Unmanned Systems Technology 009 | Ocean Aero Submaran S10 | Simulation and testing | Farnborough report | 3W-110xi b2 TS HFE FI | USVs | Data storage | Eurosatory/UGS 2016 report

18 In conversation | Jonathan Auld together,” says Auld. “We are working with satellite positioning company Veripos in Scotland, which has a global network of base stations to generate PPP corrections, currently in the marine and agriculture sectors but we expect to lean on that for autonomous systems. “All these pieces get you to 20 cm accuracy across large regions, and we then layer on top of that the protection level but at a much more challenging performance level than is typically used at the moment for aviation.” Integrity concepts This creates a positioning system that uses satellite for the rural areas where there may not be high-speed cellular support, and still supports transition to an internet link in urban areas where satellite doesn’t work as effectively. “That’s the work we are doing with Veripos, to figure out the integrity concepts in delivering the data that way,” Auld says. EGNOS and other networks are doing the same thing in the satellite link, but the problem is that those correction services don’t provide the data at the speed and quality that would allow for PPP on the receiver. “These are some of the problems we are trying to solve,” he says. “The other thing to think about is that an autonomous driving system is not just about GNSS navigation but all the other integrated systems that need positioning data. The idea is to provide the absolute positioning source that can help all those other sensors do their job as well, so some of the safety issues can be solved at higher levels in the systems. “The exact specification depends on the manufacturer you engage with, and some have been working on this for many years. But there is also a significant education component we have to consider, as the expectation of a lot of people in the automotive world is set by what they are using now in a satnav system with a map.” Agricultural applications require even higher levels of accuracy, he says. “Right now we sell our highest precision products into agriculture as smart antennas for crop management. Farmers are already using the technology but in the longer term they want to have the tractor drive itself.” As a relatively small company designing its own GNSS receiver chips, NovAtel is potentially vulnerable to larger semiconductor suppliers. “The big guys will want to sweep up as much as they can into the chips they are selling, but because the solutions don’t exist yet, there are a few years to go before the challenges are worked out, and there’s an opportunity here,” says Auld. “We think we have all the pieces to provide this high level of accuracy safely, and they are not thinking the same way as we are… yet.” There is also an issue with the receiver’s power consumption. “That is something we have to optimise and reduce, but there’s the trade-off between power and mitigating RF interference,” he says. “You can devise a lower power solution but it may be more vulnerable to jamming and interference, so if you want it to be safe all the time it won’t be as low power as it could be. “We are looking at where to draw the line – we currently have a digital custom chip with a CPU, and an analogue chip to handle the RF,” Auld says. “We are exploring a mixed-signal chip, a small surface mount module with a chipset, and a system in package – all of those have power and cost trade-offs. “There’s a window of 2020-23 that most people are zeroing in on to have the autonomous vehicle options available. Some manufacturers are pushing for more aggressive timelines, but regulatory questions and the legal framework are other issues still to be addressed.” August/September 2016 | Unmanned Systems Technology The expectation of a lot of people in the automotive world is set by what they are using now in a satnav system with a map NovAtel’s SPAN-IGM GNSS/INS satellite receiver is an example of a truth system used in automotive testing to determine the reliability of navigation data from a satellite. The challenge over the next few years is to miniaturise the technology for production driverless vehicles