Abstract
Explores the mechanical and computational integration of computer vision head-tracking with recycled electronic components to create reactive social robot kinetic sculptures. Investigates human-robot interaction response times and motor control smoothing algorithms.
Research Problem & Context
Commercial educational robots are often expensive and treat hardware as black boxes. Developing responsive, bio-inspired robotic movement using recycled electronic components requires efficient calibration of non-standard actuators.
Research Question
How can low-latency sensor filtering algorithms stabilize servo jitter when tracking human faces using uncalibrated optical sensors from salvaged electronics?
Experimental Methodology
Designed an articulated kinetic robot head constructed from repurposed electronics. Implemented an exponential moving average (EMA) filter on optical centroid coordinates calculated from a camera stream, transmitting normalized PWM signals to pan-tilt servo motors.
Hardware Architecture
- ESP32 Microcontroller
- Micro Pan-Tilt Servo Gimbal
- Optical Flow & Ambient Light Sensor Array
- Recycled Electronic Scrap Chassis & Enclosure
Software & Toolchain Stack
- C++ / Arduino IDE
- Python OpenCV Face-Landmark Tracking
- Serial Telemetry Streamer
- Custom Kinematic Smoothing Algorithms
Experimental Setup & Testing
Bench-tested tracking fidelity across angular transitions between -45° and +45° yaw. Evaluated tracking smoothness and mechanical oscillation with and without software damping.
Results & Analysis
The smoothed PID-filtered coordinate stream dampened mechanical oscillation, providing stable head tracking without high-frequency servo jitter.
Technical Limitations
Mechanical wear in salvaged gearboxes introduces small deadbands after extended continuous operation.
Primary Resources & Manuscripts
Citation Information
Cite this work
APA / Text:Alagha, A. (2024). Multi-Modal Sensory Fusion in Expressive Social Robotics (Project Report). https://abed.page/research/social-robotics-vision-fusion
BibTeX:
@techreport{alagha2024socialrobotics,
author = {Alagha, Abedalraheem},
title = {Multi-Modal Sensory Fusion in Expressive Social Robotics},
type = {Project Report},
year = {2024},
url = {https://abed.page/research/social-robotics-vision-fusion}
}