<![CDATA[Terrestrial testing of CubeSat Attitude Determination and Control System (ADCS) prototypes presents fundamental validation challenges because no laboratory platform perfectly replicates the free-floating microgravity condition of orbit. The empirical effect of such platform-induced dynamics on closed-loop reaction-wheel control performance, particularly for low-cost academic CubeSat programs, remains insufficiently characterized. This study aims to empirically compare two widely used low-cost test platforms a string suspension and a free bearing for a single-axis reaction-wheel ADCS prototype and to quantify how each platform’s parasitic dynamics constrain PID controller performance. A 1.5U CubeSat-class prototype with an ESP32-based controller, BNO055 IMU, and a Blynk Cloud IoT interface for real-time PID tuning was tested over five sessions (>1 hour of cumulative active testing). Performance was quantified using the Time-in-±5° metric, error standard deviation, settling time, and number of direction reversals. Quantitative results show that the string suspension yields a peak Time-in-±5° of 8.8% with error standard deviation of ±90.7–119.9°, driven by torsional-pendulum dynamics, while the free bearing yields a peak Time-in-±5° of only 1.2% with a stuck-and-jump signature characteristic of stiction. A Karnopp-friction simulation in Python reproduces a permanent steady-state error of ~5° under stiction, quantitatively validating stiction as the dominant non-linearity. The novelty of this work lies in the integrated combination of empirical multi-platform characterization, Karnopp-based stiction validation, and an open-source IoT-based PID tuning framework within a single low-cost experimental system, providing actionable guidance for academic CubeSat ADCS development under limited-facility conditions.]]>
Copyrights © 2026
