Unmanned Systems Technology 014 | Quantum Tron | Radio links and telemetry | Unmanned Aerial Vehicles | Protonex fuel cell | Ancillary systems | AUVSI 2017 Show report

20 In conversation | Orin Hoffman “That was another area for testing different designs, but it was simplified by using 3D printing to quickly generate new components that form the shape of the collection tube that could be snapped on, tested and modified as needed in the field. That drastically speeded up our design cycle. “In the current design we are using a larger containment vessel, up from 8 to 12 in, and have a wider opening. One thing we found was that, depending on the orientation of the fish, their spines can get caught, so making the diameter wider on the intakes helps,” he says. Next steps Now that the design is fixed, the team is raising finance through crowdfunding to take the system into production. “When we go into production we will then make custom thrusters that are optimised for low cost,” says Hoffman. “We are targeting a price of less $1000. “We know how to make an expensive robot scale to cost. The Roomba [robot vacuum cleaner] for example is a fully autonomous robot, and that type of capability with all the sensors and motors was initially $10,000; now it’s $700.” The design will evolve with more functions such as machine learning and more autonomy, but the ROV will still need a human operator, he says. “The point of the project was figuring out what kind of robot we needed, and in the future we will add new capabilities – that’s how robots in the field develop. “My gut feeling though is that there will always be some teleoperation. You are pulling the robot up to collect fish anyway, so users will always want that control, see what’s down there.” The ultimate aim is to create a sustainable supply of the fish as food. “You have to eat them to beat them,” says Hoffman. “That’s why it has to be low cost for fisherman to go out and capture them, to make the supply sustainable. “A few local restaurants serve lionfish but there isn’t a consistent supply from recreational divers [rather than commercial fishermen] so the hope is that by giving fishermen this tool in the local area we can at least develop a consistent supply.” He adds, “Designing for operation in water is hard, as is designing for pressure, and we know how to make cheap electronics, but there are a lot of things in salt water that drive costs up and reliability down, and those will always be challenges. “From our analysis these are challenges that can be met, but the next phase will be another manufacturing effort that will be even harder than the biology phase.” The next phase is to deploy up to 20 units of the second version of the prototype (the LF2) to evaluate the operation of the design. After that, the team plans to trigger the production phase, which will take 12-18 months. June/July 2017 | Unmanned Systems Technology Orin Hoffman is currently the Autonomy and Robotics lead at the US Department of Defense Innovation Unit Experimental (DIUx), whose role is to increase the DoD’s access to commercial technology, He was formerly chief technologist at Endeavor Robotics, formerly iRobot Defense and Security, which supplies UGVs to customers in the defence, public safety, and energy and industrial markets. He was also vice- president of advanced product development. As director of third- party development, he led the production of the iRobot Developers Kit, which provides open access and hardware integration support for iRobot’s family of systems. He joined iRobot as a software engineer in 2006, having been a senior lecturer at Mount Holyoke College in Massachusetts. He has a BA in Physics and Philosophy from Amherst College, Massachusetts, and an MSc in electrical computer engineering from the University of Massachusetts, Amherst. Orin Hoffman Front view of the LF1, showing the collection aperture and the probes to stun the fish