Smart Upper Prosthetic Limb

(State-level Robotics Competition)
Competition
GUJCOST ROBOFEST 2.0
Year
2022
Award
Winner of the GUJCOST ROBOFEST 2.0 in the Prosthetic Limb Category, Recieved 650K Rupees.
Summary
Led a team at Birla Vishvakarma Mahavidyalaya to develop an innovative smart prosthesis with remote sensors, integrating IoT, AI, and neural networks. Designed the prosthetic, manufactured it, and created a custom PCB for the prosthetic hand. The project aimed to provide users with full control, cloud connectivity for data storage and analysis, and a human-like experience with the sense of touch. The prototype features high-torque motors for stability, a user-friendly design, and adaptive learning capabilities for real-time muscle command response.
Goal
To build a smart AI prosthesis with IoT capabilities that offers users complete control over their limbs, connects to the cloud for data storage and analysis, and feels like an integral part of themselves, thereby closing the loop between human and prosthetic limbs.
Objectives
  • Demonstrate a forehand capable of independently actuating four fingers and a thumb, with the ability to pick up at least two different objects, showing repeated functionality without the use of ready-made kits or chassis​.
  • Pick up and hold a 400 mL glass filled with water for one minute without human intervention or spilling, demonstrating the limb's ability to grip objects securely and manage complex tasks autonomously​.

Making Process

1. Mechanical Design and Fabrication

Created a CAD model with Autodesk Fusion360, optimizing for durability, ease of repair, and weight reduction. Designed human-like fingers and a robust thumb with increased degrees of freedom.

ABS (Acrylonitrile Butadiene Styrene) was used for its durability and lightweight properties, resulting in a 22% volume reduction from the initial Proof of Concept phase. Parts were 3D printed and fabricated in the institute’s central workshop, ensuring precision and customization.

2. Electronics and Sensor Integration

Integration of Arduino Nano RP2040 Connect, sEMG (Surface Electromyography) sensors, a 16-channel servo motor controller, high-torque servo motors, tactile sensors, and haptic motors. Strategic positioning of sEMG electrodes to detect muscle activity and transmit signals to the Arduino for processing. Computations, sensory data collection, and motor commands managed using Arduino Nano RP2040, selected for its inbuilt Wi-Fi and Bluetooth capabilities.

3. Software and Neural Network Implementation

Development of a custom neural network algorithm in C++ to process muscle signals and control finger actuation, utilizing a dataset of approximately 2200 data points to distinguish between active and passive arm positions. Creation of a user-friendly web application for testing and controlling the prosthesis, allowing individual or combined finger actuation. IoT enabled for data collection, performance monitoring, and potential over-the-air updates to enhance the neural network and functionality.

Achievements

Team BVM achieved significant recognition by winning the state-level robotics competition, GUJCOST ROBOFEST 2.0, organized by the Government of Gujarat, India. The competition highlighted the team's innovation and technical prowess in developing a sophisticated prosthetic limb. The team was awarded a total prize of 650,000 INR, equivalent to approximately 8,800 USD.

Other projects

Thank you for visiting!

Template by Esme Bonnell. Edited by Jay Goswami