This paper deals with a multi-mode control architecture for robot navigation and obstacle avoidance. It presents an adaptive and flexible algorithm of control which guarantees the stability and the smoothness of mobile robot navigation dealing with unexpected events. Moreover, the proposed Planning and Re-Planning (PRP) algorithm combine the two schools of thought, the one based on the path planning to avoid obstacles and reach the target, described as cognitive, and the second using the reactive algorithms. In fact the mix of these two approaches allows us to develop a very reliable algorithm. It provides us a scalable mobile robot navigation and obstacle avoidance, with less processing. It is accomplished by making an initial path planning, then to resolve the problem of unexpected static or dynamic obstacles while tracking the trajectory. A system of hierarchical action selection allows us to switch to a reactive avoidance, then to re-plan a new and safe trajectory to reach the target. A large number of simulations in different environments are performed to show the efficiency of the proposed PRP algorithm.