartphone-based diagnostics, secure vehicle data, physically adversarial situations, and edge intelligence. Those results still shape work on other machines today.<\/p><\/div>\n\n\n\n

From signal to action<\/h2>

The research is not finished at a good classifier. We ask what decision the result supports, who should make it, and what evidence they need.<\/p>

One \u00b7 Problem<\/h3>

Start with the failure, decision, or constraint.<\/p><\/div>

Two \u00b7 Signal<\/h3>

Find the smallest useful observation.<\/p><\/div>

Three \u00b7 System<\/h3>

Build the model, workflow, or controller.<\/p><\/div>

Four \u00b7 Validation<\/h3>

Test it on the real machine, with the variation the real world brings.<\/p><\/div><\/div>\n\n\n\n

Selected 2026 work<\/h2>