Control of a Mobile Robot for Figure Recognition Applying Lambda Tuning and Machine Vision Techniques

The complexity of the environment in which a mobile robot moves are a fundamental factor for different navigation failures, even invalidating the objective. This paper approaches the design and implementation of a system for the control of a mobile robot provided by computer vision techniques, which is able to verify the direction and tracking of a target point in a specific environment. This system integrates the Lambda tuning method to analyze and design proportional integral derivative (PID) controllers and ensures robustness, stability and non-oscillating response. The images used were subjected to pre-processing using segmentation, thresholding and morphological transformation algorithms in order to improve their quality and simplify the search of figures. Consequently, the lambda method resulted in the resolution of oscillation problems allowing the user to acquire control parameters easily and with high performance. Python and Arduino perform the simulations and experimental tests of the system, verifying that the tool complies with reaching the detection of the desired figure.