Abstraci -In this paper, a real-time obstacle avoidance method is discussed fur a remote mobile robot controlled by a teleoperator. The method enables the remote user tu drive the mobile robot without collisions. It consists of a real-time environment modeling algorithm using ultrasonic sensors, and a mobile robot control algorithm for obstacle avoidance. The former solves the limitations of the ultrasonic sensors such as noise sensitivity, poor directionality, and specular reflection fur obtaining more accurate sensor data. Then the latter conducts wall or center following implemented by a fuzzy controller for obstacle avoidance using the environment model obtained by the previous algorithm. The command for obstacle avoidance is actually applied as a reflective force using a haptic device such as a force feedback joystick. The remote mobile robot, ROBHAZDTrnuBot in HAZardous environment -Double Track) developed by KIST(Korea Institute of Science and Technology), is employed as the mobile robot model. The normal and the maximum speed of the ROBHAZDT are 3 . l M h and 7.2kmm, respectively. The weight of the robot is 50kg. Simulations with the ROBHAZDT model are carried out tu verify the performance of the proposed method.
A new grid map-merging technique which consists of virtual emphasis by one-way observation, curvature-based map matching and particle swarm optimisation is proposed. The proposed technique can improve not only the accuracy of map merging, but also the flexibility of multi-robot systems. The improved performance is verified by showing higher similarities than the existing map-merging techniques in experiments.Introduction: Multi-robot systems (MRSs) have several advantages over single-robot systems such as efficiency and flexibility [1]. When MRSs are applied to performing tasks in unknown environments, robots should share their information on the surrounding environments to realise the advantages. As one of the essential issues for sharing information, map merging is to fuse maps which have been built by different robots [2]. If the robots do not know their initial poses to one another, map merging is challenging because the map transformation matrix (MTM) among them cannot be initially acquired. To acquire the MTM, mutual observation among robots can be used, but it may cause a limitation of operating MRS because robots should meet each other at the same time [3,4]. Map matching with the overlapping area between maps can also be used, but it may be caught on local maxima if overlapping areas are insufficient [5,6].This Letter proposes a new map-merging technique which consists of three steps. First, the structures represented by occupied grids are virtually emphasised along two-dimensional (2D) normal distribution with the robot position estimated by one-way observation. Secondly, curvature-based map matching with the emphasised structures is conducted to compute a coarse MTM. Finally, map matching based on particle swarm optimisation (PSO) is applied to computing a fine MTM. Since the proposed technique uses one-way observation instead of mutual observation, the flexibility of the MRS is improved. Besides, since a coarse MTM is used to reduce the search space for PSO-based map matching, the proposed technique can not only avoid local maxima, but also produce more accurate MTM.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.