A Multi-Robot System Architecture for Trajectory Control of Groups of People

Abstract

How to conduct a group of humans towards a target destination by a team of robots is the key-problem discussed in the present context. A suitable multi-robot system (MRS) architecture has been investigated and implemented for guiding groups of people. The present system can be seen as guiding-tours , nevertheless further than such concept this implementation can be though, or it is closer to the model given by several dogs flocking herds of sheep, guiding them towards a target ed place. Dogs and sheep have a minimal way of explicit communication. Sheep herd’s trajectory is controlled by a team of dogs (even one). The dogs do bark and/or approach to the herd if there is any situation disordering guidance. The proposed context differs from it, since does not exist any type of explicit signal for guidance, and trajectory control is given by a way based on natural reactions of angle-velocity motions between humans and robots. An extensive theoretical description and experimental results are discussed. Recent progress in robotics and artificial intelligence has made possible to bu ild interactive mobile robots that operate highly reliably in crowded environments. There exist in the research field community several works concerned with guiding-tours, nevertheless tackling different problems and/or deploying different architectures. Few successful works concerned with guidingtours have been developed (Nourbakhsh et al., 1999; Thrapp et al., 2001; Burgard et al., 1999). Our context has some differences, such as our system is compounded by a team of mobile robots, due to the importance of th e task our architecture is centralized and deliberative, the MRS controls the people trajectory motion, and communication between robots and people is based on motion reactions. Section 2 discusses in the limits and scope of this research; section 3 details the architectural framework of th e MRS. In section 4 a methodology for people localization by the MRS is presented in deeply. The section 5 briefly describes a proposed strategy for control of the conduction task, while in section 6 the part of robots motion planning is trea ted. Finally, section 7 and 8 shows simulation results of the proposed methodologies, and the conclusions respectively in each section.