Unmanned Systems Technology 022 | XOcean XO-450 l Radar systems l Space vehicles insight l Small Robot l BMPower FCPS l Prismatic HALE UAV l InterDrone 2018 show report l UpVision l Navigation systems

12 HP has launched its first metal 3D printer, marking a significant move towards mass production of custom parts (writes Nick Flaherty). The company says its Metal Jet printer can provide up to 50 times more productivity at a far lower cost than other 3D printing methods. The printers are already being used by GKN Powder Metallurgy and Parmatech for the factory production of final parts in stainless steel. The printer has a relatively large bed size, of 430 x 320 x 200 mm, with six print heads, each with four nozzles on two print bars to provide redundancy. A volume of space, or voxel, can be addressed by up to four nozzles. The base is a layer of metal powder, and a latex binding agent is jetted onto the powder bed on a 21 micron grid to build up the item, which has a homogeneous microstructure. At the end of the process the part is sintered using standard methods. The stainless steel- finished parts meet or exceed ASTM and MPIF standards. Using the latex binder means the cost of the nozzles is 100 times lower than that of piezoelectric nozzles, allowing the redundancy in the print head. This has 63,000 nozzles firing up to 500 million drops every second. GKN Powder Metallurgy is using the printer in its factories to produce functional metal parts for automotive designs. The company produces more than three billion components a year, and expects to print millions of production- grade Metal Jet parts for its customers as early as next year. Dr Martin Goede, head of technology planning and development at Volkswagen, said, “By 2025, the Volkswagen Group will have introduced 80 new electric models. A single car consists of 6000-8000 different parts. “A big advantage of an additive technology is it allows us to produce many of these parts without first having to build manufacturing tools. By reducing the cycle time for the production of parts, we can produce a higher volume of mass production very quickly.” Commercial Metal Jet printers will cost $399,000, and will begin shipping in 2020 to early customers. Printing for the masses Manufacturing October/November 2018 | Unmanned Systems Technology Researchers from the Salk Institute and the University of California San Diego have used a new type of algorithm to train gliders to autonomously navigate atmospheric thermals, soaring to heights of 700 m (writes Nick Flaherty). A reinforcement learning algorithm was used to manage vertical wind accelerations and roll-wise torques, and provide a navigational strategy that applies directly to the development of autonomous soaring vehicles. Reinforcement learning is a type of machine learning, where a software agent learns how to behave in an environment based on performed actions and the results. “This is an important step towards artificial intelligence in UAVs – how to soar autonomously in constantly shifting thermals like a bird,” said Professor Terrence Sejnowski, head of Salk’s Computational Neurobiology Laboratory. Massimo Vergassola, a professor in the UC San Diego Department of Physics, said, “This is a novel instance of learning a navigational task in the field, where learning is severely challenged by a multitude of physical effects and the unpredictability of the natural environment.” The team equipped several 2 m- wingspan gliders with a dedicated flight controller that implemented autonomous flight policies using precise control over bank angle and pitch. The navigational strategy evolved from data collected from the vertical wind acceleration and vertical wind velocity gradients across the gliders’ wings. UAVs learn how to soar Airborne vehicles The new type of algorithm is said to be an important step towards putting artificial intelligence in UAVs HP’s Metal Jet 3D printer is already being used to make metal parts for the automotive industry