Unmanned Systems Technology 024 | Wingcopter 178 l 5G focus l UUVs insight l CES report l Stromkind KAT l Intelligent Energy fuel cell l Earthsense TerraSentia l Connectors focus l Advanced Engineering report

57 Consumer Electronics Show 2019 | Report Navatics attended the show to exhibit its Mito UUV, a 3.45 kg ROV designed for underwater photography and video recording. The system moves around at up to 3.8 knots using four thrusters, and is programmed with the company’s proprietary active stabilisation algorithm. This allows it to ‘hover’ in place amid currents of up to 3 knots, with a tilt lock of ±45° to maintain a consistent angle and movement of the camera. A tether extends from the dorsal hull to a floating buoy, which integrates a wi-fi modem to enable users to control the UUV up to 500 m away from the buoy. The tether extends up to 50 m, covering the Mito’s maximum recommended depth of 40 m, while the buoy carries a battery to give up to four hours of operation and has a solar panel to help extend usage beyond that, depending on levels of sunlight. The solar panel provides the tether with an extra 1-2 hours of use on top of its normal battery life,” said Andreas Widy. “The MITO itself is powered by its own internal battery that lasts for around 4 hours. The camera records 4K video at 30 fps onto a 64 Gbyte micro-SD card, while transmitting it in 1080p in real time to the user’s control station. Two lights, producing 1000 lumens each, provide illumination for the cameras if the UUV should be directed into underwater caves or structures. The system is currently rated to operate in water temperatures ranging from -5 to +40 C, and will be available for shipping this March. SmoothTalker exhibited its range of cellular signal boosters for vehicles. The boosters can operate at a range of frequencies and bands, including 3G, 4G, 4G LTE, 700 MHz and 1.9 GHz. These signal amplifiers use SmoothTalker’s proprietary Stealth Tech configuration, which is designed to prevent issues that can arise between boosters and cellular towers, such as network interference. These safeguards include powering down when approaching a cell tower, active gain control and anti-oscillation controls. AlphaICs unveiled its new alphaEdge processor board for SAE Level 2+ self-driving vehicles and other autonomous ground and air vehicles. The system comes in several versions, including a solution that consumes 10 W and another that consumes 40 W, with eight camera inputs, support for automotive interfaces across CAN, UART, GPIO, I2C and SPI protocols, HDMI, USB 3.0 and Gigabit Ethernet. It also comes embedded with a software tool chain to support a range of AI frameworks, including TensorFlow, Caffe2, ONNX, and PyTorch. AlphaEdge also uses the company’s proprietary Real AI Processor (RAP) inference chip, which is programmed to assist with supervised machine learning in applications such as image classification, ADAS, autonomous decision-making and dialogue classification, providing update rates at the millisecond level. RAP is based on a computing paradigm using self-learning ‘agents’ (rather than scalars or vectors, as in CPUs and GPUs respectively), each of which consists of a group of interconnected tensors. This, the company said, provides the architecture for a wider divergence of threads and thus accelerated reinforcement learning. Eyedea displayed the Pitta, its autonomous ‘selfie’ UAV. The 200 g quadrotor device uses a 4K camera that records video at 30 fps and captures photos at 16 MP resolution, with a flight endurance limit of 15 minutes while recording. Powered by a lithium-polymer battery, the system maintains stability during flight using triple-axis accelerometers, gyroscopes and magnetometers, software for image stabilisation in the main camera. An additional camera faces downwards for optical flow positioning, which uses visual image inputs to track the movement of the UAV over the ground and output movements to counteract external forces such as wind to keep the craft in place. The 170 x 170 mm craft has further software embedded to enable object detection, visual tracking and a ‘follow me’ mode to record users as they move around. Video files are recorded using H.264 encoding and can be output as AVI or MP4 files. Unmanned Systems Technology | February/March 2019 Navatics’ Mito photography and video UUV