Contrôle de la formation de plusieurs robots mobiles à roues non holonomes

Abstract

Formation control of wheeled nonholonomic mobile robots has advanced significantly in the past few decades, and currently is regarded as a crucial research subject in the domains of multi-agent systems and robotics. This resulted from its potential in a wide range of real-world applications, including search and rescue operations, exploration and large object transportation. Where it has be proven that it can be more efficient in achieving such complex missions compared to using a single robot, as it allows for parallel execution of tasks and increased overall system capabilities. However, formation control of nonholonomic wheeled mobile robots can face several challenges. For example, the nonholonomic constraints that restricts the robot's motion, which require careful consideration in the modeling phase and during the control design. As well as, the presence of uncertainties and disturbances in the robot's dynamics which require the design of robust and adaptive control strategies. In this thesis, the formation control problem is investigated by proposing three different control approaches. Firstly, we explore the leader-follower formation strategy via fuzzy fractional integral sliding mode control. This control scheme enables the follower robots to accurately track the leader and achieve the desired formation pattern, despite the presence of external disturbances and uncertainties. Secondly, an adaptive distributed fractional fast terminal sliding mode control is introduced. This controller aims to accomplish a rapid and finite-time convergence of the robot towards the desired formation. The controller is developed using the consensus approach, which make it suitable for large multi robot systems as it requires less communication links between the robots. Lastly, a discrete predictive sliding mode control is developed for formation control of nonholonomic robots. Such control synthesis can lead to a well accomplished formation with a robust, chattering free and constrained control laws. To demonstrate the effectiveness and efficiency of the proposed controllers, comparative studies are conducted. The results highlight improved formation tracking, robustness to uncertainties and disturbances, and ability to achieve complex formation patterns.