**2.1 Reactive approach**

Many reactive approaches have been proposed to resolve the problem of autonomous robot navigation. In the following, we mention the most important among them.


in reactive schemas, without using planning techniques, in order to consider the

To deal with uncertainty, autonomous navigation involves using systems for navigation control that must be not too computationally expensive, as this would result in a sluggish response. In this case, we choose the fuzzy logic technique because it is faster when the

In the field of multi-agent systems, several works have encouraged researchers to develop models simulating robot's behaviours in order to achieve a known target (Posadas et al., 2008). They are more flexible and fault tolerant as several simple agents are easier to handle and cheaper to build. Indeed, the term "agent" has been defined as hardware or a software system with certain properties such as autonomy, social ability, reactivity and pro-activity (Ferber, 1995). In fact, there are similarities between robot and agent because they share the same characteristics. So, the mapping from a robot to an agent seems straightforward because each robot represents a physical entity, independent from other robots, with a

In this paper, we present a reactive anticipation model based on fuzzy logic that takes into account: (1) the reactive navigation in an environment composed of local minimum and moving obstacles, and (2) the anticipation of blocking situations and the prediction of the nature, the velocity and the position of obstacles in the future. This information is used to predict conflict situations without using a motion planning method. In order to validate our work, we evaluate our approach by simulating various scenarios. We also give a comparative

In the remainder of the article, we give, in section 2, an overview of the existing approaches for conflict resolution applied to autonomous robot navigation. In section 3, we present the most used techniques to deal with the uncertainty of perception and we justify our choice of the fuzzy logic method. In section 4, we describe our model that combines reactivity with anticipation in order to deal with the problem of conflict resolution without using a motion planning. In Section 5, we present our experimental results. Finally, we conclude in section 6

When navigation occurs in an environment that is totally or partially unknown or even dynamically changing, higher degree of autonomy for a mobile robot is required. Thus, a mobile robot should be able to take decisions on-line and to minimise conflict situations in

In this context, a wide range of approaches have been adopted to suggest solutions to the problem of space conflict when a robot shares the same environment with other actors.

Many reactive approaches have been proposed to resolve the problem of autonomous robot

 Potential Fields (Huang et al., 2006; Khatib, 1986): this approach relies on creating an artificial repulsion field around obstacles and an attraction field around the goal. Vector Field Histogram (Ulrich & Borenstein, 2000): it uses a heading dependent histogram to represent the obstacle's density. So, the robot can move in the direction

where there are less obstacles in order to minimise its interaction with them.

such uncertain and dynamic environments using only sensors' limited information.

navigation. In the following, we mention the most important among them.

study between results obtained by our model and those of some other approaches.

environment's evolution in the future.

and we give perspectives for this work.

**2.1 Reactive approach** 

**2. Existing approaches for conflict resolution** 

specific task.

frequency of the environment's changes is high.


The main drawback of these strategies is that the robot gets into an infinite loop or local minimum when it is moving among multiple obstacles or in conflict situations when it shares the same resource with moving obstacles. To overcome this problem, many methods are proposed. We mention:

