Patrick Slade is an assistant professor of bioengineering at Harvard University’s John A. Paulson School of Engineering and Applied Sciences (SEAS). His research combines wearable robotics, biomechanics, sensing, and artificial intelligence to help people move more easily and safely in their daily lives. His lab develops technology that can measure human movement, adapt assistance to individual users, and support people with mobility and navigation challenges. The following Q&A is based on an interview with Slade. Edited for clarity, length, and context.
Q: How will AI improve wearable technology and assistive devices?
A: One of the things we’ve learned in wearable robotics is that just because you put a motor on a person doesn’t necessarily make it easier to move. In some cases, it can actually make movement difficult because the device is not working in harmony with the person.
We all move a little differently. People differ in strength, balance, motor control, and movement patterns, and these differences are even more important in patient populations. A device that helps one person may not help others in the same way.
AI gives us the tools to better understand those differences. Wearable sensors and machine learning models can be combined to estimate things that are difficult to measure outside the lab, such as energy expenditure, gait speed, movement asymmetries, and joint loads. You can then use that information to customize your assistance, adjusting when it’s delivered, how much it’s delivered, and what outcomes you’re looking to improve.
Q: What are the limitations of current wearables such as smartwatches?
A: Consumer wearables are great at collecting data, but some of the metrics they report can be shockingly inaccurate. Estimates of energy expenditure, or calories burned, can vary by as much as 40 to 80 percent.
Part of the problem is that many devices rely on indirect signals such as heart rate or wrist movement. These signals can be helpful, but they don’t always reflect the mechanical work your body is doing. When you walk or run, much of your energy expenditure comes from your leg muscles, not from your wrist movements.
Our approach is to use signals that are more closely related to biomechanics. for example, A smartphone that you carry in your pocket Acceleration can be measured. That information can be combined with machine learning models and established physical principles to more accurately estimate meaningful measurements such as energy consumption.
The goal is not just to collect more data. It’s about identifying the most important signals and using AI in a way that reflects how the body actually works.
Q: What are the biggest challenges in developing wearable and assistive technology?
A: Personalization is one of the core challenges of our research. Everyone has different movement patterns, muscle strength, balance, motor control, anatomy, and goals. These differences are especially important for people who have had a stroke, have knee osteoarthritis or knee pain, are elderly, and are blind or visually impaired.
Fixed controllers can help some and hinder others. Instead, we want technology to be able to learn how a particular person moves and determine what kind of support would be most helpful.
For wearable robots, that could mean adjusting when assistance is provided, the amount of force applied, and the outcome the system is trying to improve. It may be possible to adjust the device to reduce energy expenditure, increase walking speed, improve leg symmetry, or reduce stress on painful knees.
These systems are also needed to adapt to external, controlled laboratory environments. People walk at different speeds, change direction, climb stairs, encounter uneven ground, and perform activities that may not appear in the training dataset. Our existing knowledge of human movement helps make technology more robust and reliable.
Q: Can you give us a real-world example of AI-powered assistive technology?
A: One example is our work with people who have experienced stroke. Wearable sensors and AI can be used to estimate walking speed and asymmetry between paretic and non-paretic legs. This information helps personalize wearable robots and determine how the device assists people.
Another use includes people with knee osteoarthritis or knee pain. The study estimated the load passing through the knee and investigated how assistive devices can reduce it.
We are also developing navigation technology for visually impaired people. The system uses your smartphone’s camera, GPS, and motion data to identify features such as sidewalks, curbs, crosswalks, and obstacles. It then provides audio feedback through open-ear headphones to guide the user.
Across these projects, AI enables technology to interpret complex information and adapt assistance to people and environments.
Q: What does the future hold for AI-powered wearable and assistive technology?
A. I think it’s realistic that within the next few years people will be able to buy wearable devices that use AI to automatically personalize themselves. These systems can support the elderly, people with mobility impairments, workers who perform physically demanding tasks, or people who need help staying active.
An electric bicycle is a simple example. It’s not a replacement for cycling. Provides enough support to make cycling more accessible. Wearable robots could play a similar role by facilitating movement, reducing pain and fatigue, and helping people continue doing activities that are important to them.
