Vibration-Adaptive Velocity Control for Mobile Robots Using Fuzzy Logic: Online IMU-Based Approach

Abstract

This thesis presents the design, implementation, and experimental validation of a vibration-adaptive velocity control system for a teleoperated mobile robot. The system is built within a shared-control architecture in which a human operator commands the trajectory of the robot through a wireless game controller and a Mamdani fuzzy logic controller modulates the magnitude of the velocity command in response to vibration measured by an onboard inertial measurement unit. The same scalar scaling factor is applied to both the linear and angular components of the velocity command, which yields trajectory-independent modulation as a structural property of the architecture. The system is implemented as a distributed ROS 2 application across three networked hosts, an iRobot Create 3 mobile robot, a Raspberry Pi 4 companion computer, and a development laptop, using an MPU-9250 inertial measurement unit as the sole sensing modality. The control pipeline comprises four custom nodes performing IMU acquisition, vibration feature extraction, fuzzy inference and command fusion. The vibration feature is the standard deviation of the acceleration magnitude over a sliding window at 100 Hz; the fuzzy controller is a three-rule Mamdani system with centroid defuzzification. Experimental validation across nine runs in three conditions demonstrates that the controller reduces the executed velocity to 0.66 ± 0.03 of the operator’s commanded velocity during disturbance, while the operator’s commanded velocity remains unchanged. The modulation acts identically on linear and angular motion. The experimental disturbance is operator-applied chassis vibration rather than terrain-induced vibration; demonstration of the controller under real terrain conditions is identified as the principal direction for future work. The work contributes a transparent, interpretable, training-data-free implementation of shared-control velocity modulation on consumer-grade mobile robotic hardware. All the codes used for the purpose of this thesis and that are detailed throughout this document are available online at: https://github.com/paulaep/Vibration_Adaptive_Velocity_Control_for_Mobile_Robots_Using_Fuzzy_Logic.git