Raeyaan Muppaneni
Built a low-cost bionic robotic arm controlled by muscle signals using AI, achieving 93% accuracy. Presented at IEEE WcCST-2026 in India.
Where Raeyaan Started
His Background
- • 11th grader at Irvington High School, Fremont
- • Interested in AI, robotics, and engineering
- • Experience with Python and Java
- • Active in robotics camp and math competitions (AMC 12)
- • No prior research or publication experience
His Goals
- • Publish research in a major journal
- • Build something at the intersection of AI and robotics
- • Win science fair competitions
- • Build a competitive edge for college admissions
- • Ultimately: "Winning ISEF"
The Problem He Wanted to Solve
Over 804.5 million people globally are affected by motor impairments. Conventional prosthetics are expensive, inaccessible, and lack real-time responsiveness. Raeyaan wanted to build an affordable, AI-powered robotic arm that anyone could use.
The Research
Working with YRI mentors, Raeyaan built a complete system from scratch: a low-cost bionic robotic arm that reads muscle signals (EMG) from the forearm, classifies gestures using AI, and moves a servo-driven arm in real time. The entire system runs on a Raspberry Pi, making it affordable and accessible.
Real-Time Control of a Low-Cost Robotic Arm Using EMG Signal Classification by AI-Based Machine Learning on Raspberry Pi
Conventional prosthetics are expensive and lack real-time adaptability for people with motor dysfunction
EMG signals captured via MyoWare sensor, transmitted wirelessly to Raspberry Pi via ESP32 and MQTT protocol
KNN, Random Forest, Multi-Layer Perceptron, SVM on three feature pipelines (raw EMG, TSfresh, MFCC)
Random Forest on raw EMG envelopes achieved 93% real-time accuracy on Raspberry Pi deployment
Hardware + Software Integration
What makes Raeyaan's project exceptional is that he didn't just train a model — he built the entire hardware system. The robotic arm uses an ESP32 microcontroller, MyoWare EMG sensors, and a Raspberry Pi 4 running real-time inference. The arm recognizes four gestures — hand lift, hand twist, wrist lift, and neutral — and responds with proportional servo movement. Total hardware cost: a fraction of commercial prosthetics.
ML Models
Feature Pipelines
Real-Time Accuracy
Gestures Recognized
The Outcome
Presented at World Conference on Computational Science and Technology
IEEE WcCST-2026, Chandigarh University, India
March 26-27, 2026
850
IEEE Computational Intelligence Society
No research experience, interested in robotics and AI but no clear project direction
Built a working bionic arm, published at IEEE conference, 93% real-time ML accuracy on edge hardware
The Bigger Picture
People globally affected by motor impairments who could benefit from affordable assistive tech
Real-time classification accuracy on a $35 Raspberry Pi — no expensive hardware needed
Published and presented at an international IEEE conference as a high school junior
Ready to Start Your Research Journey?
Join the YRI Fellowship and work with expert mentors to conduct original research, get published, and build a profile that stands out.
Apply Now