The National Conference for Postgraduate Research 2016, Universiti Malaysia Pahang 330 Human-Robot Interaction Using ROS Framework for Indoor Mapping Missions M.S. Hendriyawan Achmad 1 , Mohd Razali Daud 2 , Saifudin Razali 3 , Dwi Pebrianti 4 1 Department of Electrical Engineering, Universitas Teknologi Yogyakarta, Yogyakarta, Indonesia 1-4 Robotics and Unmanned Research Group (RUS), Instrument & Control Engineering (ICE) Cluster, Faculty of Electrical and Electronics Engineering, Universiti Malaysia Pahang, 26600, Pekan, Pahang, Malaysia 1 hendriyawanachmad@uty.ac.id, 2 mrazali@ump.edu.my, 3 saifudin@ump.edu.my, 4 dwipebrianti@ump.edu.my Abstract −The necessity of effective integration tools in robotics research play an important scientific contribution and highly recommended. However, robotics researchers have limited time during the experiments with definite targets and ultimately there is no choice but to use integration tools as software framework to avoid re-inventing the wheels. This research examines how humans work with teleoperated unmanned mobile robots to perform interaction in order to explore and build an indoor map by utilizing robotic software framework known as Robot Operating System (ROS). ROS infrastructure tools are involving together from the file system level to the community level, enables independent decisions about development and implementation. These experiments focus on two major area; the way human delivers the targets coordinate to the robot using ROS framework, and the way robot conducts SLAM as a feedback to the human related to current location and occupied map within ROS platform. Hector SLAM plays an important role in 2D localization using 2D LIDAR sensor which is needed by Octomap in order to construct 3D mapping simultaneously using Kinect sensor. The results showed that human-robot interaction using the ROS-based teleoperated system for mapping task is easy to configure and propose an effective tool in term of large-scale service robot development. Keywords − ROS; Kinect; 2D LIDAR; Hector SLAM; Octomap; Human-Robot Interaction 1. INTRODUCTION The robotics technology has developed rapidly and applied to many fields, such as industry, education, health, household, entertainment, and the military [1]. An interaction between humans and robots in completing the task is one of the benefits of robots to make people's lives better. In some cases, robots will share the same workspace and work closely with humans to accomplish collaboration tasks as part of their day-to-day work [2]. The aim of Human-Robot interaction research is to determine models of interaction in term of hardware & software development to obtain effective collaboration between humans as masters and robots as slaves [3]. Robots with many sensors used for exploration and mapping tasks without software framework will get a higher level of complexity compared to utilizing of software framework. ROS was developed to address a set of big challenges of complexity being faced when developing large-scale of service robots [4]. There are many studies discussing ROS- related to humans – robots interaction to accomplish collaboration tasks [5]. The Simultaneous Localization and Mapping (SLAM) technique is a common issue in mobile robots discussion. The basic idea of SLAM is to use the robot for explorations mission at an unknown location and construct the map base on obstacles surrounding it. In reality, the non-ideal behavior of sensors and actuators produce very complex problem and need to be addressed [6]. Researchers have been studying localization and mapping techniques by using many types of sensor devices, such as sonar depth sensors [7], visual sensors [8], and laser depth scanners [9]. Nevertheless, the techniques tend to devices-oriented due to the differences of work of the sensors and the characteristics of each device as well [6]. The idea to puts SLAM technique on the mobile robot equipped with multiple sensors is a challenge due to software integration complexity as has been done in [10]. This paper proposed the methods how to implement the ROS framework to accommodate human-robot interaction in order to carry out indoor exploration and mapping tasks (SLAM) utilizing the differential mobile robot. For localization function, we used Hector SLAM algorithm. And for 3D mapping, we used Octomap algorithm. Both of techniques are open-source and widely used together with ROS.