Unmanned Systems Technology 004 | Delair-Tech DT18 | Autopilots | Rotron RT600 | Unmanned surface vehicles | AMRC | Motion control | Batteries

36 and incorporating sensor data from the gyroscope, navigation sensors, airspeed indicator and other sensors on the aircraft, including data from the engines, and putting that through a controller to determine the exact state of the aircraft. This allows the precise servo and engine control needed when changing between modes while in flight. A key part of the development of this capability was identifying the different failure modes that can occur, and mitigating those modes in the filtering algorithms and the operating system to achieve the high reliability. Other developers have customised an autopilot for particular platforms, particularly larger aircraft, by using dedicated software to get the faster response times. The inner-loop algorithm that handles vehicle stability is programmed directly into the processor rather an operating system, using reduced instruction set commands. These inner-loop algorithms are analysed under worst-case load conditions, with the timing requirements determined by the different ways the vehicle behaves, in a process called stability node analysis. So, for example, a fixed-wing system would be expected to meet control deadlines at a frequency of at least ten times that of the vehicle ‘short period’ with plenty of spare cycles available. The short period is the time taken for the active surfaces of the aircraft to change and settle down to a new position, and this 10x requirement is more than adequately achieved with current technology.  This approach has also been tested on rotor craft. As an example of the challenge of developing such algorithms, one particular single-rotor craft, with all the associated gyroscopic complexities, spent 18 months in the US without achieving stable controlled flight. Instead, using stability node analysis and programming the control loops directly into the processor allowed full automated control within three months of development, and with a far lower budget than the original control system. Redundancy Guidance and navigation algorithms in the flight management systems can also benefit from this efficient coding, although the latency is not such an issue in these applications, as the clock speeds of modern processors are more than adequate. After speed of response, the second approach to safety is to use different forms of redundancy in the design of the system, by separating out the core autopilot functions or adding multiple versions of the same boards and software. Dual redundant systems use two boards running the autopilot algorithms in a master-slave configuration. The slave board takes over if there is a problem Autumn 2015 | Unmanned Systems Technology Focus | Autopilots Weight is a key consideration for an autopilot, and the Easypilot weighs just 4.5 g (Courtesy of UAS Europe) The Arcturus Jump hybrid UAS operates as a fixed-wing and a rotary aircraft, which is a challenge for its autopilot (Courtesy of Arcturus UAV)