Counter-Rotating Hoop Stabilizer and SVR Control for Two-Wheels Vehicle Applications

Abstract

The matter of stabilization has always attracted attention from both academia and industry. In the case of bikes, kickboards, and scooters, simple stabilizer designs have been explored but are rarely applied in commercial products. In this research, a small-scale, customizable, two-wheels design, is proposed and connected to a retroactive closed-loop control unit for the automatic correction of the destabilization during the motion. The design is based on two counter-rotating wheels, spun into motion at the beginning of the ride and controlled by an Inertial Measurement Unit (IMU) sensor. The two DC motors controlling the spinning of the balancing wheels are adjusted by means of Pulse Width Modulation (PWM) input in the 0 ~ 255 PWM range. The control is based on an ARDUINO Uno Rev3 microcontroller and on a Support Vector Regression (SVR) model coupled with a Radial Basis Function (RBF) kernel. If an angular deviation outside the user-defined range is detected by the IMU sensor, the trained SVR-RBF model predicts the required PWM value to reestablish equilibrium and sends the signals to one or both DC motors. The proposed architecture was trained and validated in a ±21º range, resulting in a 100% correction accuracy up to a ±23º range, whereas, for greater angles up to ±30º, a drop in performances was observed. In addition to that, when a random acceleration in the ±6°/s 2 range was applied, the proposed design showed a remarkable capability of predicting the right PWM values, for both reaction wheels, capable of reestablishing equilibrium in the system within an average intervention time equal to 1.28s.

Analysis of Using Machine Learning Techniques for Estimating Solar Panel Performance in Edge Sensor Devices

“…Paper [11] proposes a real-time non-intrusive load classification (RT-NILC) IoT-based system with an Arduino-based data acquisition system. In [12], a small-scale two-wheel system connected to a control unit is developed using an ARDUINO Uno Rev3 microcontroller and a Support Vector Regression (SVR) model.…”