Blog
Blog

25th March 2025


As humanoid robots become more advanced, motion capture is playing a crucial role in teaching them how to move like humans. This article explores how Xsens technology is used to capture precise human motion data, enabling robots to learn, adapt, and perform complex tasks safely and efficiently through AI and machine learning.

In recent years, the convergence of robotics and AI has accelerated dramatically, paving the way for the rise of humanoid robots—machines designed to resemble humans and perform tasks that enhance our daily lives. Engineers are uncovering new ways in which robots can support human activity, from domestic chores to rehabilitative care and even in the workplace.

The field of humanoid robotics is advancing rapidly, with applications ranging from companionship and hospitality to high-risk environments too dangerous for human workers. “The main benefit is to increase productivity and safety,” says Dennis Kloppenburg, Product Manager at Xsens. “If a robot can take on tasks in hazardous environments, the risk to human workers is greatly reduced—and unlike humans, robots can operate around the clock, further boosting efficiency.”

These applications are not limited to manual labor. “We already have robots that vacuum our floors, but imagine if that same robot could assist with repetitive tasks in an office setting,” Dennis adds. While many robots today are specialized in performing a single task, recent advancements have made versatility and adaptability more achievable than ever.

Designing robots in humanoid form enables engineers to better explore the boundaries of robotic functionality, allowing for more natural, efficient, and flexible movement.

“The rapid progress in AI has propelled this industry forward,” Dennis continues. “Robots aren’t just programmed for a single task anymore—they can now learn to adapt to different scenarios using AI.” This progress also allows robots to perform tasks requiring fine motor skills, such as grasping and manipulating objects—skills that can only be developed through AI and machine learning, supported by motion capture technology.

Training a robot

“The top 10 humanoid robot companies use Xsens motion capture technology to train their robots,” Dennis notes. “With Xsens, you can measure, observe, and analyze human movement—and pass that knowledge on to robots.”

Here’s how the process works:

  • Select your subjects – Capturing motion data from a diverse group of individuals—different heights, sizes, and movement styles—helps create a robust model of what’s considered ‘normal’ human movement.
  • Record the data – Consider the tasks the robot will be expected to perform. Human subjects go about daily activities, such as working in a factory or picking up objects, to generate hours of motion data for the robot to learn from.
  • Apply and adapt – With this data, robots mimic these actions, using AI to generalize and apply what they’ve learned across a variety of real-world scenarios.

The role of training data

AI allows robots to learn by mimicking data gathered from human subjects. “AI uses Xsens motion data as a foundation for learning,” Dennis explains. Motion trackers placed on the body collect information on joint position, velocity, and rotation—critical inputs for replicating lifelike movement.

“Xsens is the gold standard for scalable data acquisition,” says Dennis. “Known for its high accuracy, low latency, and ability to scale, it allows seamless capture of motion data from hundreds of workers performing tasks for robots to learn from. This powerful combination enables robot training up to six times faster than alternative tools.”

Inertial measurement units (IMUs) in the Xsens system ensure high-fidelity motion data, capturing even the subtlest movements. Unlike optical systems, Xsens suits require no camera setup, allowing human subjects to move freely in any environment. “One person’s data can train multiple robots,” Dennis notes. Xsens systems are also immune to magnetic interference—critical in environments where electromagnetic fields are common.

As the conversation shifts from building humanoid robots to maximizing their capabilities, the question becomes: how do we get there? For Dennis, the answer is clear: “With our data, there’s no debate—Xsens is the unmatched solution for scalable humanoid robotics.”

Learn more about Xsens motion capture for humanoid robots

Related articles