Unmanned Systems Technology 025 | iXblue DriX I Maintenance I UGVs I IDEX 2019 I Planck Aero Shearwater I Sky Power hybrid system I Delph Dynamics RH4 I GCSs I StreetDrone Twizy I Oceanology Americas 2019

46 The company is focusing on convoys of two trucks for the time being. Further development of the technology aims to solve the issue of a shortage of drivers in the trucking industry, owing to the long hours and associated stress. Sanitation Clearpath Robotics’ Warthog UGV has been used in a joint project between researchers from Northeastern University in Boston, in the US, and the University of Alicante in Spain to develop an autonomous outdoor waste collection system. “They’ve acquired a Warthog for the project, and there were some modifications they wanted to make when they bought it,” says Robbie Edwards, engineering manager at Clearpath Robotics. “They needed a UGV that was big and robust enough to integrate a forward- facing Lidar, and a UR10 robotic arm from Universal Robots, which is that company’s biggest arm.” The modified Warthog is now serving as a proof of concept and as a platform for developing the code and algorithms to autonomously identify, grab and remove waste items from the ground. The UGV’s chassis measures 1.34 m from front to back, and 1.38 m from side to side. Its base weight, including battery pack, is 280 kg, with a 272 kg payload limit and the ability to clear obstacles up to 25.4 cm tall. The Warthog is capable of a maximum speed of 18 kph, with endurance varying depending on the battery type – 2 hours can be achieved with a lead-acid battery pack, while up to 3 hours is possible when running on lithium-ion batteries. Localisation is provided by a Sick LMS511-series 2D Lidar installed on the front of the Warthog, which provides a 190° horizontal FOV. The UR10 arm is fitted with a Robotiq 2-Finger 85 gripper, which has a ‘grasping’ aperture of 85 mm and a maximum payload capacity of 5 kg (although the UR10 itself can technically carry up to 10 kg). Three Intel RealSense D415 cameras have also been installed. Two of them are fixed to each side of the Warthog’s front, and point downwards to cover the targeted cleaning area, while the third is mounted on the gripper and acts as a ‘hand-eye’ camera. “Not only was the Warthog big enough to carry the UR10, but it retained enough space in the back for a bay where trash could be placed and taken to a tipping area,” Edwards adds. Navigation was achieved largely by using the algorithm stack that came with the pre-installed ROS-Indigo driver software. However, to reduce control irregularities caused by the Warthog’s skid-steering, the team installed a differential drive system (by putting the rear wheels on casters) and increased the update rate of the controller to 40 Hz. In addition, Grasp Pose Detection, a publicly available grasping software stack, was used to program the arm to determine grasp poses (and therefore the most successful angle and approach for the arm to take towards the trash). The software is based on Lidar point- cloud data, with data from the two cameras being converted into Lidar data using a software node in ROS- Indigo. For cleaning tight interior spaces, Neovision Industrial Vision Systems and M-Tecks Robotics have partnered to develop the Jetty robot. This uses six tracks around its body to drive through air vents and other ducts of varying shapes and sizes, with a dry-ice jet to clean such spaces without having to use toxic chemicals. The system can travel at up to 0.25 kph and weighs 35 kg, not including tether cables for comms with the GCS and for supplying the power and dry ice. A camera mounted on the Jetty feeds 1280 x 960 live video at 30 fps to the operator for controlling and guiding the vehicle through ducts between 36 and 71 cm in diameter (or up to 130 cm in diameter if optional arm extensions are used). April/May 2019 | Unmanned Systems Technology Insight | UGVs Sanitation project demo version of the Warthog UGV, which is large enough to integrate a heavy robotic arm, a 2D Lidar and a bay for storing and carrying refuse

RkJQdWJsaXNoZXIy MjI2Mzk4