Date of Defense

17-11-2025 10:00 AM

Location

F1-1043

Document Type

Thesis Defense

Degree Name

Master of Science in Mechanical Engineering (MSME)

College

COE

Department

Mechanical and Aerospace Engineering

First Advisor

Dr. Farag Omar

Keywords

Real-Time Robotics, Trajectory Prediction, Air Hockey Robot.

Abstract

This thesis presents the development of an intelligent air hockey robot that combines precise mechanical design, computer vision, and adaptive control within an AI-based hierarchical decision architecture. The system integrates synchronized stepper motors, high-speed image processing, and a real-time decision framework to achieve competitive gameplay performance. The robot detects the puck using adaptive HSV color segmentation, supported by dynamic calibration that maintains accuracy under different lighting conditions. A two-stage trajectory prediction model, based on exponential decay velocity estimation, enables anticipation of puck motion and improves response time during fast gameplay.

At the decision level, a fuzzy-logic supervisor governs the robot’s strategic behavior, determining when to defend, attack, leave, or return to home position according to real-time game conditions such as puck position, speed, and predicted path. The lower control layers execute these decisions through precise motion coordination using Arduino-based stepper motor control. Experimental evaluation confirms significant improvement in interception accuracy, stability, and adaptability compared with conventional reactive systems.

The proposed adaptive fuzzy-hierarchical framework demonstrates how intelligent decision-making and predictive control can enhance robotic performance in highly dynamic environments. Its structure can be extended beyond entertainment robotics to other real-time applications that require perception, prediction, and adaptive response.

Share

COinS
 
Nov 17th, 10:00 AM

ENHANCED AIR HOCKEY ROBOT PERFORMANCE THROUGH ADAPTIVE CONTROL ALGORITHM USING AI-BASED HIERARCHICAL DECISION ARCHITECTURE

F1-1043

This thesis presents the development of an intelligent air hockey robot that combines precise mechanical design, computer vision, and adaptive control within an AI-based hierarchical decision architecture. The system integrates synchronized stepper motors, high-speed image processing, and a real-time decision framework to achieve competitive gameplay performance. The robot detects the puck using adaptive HSV color segmentation, supported by dynamic calibration that maintains accuracy under different lighting conditions. A two-stage trajectory prediction model, based on exponential decay velocity estimation, enables anticipation of puck motion and improves response time during fast gameplay.

At the decision level, a fuzzy-logic supervisor governs the robot’s strategic behavior, determining when to defend, attack, leave, or return to home position according to real-time game conditions such as puck position, speed, and predicted path. The lower control layers execute these decisions through precise motion coordination using Arduino-based stepper motor control. Experimental evaluation confirms significant improvement in interception accuracy, stability, and adaptability compared with conventional reactive systems.

The proposed adaptive fuzzy-hierarchical framework demonstrates how intelligent decision-making and predictive control can enhance robotic performance in highly dynamic environments. Its structure can be extended beyond entertainment robotics to other real-time applications that require perception, prediction, and adaptive response.