Graduate Research Assistant
Dr. Tejaswi Gowda, Arizona State University
Embedded Files
Mesquite is an open-source, low-cost motion capture system built around a distributed network of 15 wearable IMU sensor pods and a hip-worn smartphone for spatial anchoring. A Raspberry Pi Zero W serves as the central hub—aggregating orientation data from the pods, synchronizing it with the phone’s WebXR SLAM world frame, and streaming the reconstructed skeleton to any WebSerial-compatible browser in real time
Modular Hardware Design: Each pod integrates a 9-axis ICM20948 IMU with an ESP32 microcontroller running dual-core FreeRTOS. Pods sample at 1000 Hz and output fused quaternions at 100 Hz, powered by 400 mAh Li-Po batteries for 4–5 hours of continuous operation.
Robust Wireless Network: A dedicated Wi-Fi router connects pods, smartphone, and hub—avoiding Bluetooth congestion and enabling up to 15 sensors with <15 ms end-to-end latency and 99.7% packet delivery.
Web-Based Visualization & Recording: A Three.js-based web interface offers real-time 3D rendering, calibration tools, and export to BVH/CSV formats, accessible on desktop or mobile without additional software.
Benchmarking against an OptiTrack optical system demonstrates angular accuracy within 2–5° for most joints and position drift under 2% over five minutes, while maintaining steady 32 FPS capture . This professional-grade performance is delivered at approximately 5% of the cost of commercial alternatives, confirming Mesquite’s suitability for research, animation, and clinical applications
In above mentioned graphs the blue colour is for mesquite and red color is for optitrack.
This project has been approved in Foundation of Open Source Software United. 👉 Link
Video2.mp4Mesquite’s affordability and portability enable its use in biomechanics and rehabilitation studies, indie game animation, healthcare monitoring, sports performance analysis, and educational settings. The system’s open-source design invites community extension—ranging from hand-tracking modules to cloud-based analytics—making high-fidelity motion capture broadly accessible
mvot.mp4All the resources of the mesquite are available on the Github repository mentioned below. Here in the video the orange skeleton is existing system optitrack and the blue skeleton is for mesquite.
Page updated
Report abuse