Page - 22 - in Joint Austrian Computer Vision and Robotics Workshop 2020

usually covered with rubble and debris and can ex- tend over several floors. Therefore the base plat- form of the mobile robot must be able to manoeu- vre in rough terrainandshouldbesuitable forclimb- ing stairs. As stated in [22] track based robots are designedforoperation inuneven,debris-covered ter- rain and are therefore ideally suited for natural dis- asters, moreover these tank like tracks add stability to the whole robot system [25]. Manipulation of ob- jects is also often required by S&R robots, whether fordemining, interactionwithvictimsor forgenerat- ing an uniquecameraangle [24]. Nowthat thesystemrequirementshavebeendefined, followingchapterexamines theapproachof theUAS to integrate these requirements intoaS&Rrobot. 2.SystemConcept Following section explains the hardware and sen- sor setup of the robot of the UAS as well as the soft- ware architecture for successful participation in the EnRicH 2019 trial or other S&Rapplications. 2.1.Systemdesign Figure 1 gives an overview of the implemented hardware components which are described in detail in the next section. As visualized the setup consists Figure1. Systemdesignoverview of two main parts, the ”Mobile Robot” and the ”Op- erator Station”. Although mobile S&R robots have autonomouscapabilitiesanoperatorstationisneeded toprovideasafetyfallbackandteleoperationsystem. Themobilerobot itselfneedstobeequippedwithnu- meroussensorsrangingfrom2Dand3DLIDARsfor obstacle avoidance and mapping, an robotic arm in- cluding an end effector (EEF) for manipulation and front and rear facingcameras for teleoperation. Theremainderof thischapterdescribes implemented hard- and software needed to provide a mobile S&R robot. 2.2. Implementation -Hardware The track steered mobile robot ”Tracker” of com- pany taurobGmbH[31] isusedas thebasicbuilding block. Due to adjustable crawler tracks, a high de- gree of off-road mobility is provided for maximum versatility [31]. In addition a 4 degrees of freedom (DOF) robot arm, of taurob GmbH, for manipula- tion tasks was mounted. The Robot Operating Sys- tem (ROS) API provided by taurob GmbH was the decisive factor for this mobile robot platform. The ”Tracker”, manipulator and the current sensor setup aredepicted infigure2. Figure2. Robbie hardware setup (1) Velodyne PUCK-VLP16, (2) Wifi/ 4G Antenna, (3) Rear-Facing Intel Realsense D435, (4) GarminGPSModule, (5)LEDHeadlight, (6)UeyeUI-3240LECamerawithCameraMount, (7) Operation Indication Light, (8) Robotic Arm, (9) SSM1+ Radiatiometer with mounted Probe, (10)EmbeddedPC, (11)Rear-FacingInternalCameraandSICKTIM-551-2050001, (12) taurob Tracker, (13) Front-facing TIM-551-2050001, (14) Intel Realsense D435 mounted on EEF, (15) EEF, (16)Front-Facing InternalCameraand InternalHeadlights Toallowamaximumlevelofflexibilityamodular hardware setup consiting of a sensor rig was chosen, thus enabeling easy replacement of sensors as well as software toenabledifferentS&Rtasks. AGarmin GPS module was attached to the sensor rig for out- door localization. For GPS-limited indoor scenarios depth and range sensors were used for localization and mapping. Therefore, a 3D Light Detection and Ranging (LIDAR) Velodyne PUCK VLP-16 [32] and two SICK TIM-551-2050001 [30] 2D LIDARs, one at the front and one at the rear, were attached to the robot. In contrast to the 2D LIDARs mounted in a planar arrangement, the 3D LIDAR was mounted at an angle of 20◦ to Robbie’s direction of travel. 22