testDesign Decentralized Controller for a Group of Cooperative Robot to Pushing a Box in Presence of Network Constraints

The problem of pushing objects by a group of cooperative robots has many applications on land and sea level and due to its importance, it has become a standard problem for evaluating various theories of robot cooperation. In this case, each robot produces distributed control force to push the object in the desired direction. The proposed methods for distributed control of an object on a time-varying path require information about the position of the robots relative to the object. The problem of the lack of sufficient knowledge of each robot of how the robots are positioned relative to the body can be solved by proposing a consensus issue on positional moments. In this case, the robots must reach a consensus on these moments by exchanging information through the communication network between them. The effect of communication network between robots on the process of reaching consensus and the effect of delay in consensus on the results of control of object on the desired path is the subject of this article. In this paper, the appropriate control law for achieving consensus in the absence of full connection between all bots, delay and the probability of data loss in the communication network is presented. The maximum allowable network delay is also specified to prevent the instability of object motion control. The simulation results show the capability of the proposed method for controlling the velocity of the object on the desired variable path and show the effect of network constraints on the performance of the controller.