OPTICAL SENSOR-BASED BALL TRACKING AND CENTRALIZATION USING A PID-CONTROLLED STEWART PLATFORM
This research work entails optimizing a model of a Stewart Platform system. The proposed system is designed to autonomously centralize a rolling ball on its top plate using a multiplexed array of optical sensors and a PID control algorithm. The Stewart Platform, a parallel manipulator, comprises a fixed base and a movable top plate connected by six servo-actuated extensible legs. By varying the lengths of these legs, the platform achieves roll (rotation about the x-axis) and pitch (rotation about the y-axis) motions to influence the movement of the ball. Conventional ball-balancing systems rely on resistive touch surfaces, inertial sensors (gyroscopes and accelerometers), or vision-based setups involving external computers. The problem that is inherent with these existing designs is that they require enormous computational power for floating point calculations that come with image processing and reading from resistive panels. Our proposed approach deviates from these existing designs by introducing a simple sensing technique. This technique uses a two-dimensional grid of optical sensors beneath a transparent top plate. As the ball moves, it obstructs light received by specific sensors, allowing the controller to determine its position through binary (digital) readings�eliminating the need for computationally expensive. This significantly reduces processing load and allows the system to operate on low-cost microcontrollers. A PID controller, tuned via the Ziegler-Nichols method, regulates the platform's motion to minimize positional error. To achieve this work the complete system is modeled and simulated using MATLAB Simulink's Multibody environment. Simulation results show that the platform effectively stabilizes the ball for disturbances like base tilts up to 12 degrees. Furthermore, performance comparison with previous work indicates a 5-second improvement in settling time, highlighting the efficiency and responsiveness of the proposed design.
Authors : Jival, A. and Nwaobasi, C.C.
Category : Open Access Volume (Issue) : 11(2) Date Uploaded : 4th August 2025