Unmanned Systems Technology 004 | Delair-Tech DT18 | Autopilots | Rotron RT600 | Unmanned surface vehicles | AMRC | Motion control | Batteries
with the master, and usually then implements an emergency protocol such as staying in one place in the air, finding a safe place to land or returning to base. This approach requires more monitoring of the master board to determine if there is a problem that requires a switch to the other board. This is different from a dual-processor system, which separates the main processor (the FCS) from the sensor and data processing unit, often called the attitude heading reference system (AHRS). The AHRS handles data from the inertial measurement unit, the magnetometer, GPS, altimeter, airspeed sensor and so on, while the FCS handles the flight management and comms, with the control and navigation algorithms in the autopilot. But the autopilot can also include the payload management and comms, which can make the autopilot software more complicated. There is also the triple redundancy approach, which as its name suggests uses three boards, each with their own processor, plus a dedicated decision- making board for implementing voting. This voting scheme takes the output of each board and checks that they all agree. If they don’t, the results of the two that do agree are used, and the third board is taken out of service, but the mission can still continue with two boards in operation. Using three sets of hardware can make this a more expensive approach but provides greater reassurance of safe operation in all circumstances, which can be needed when using valuable payloads. The need for redundancy can sometimes be directed by the regulatory requirements. Regulatory bodies are requesting it, and each country can have a different requirement – for example, for operations over an urban area, dual redundant downlinks and redundant autopilot boards are often needed. Increasingly these days, a technique called virtualisation is being used in larger platforms to provide the same separation of functions as multiple boards. Instead of having separate boards, the functions can be run in separate parts, or partitions, of a more powerful processor. Each partition has access to its own protected memory space and can be used to implement a dual or triple redundant architecture on a single board. This approach is often based around a commercial real- time operating system (RTOS) that can access parts of the processor to help the virtualisation, and allows the developer to write, test and qualify code in a high-level language such as C. However, using just a single board, virtualisation can be vulnerable to power or component failures, although
Made with FlippingBook
RkJQdWJsaXNoZXIy MjI2Mzk4