In this paper we use the vibrating potential method 2) (VPM) in an attempt to solve the obstacle avoidance problem. The paper reports swarm behavior for finding suitable avoidance that is derived from the VPM through self-organization. In the VPM, AGV consists of a group of units that has elementary sensor, action, controller, and communication functions. The communication media is through wave propagation on the vibrating potential field. The wave motion propagates from each of these units, and informs the unit's status to other units. From this wave motion, the AGVs obtain information about their environments and are thus able to choose appropriate behavior (i.e., obstacle avoidance and goal pursuit.) This system is advantageous to users because the AGVs do not require programming oriented to a specific problem. We provide illustrations in the form of computer simulations. These show AGVs as they pursue goals, avoid obstacles, and as they leave and rejoin a group.