Multi-Modal Lower-Limb MoCap System for Teleoperation-Oriented Data Acquisition and Browser-based Visualization

Human beings have long sought to translate their own movements into digital commands, a pursuit that sits at the heart of modern robotics and virtual reality. To make a robot walk or a virtual avatar move with human grace, engineers need a way to capture the complex geometry of a person's legs in real time. This process, known as motion capture, traditionally relies on expensive, room-sized camera systems or heavy, full-body suits that cost thousands of dollars. For researchers and students who want to experiment with controlling robots or visualizing movement, these high-end tools create a barrier to entry. The challenge, therefore, is not just to capture motion, but to do so with a system that is affordable, reproducible, and capable of streaming data directly to a computer screen without complex setup.

A team of researchers at Nanjing University of Science and Technology has addressed this challenge by building a low-cost, wearable system designed specifically to track the movement of the lower body. Their work, detailed in a recent development report, focuses on creating a pipeline that takes raw physical data from a person's legs and turns it into a smooth, visual representation in a web browser. The system is not a final, perfect product for industrial use, but rather a proof-of-concept that demonstrates how a small laboratory can build a functional motion capture tool for around three hundred and fifty to three hundred and sixty Chinese yuan. The goal was to create a device that could stream joint angles, the tilt of the pelvis, and distance measurements to a digital monitor, allowing a user to see their own movements mirrored in real time.

The physical device resembles a lightweight, skeletal frame that straps onto the thighs and shins. Instead of using heavy industrial sensors, the researchers utilized commercially available electronic modules that are easy to source and replace. The core of the system relies on a clever use of standard servo motors, the small electric motors often found in model airplanes and robotics kits. In this setup, the motors are not used to move the legs; instead, they are turned off and used as passive sensors. As the wearer moves their hip, knee, or ankle, the joints physically rotate the motor's internal gears. The system reads the electrical signal generated by this rotation and translates it into an angle, effectively turning the motor into a ruler that measures how far a joint has bent. To track the body's orientation in space, a small inertial sensor is attached to the pelvis, measuring how the hips tilt and turn. For measuring height or distance to the ground, the team initially tested ultrasonic sensors, which work like a bat's echolocation, but found them too slow and prone to errors. They switched to a different type of sensor that uses a beam of light to measure distance instantly, which proved much more reliable for keeping up with fast movements.

All of these sensors feed data into a small computer board that acts as the system's brain. This board sends a steady stream of information over a wireless or wired connection to a computer running a simple software bridge. This bridge translates the raw numbers into a format that a web browser can understand. On the screen, a digital avatar, or "twin," mimics the wearer's movements. The software includes several layers of protection to ensure the image on the screen looks natural; it filters out tiny, meaningless shakes in the data, ignores impossible values, and smooths out the motion so the digital legs do not jitter. The researchers tested this entire chain of hardware and software by having a person stand, walk, swing their legs, and jump. The system successfully captured the motion of the hips and knees, and the digital avatar on the screen moved in sync with the person, including the rise and fall of a jump.

However, the experiment also revealed a significant physical limitation in the design. While the sensors for the hips and knees worked well, the attachment for the ankle was not rigid enough. When the person moved their foot, the strap holding the sensor shifted slightly, causing the data to become unreliable. Because the researchers were honest about this flaw, they simply locked the ankle data to a fixed position during the demonstrations to prevent the digital avatar from glitching. This decision highlights a key finding: the system is excellent for capturing large movements of the upper leg and pelvis, but it cannot yet provide precise measurements for the ankle without a stiffer mechanical design. Furthermore, the researchers were careful to note that while the system works, they have not yet compared its accuracy against high-end professional cameras or medical-grade tools. They have not proven that the numbers are perfectly correct, only that the system is stable enough to run for over thirty minutes without crashing and that it can visualize movement in real time.

The value of this work lies not in a breakthrough in measurement precision, but in the creation of a transparent, accessible path for data collection. The team demonstrated that by using simple, off-the-shelf parts and careful software design, it is possible to build a motion capture system that is affordable for students and small labs. They showed that specific engineering choices, such as using a light-based distance sensor instead of sound, and organizing the software into distinct layers that do not block each other, are critical for making these low-cost devices work reliably. The report concludes that while the current ankle design needs improvement and the system requires manual calibration for each user, the foundation is solid. It offers a working, end-to-end example of how to turn a person's walk into a digital stream, paving the way for future experiments in robot control and virtual interaction without the need for a multi-million-dollar laboratory.