Unmanned Systems Technology 003 | UAV Solutions Talon 120 | Cable harnesses | Austro Engine AE50R and AE300 | Autonomous mining | AUVSI 2015 show report | Transponders | Space systems

14 B io Inspired Technologies calls itself an ‘adaptive’ technology company, and as such it is looking to exploit approaches used in nature for electronic systems for UAVs and other autonomous systems. For the past six years it has been developing technologies such as memristors – resistors with a memory (see page 16) – and neural nets in integrated circuits for the US Department of Defense’s DARPA research agency. The resulting devices mimic natural processes such as neural networks, which can emulate the way neurons operate in the brain. The company’s chief executive Terry Gafron points out that these devices are not traditional processors that use the established Harvard architecture, which separates instructions from data. The whole process of making a decision with a conventional processor is binary, it’s a voltage that’s either a ‘1’ or a ‘0’, but that’s not how nature behaves, and the idea is that by emulating nature – even poorly – we would be able to build much more successful control systems. To that end Bio Inspired has been using memristor-based neural networks for image recognition at the heart of the control system for an autonomous vehicle. Instead of capturing an image and trying to compare it exactly to an image in a database, a neural network quickly produces a probability of how close the match is, even if the image is unclear. This gives a quicker response from a smaller database with dramatically less processing power. The memristor work is unique in that the ion-conducting silver chalcogenide material on which it is based acts as an exact analogue of a biological synapse, a structure in the nervous system which allows a nerve cell to pass an electrical signal to another cell and which plays a role in the formation of memory and therefore learning. Gafron says, “If we start building biological machines and circuits that learn, that recognise shapes, then we are closer to the way a house fly operates. When you look at the capabilities of a house fly, it’s infinitely more capable than a fighter aircraft, because the systems are capable of making decisions based on ambiguous The chief executive of Bio Inspired Technologies tells Nick Flaherty how the natural world gives a guiding light for its UAV control systems Summer 2015 | Unmanned Systems Technology Nature trail