The main objective of this thesis is to validate the general hypothesis that car drivers select a mode of action (i.e., make the decision to overtake or not) and regulate it (i.e., control their automobile overtaking) by perceiving overtaking possibilities called affordances. These affordances are intrinsically scaled to the capacity for action of the vehicle driven (e.g., speed or acceleration). From the mathematical formalism of the affordance developed by Warren (1984) and the new conceptualization of the affordance provided by Fajen (2007a), it is about putting the theory of the affordances to the test of automobile overtaking through three experiments carried out on driving simulator. In a first experiment, we showed that car drivers take the decision to initiate an overtaking on the basis of an affordance defined by the ratio between the properties of the agent-environment system (minimum satisfactory velocity to succeed an overtaking (MSV)) and the agent property, i.e., the action limit (maximum speed of the vehicle driven (Vmax)). In a second experiment, we tested the preferential use of a second overtaking affordance able to specify earlier and more precisely the overtaking opportunities. We have highlighted that the decision to initiate an overtaking depended mainly on a high order affordance, the MSA/Amax, defined by the ratio between a property of the agent-environment system (minimum satisfactory acceleration to succeed an overtaking (MSA)) and the property of the agent (maximum acceleration of the vehicle driven (Amax)). In a third experiment, we studied the selection and regulation of an action mode questioning the ability of drivers, when performing a same overtaking, to use various limits of action specific for each gear ratio. According to our hypothesis, drivers are able to exploit the overtaking opportunities defined by the ratio between the properties of the agent-environment system (minimum satisfactory acceleration to succeed an overtaking (MSA)) and the agent’s properties (maximum acceleration of the vehicle driven in 4 th gear (Amax 4th) and maximum acceleration of the vehicle driven in 3rd gear (Amax 3rd)). The formalization of these possibilities for action allowed us to expand the scope of the affordance-based control (Fajen, 2007a) to the overtaking tasks which are very close of situations encountered in daily life and to take a new look at this kind of tasks. Future research will be needed to demonstrate the causal function of the overtaking affordances implemented in the process of selection and regulation of action. This step will imply to identify very precisely the optical invariant underlying the overtaking and to study how they are taken into account, especially when a competition between affordances occurs.