University of Miami's AI Cycling Coach Reads Your Effort

A University of Miami lab is testing an AI cycling coach that watches a rider's power, cadence and heart rate and speaks up only when encouragement may help.

•November 13, 2026
University of Miami's AI Cycling Coach Reads Your Effort Research

Summary: University of Miami researchers Anna Queiroz and Meshak Cole are piloting an AI cycling coach that reads power, cadence and heart rate on a stationary bike and decides when to give a cue. A study compares AI coaching with no coaching.

aicoachingsportstechumiamihumanperformance

A good coach watches more than the clock. A University of Miami team is building an AI cycling coach that does the same.

The prototype comes from Anna Queiroz and Meshak Cole. Queiroz is an associate professor in the School of Communication’s Department of Interactive Media.

Cole is an M.F.A. graduate student. Their system joins physiological sensing, motivational psychology and artificial intelligence.

A Coach That Listens to the Rider

The setup is simple. A rider sits on a stationary bike with earbuds, a chest strap and sensor-enabled pedals, all linked to a phone app.

The system tracks power, cadence and heart rate. It uses those signals to decide when encouragement may help and what kind of message to deliver.

The app also checks every few seconds whether a cue is due. It picks one of several coaching approaches and decides how long to wait before speaking.

If a rider’s power sags below baseline for about half a minute, the system can send a message. It is meant to reframe the struggle.

The text appears as a subtitle on a virtual handlebar display. It is also converted to speech that the rider hears through earbuds.

Cole said the idea came from his own workouts. “It makes personalized coaching more accessible by delivering adaptive guidance tailored to each individual’s needs, abilities and progress,” he said.

Each rider starts with a short pretest about values, interests and some personality traits. “Depending on performance by each individual, the algorithm decides which cues to use,” Cole said.

Riders pedal in sessions of about 20 minutes. During a test they may hear a pacing cue such as, “You are finding a powerful rhythm and committing to it, that’s the smart way to build your effort.”

Or they might get a breathing cue. “Drive your breath from deep in your belly, letting each forceful exhale clear the way for fresh energy.”

Riders can also ask for a boost. The app decides how to answer that request.

That responsiveness is the point. Queiroz, who directs the Future Realities Research Lab, contrasts it with popular home-fitness products.

“What is available right now, such as the Peloton model, provides a preprogrammed experience and prerecorded classes that are not personalized,” she said. “This is different.”

She said the program is meant to “bring a customized coaching experience without the cost of having a personal trainer or coach.”

A Pilot Study Under Way

The pilot began last spring semester. Queiroz hopes to monitor about 90 participants by the end of this spring.

Participants are placed in one of three groups. One group gets AI coaching, one gets none, and one gets some coaching before the ride.

That design lets the team compare progress across groups. Cole said the hope is that riders in the AI group will improve their performance.

The study is early, and no results have been reported. A few dozen riders cannot settle whether adaptive coaching beats a human trainer or a good playlist.

What it can show is whether the system’s timing helps. That is the hard part of any coaching tool.

A Miami Research Thread

The project fits a pattern at the University of Miami, where researchers are applying AI to human problems. MAIN has covered a UM study using AI to predict Parkinson’s decline early.

It also covered a UM professor using AI to invent new languages. The cycling project is another case of AI built around how people feel and respond.

Miami’s wellness scene is thinking along the same lines. MAIN reported on an AI wellness masterclass bringing AI and human performance together.

Personalized coaching without a personal trainer’s price could matter far beyond cycling. Queiroz and Cole are testing whether AI can supply the encouragement as well as the data.

Sources:

  1. University of Miami News: AI-driven cycling system adapts coaching to a rider’s effort
  2. MAIN: UM Study Uses AI to Predict Parkinson’s Decline Early
  3. MAIN: This UM Professor Uses AI to Invent New Languages
  4. MAIN: AI Wellness Masterclass Brings AI and Human Performance Together