Nowadays more and more automotive groups are working on developing driverless, autonomous cars and vehicles. One of these efforts is the SARTRE (Safe Road Trains for the Environment) project by the Swedish vehicle manufacturer Volvo. These convoys run on ordinary motorways. The first member of the convoy, the leader truck is human-driven. The truck is equipped with various sensors, whose data is transmitted towards the remainder members of the convoy. The followers are able to track the leader’s path based on his sensory data. Great advantage of such a convoy is that it relieves partly car and truck drivers and motorways, because the safety gap can be reduced due to the shorter reaction time robots have. Smaller following distance causes beneficial drag coefficient, which results in great savings on fuel expenses, not to mention the reduction of the environmental impact.The goal of my thesis is to implement the formation control of a convoy consisting autonomous mobile robots on the University’s test table. Based on the Volvo’s project I decided to prepare the followers to be able to switch roles (order) in the convoy on the fly.The system introduced above requires the realization and harmonization of various subsystems. At least 3 directable mobile robots are indispensable, and they can’t be controlled without reliable communication or without fast and accurate image processing module. The lead robot tracks a predefined track, which is a priori unknown for the followers: they are informed about the leader’s motion only from the online visual information of the camera.The emphasis of this thesis is on formation control of the followers and the framework, which facilitates the operation of the whole system. The main program is written is MATLAB, towards which the main expectation was to implement a complete simulation environment, which is capable of computing the actuator signal and modelling the behaviour of the agents. My task was also to investigate, and when possible, realize the convoy on hardware robots.