Unmanned Systems Technology 001 | UAV Factory Penguin C | Real-time operating systems | Hirth S1218 two-stroke twin | Base stations | ASV C-Enduro | Composites | Datacomms

Canada QNX +1 613 591 0931 www.qnx.com France Esterel +33 130 686160 www.esterel.com Germany Wittenstein +49 7931 4930 www.wittenstein-aero.de UK Green Hills Software +44 (0)2380 649660 www.ghs.com Lynx Software Technologies +44 (0)118 965 3827 www.lynx.com Wind River +44 (0)1793 831831 www.windriver.com US Express Logic +1 858 613 6640 www.rtos.com Freescale Semiconductor +1 512 895 2000 www.freescale.com Green Hills Software +1 805 965 6044 www.ghs.com Lynx Software Technologies +1 408 979 3900 www.lynx.com Mentor Graphics +1 800 592 2210 www.mentor.com Micrium +1 954 217 2036 www.micrium.com Rockwell Collins +1 319 295 5100 www.rockwellcollins.com Tenasys +1 503 748 4720 www.tenasys.com Texas Instruments TI-RTOS +1 512 434 1560 www.ti.com Wind River +1 510 749 2411 www.windriver.com Some examples of real-time operating system suppliers development, as it is providing a secure and real-time environment for complex systems with limited processing power and memory space, and also includes an RTOS. However, several open standards groups such as Thread, AllSeen and the Open Interconnect Consortium are emerging with real-time solutions for such systems and are seen as fragmenting the market. The best way to kill interoperability is to create too many open standards, but in the safety-critical world that doesn’t happen – everyone in the business is in FACE to ensure that the people building unmanned vehicles have interoperable standards. The future Specifications are emerging though that may be used in autonomous systems. For example, COAP – the constrained application protocol – from the Internet Engineering Task Force sits above the IP layer, and is designed to carry out reliable transactions for low-power applications, while the Routing Protocol for Low-Power and Lossy Networks (RPL) is a routing protocol intended for constrained systems. It may take another five years though before COAP and RPL make it into real-world systems There is a continuing tension between technical capability and certification. For example, research shows that a control system that can adapt to inputs from sensors around the aircraft can handle potentially catastrophic events. For example, if a wing is lost for whatever reason, the system would adapt to keep flying under RTOS control. Under current safety standards, certifying that kind of system would be extremely difficult. Nonetheless, that level of adaptation is quite an achievement, says one developer who is calling for more research in this area, even though it puts more demand on the software until the control systems start to look more like artificial intelligence. Conclusion A lot of this work has been done for unmanned systems that are more advanced than their (more complex) autonomous counterparts. There are the additional challenges of safety and security with autonomous systems, and many air vehicles are not regarded as autonomous as they rely on the ground station and the data link. Providing the right infrastructure from an RTOS up through all the applications for a rugged, secure and reliable real- time implementation is still a major challenge for RTOS developers. They are at the forefront of the certification issues, and can see more of those ahead for demonstrating with absolute certainty that autonomous vehicles can operate safely. The fragmentation of different standards is a real risk at this point. While standards such as FACE are driving UAV autonomous systems, they are not necessarily an important part of projects such as Google’s driverless cars or autonomous aerial base stations, which are more cost-sensitive and constrained in terms of resources than military systems such as those being developed by the US armed forces. However, ensuring that these more commercial developments meet the same levels of safety and security for the software, from the RTOS upwards, is a key issue. Eventually autonomous systems will be certified, but it will be very difficult. NASA for example has been funding research into how safety certification might go ahead, as well as experimenting with autonomous driving on the Mars rovers as there are no safety restrictions here. Acknowledgements The author would like to thank David Kleidermacher of Green Hills Software, Alex Wilson of Wind River UK and John Blevins of Lynx for their input into this article. 35 Unmanned Systems Technology | November 2014