Mobility is the character or property of an object that has the capacity or the ability to move or be moved . It characterizes all things that can rapidly change shape, appearance or is unstable, variable or fluctuate in time and space. The introduction of mobility in networks is a milestone in the description and implementation of communication systems that must take into account the mobility of one or more entities that comprise this system. The evolution of existing communication paradigms shows that it is increasingly difficult to design a communication system integrating the end user and not to mention mobility. The increasing use of mobile and wireless networks is a perfect example of this evolution. If we consider that mobility exists in both physical and networks logical communication in these networks should reflect this mobility. Communication protocols must take into account the instability, variability and changes imposed by the mobility of network components. Many communication protocols, especially the protocols for ad hoc networks, were designed and evaluated according to their resistance and mobility. Corson and Macker highlight the negative aspect of mobility in RFC 2501 and even suggest that resistance against mobility should be a criteria of evaluation of routing protocols in mobile ad hoc networks (MANET). In the early 2000, the research shows that mobility is not all a drawback by pointing out that mobility can increase the capacity, security and covered areas. These results raise two important points: (i) The contribution of mobility described in these articles are considered at application-level if we refer to the OSI model and (ii) mobility is undergone by the users or network entities. The two points raised above will be addressed from different point of view. Regarding the first point the results I show here consider two aspects. The first issue focuses on the benefits of mobility at the application level and more specifically for coverage in wireless sensor networks. The second aspect considers the use of mobility as a network primitive and uses it to increase the performance of the protocol stack. In these two contributions, unlike the majority of the work proposed in the literature, mobility is CONTROLLED. This property allows one to control or anticipate instability, variability and fluctuations in the network but also to modify the topology and coverage in wireless sensor networks. The use of controlled mobility at applications level (for wireless sensor networks) and to control of quality of service in protocol stack (for ad hoc networks) are two contributions which are discussed in this thesis. In this context, algorithms exploiting controlled mobility are designed and properties of the algorithms are proven. In addition, implementations on mobile robots platforms show the feasibility of these algorithms in a real environment.
