Post by Gowri Priya Aji

Human-Robot Interaction (HRI) traditionally relies on resource heavy, cloud dependent computational models. For our final year engineering project, our team set out to challenge to this by building a localized, resource optimized alternative. Meet M.O.R.P.H.I.S. (Modular Optimized Robotic Platform for Human Inspired Simulation). Instead of relying on continuous internet connectivity or heavy AI frameworks, we designed an event driven, edge computed humanoid head architecture capable of real time human presence detection and physical articulation. Key Architectural Highlights: • Hardware Foundation: Adapted an open source humanoid head design, modifying the internal mechanical constraints for streamlined fabrication and component mounting. • Edge Computing: Driven locally via a Jetson Nano architecture running on an Ubuntu framework. • Signal Processing: Integrated a threshold based ambient audio sampling system to detect presence without complex, high power speech processing models. • Actuation & Sync: Utilized a PCA9685 PWM driver over I2C to precisely coordinate multiple servo actuators, achieving real-time synchronization between text-to-speech output signals and mechanical jaw/eyebrow movements. • Application: Engineered with an onboard data-handling mechanism, allowing the platform to act as a standalone, offline assistant for information kiosks and reception desks. Developing the hardware customization and software mapping from scratch forced us to balance minimal actuation with expressive feedback proving that functional HRI can be highly optimized and cost effective. An incredible engineering journey alongside: Aswin S , Febina Sara Sam and Jerin Varghese Deep gratitude to our project guide, Dr. Binson V A , for his constant technical direction and support throughout this development phase. #Robotics #Automation #EmbeddedSystems #HumanRobotInteraction #JetsonNano #EdgeComputing #HardwareEngineering #TechInnovation #OpenSourceHardware #RoboticsProject #HumanoidRobot