This work presents the development of a discrete speed controller for an autonomous vehicle for inspection of flexible ducts. To this end, the techniques of LQG control (Linear Quadratic Gaussian) with Multi-Sensor Fusion using Kalman Filtering were used. The vehicle should reach the desired speed with minimal overshoot in order to prioritize a visual inspection of good quality. To get the velocity information, the vehicle is instrumented with a pressure sensor and two odometers. The proposed controller is expected to ensure a greater robustness to the control of the vehicle due to fusion of information of sensors of different natures. The project was designed and simulated in Matlab/Simulink, and shows step-by-step development of this controller.