Unmanned Systems Technology 023 I Milrem Multiscope I Wireless charging I Logistics insight I InterGeo, CUAV London & USA show reports I VideoRay Defender I OS Engines GR400U-FI I Ultrabeam Hydrographic Ultra-2 I IMUs

45 such obstacles can lead to the Enterprise Manager reallocating a given job to another UGV to save time and improve facility efficiency. This sense-and-avoid capability has led to Omron to refer to its LD robot as an ‘AIV’ (automated intelligent vehicle) to emphasise that its system, unlike logistics AGVs, does not need tape or rails to be installed around a factory or warehouse, which can take time and money. “The LD robot has two driving wheels and four caster wheels,” Adam explains. “The caster wheels support the load, while the driving wheels push the mobile robot, and can rotate the robot to move in the direction needed. Each wheel has its own suspension, making the platform more stable, even amid small irregularities on the floor.” The battery gives about 13 hours of operation, with a maximum speed of 1.8 m/s, and each LD robot can decide when it is time to join the automatic docking station and recharge. Each hour of recharging gives enough battery power for five hours of operation. For further logistics organisation at the upper end of supply chains, South African company Dronescan is working to mature UAV-based stock-keeping. Its custom Hadeda payload design is typically for use on a DJI Matrice M100 craft inside warehouses with long, tall shelves that need inventory management. The payload itself is built around a custom circuit board, which holds the Arduino autopilot as well as power connections for key components. Most important of these is an image-based optical sensor for scanning inventory barcodes, which is equipped with a light for night-time scanning and a two-axis gimbal for ‘sweeping’ motions, to remove the need to line the UAV up perfectly with the barcodes. The payload can also be reconfigured for RFID tag scanning or integration with sensors such as laser rangefinders to scan the number of boxes across a shelf. “While developing a scanner small enough to be carried as a payload, we found the M100 to be the ideal development platform to get the scanning payload off the ground,” says Craig Leppan, co-founder of Dronescan. “Its open architecture allowed us the general ‘configurability’ to integrate the necessary power and comms interfaces, as well as the platform stability to show that no costly piloting training or anti- collision sensors were needed. And even with the constraints of battery life and the current limitations of autonomy, it can be a productive tool in warehouses.” The company says scans could be conducted at least five times quicker using a UAV-mounted scanner payload compared with conventional methods such as using forklift trucks or similar to raise personnel equipped with scanning devices to record stock. The fastest Hadeda demonstrations have recorded one scan every 2 s. “Autonomy still eludes this industry, as you can’t use GPS inside warehouses,” Leppan adds. “Lidar is sometimes brought up by other players for in- warehouse simultaneous location and mapping, but that just brings down the primary UAV constraint – battery life. “Advances in battery life are very much the game-changer for pushing the potential for autonomous logistics as we see it. Alternatively, advances in fuel cells for small UAVs could take us from 10 minutes of scanning between battery swaps to hours.” Moving freight Advances in autonomous shipping are being driven not only by improvements in sensors and autopilots, but as a Logistics | Insight The Omron LD mobile robot senses obstacles using lidar, and receives jobs and directions throughout a facility’s floors from software over a wi-fi link (Courtesy of Omron) Unmanned Systems Technology | December/January 2019 The Hadeda payload can scan warehouse inventory barcodes three to five times faster than traditional methods when mounted on a UAV (Courtesy of Dronescan)