114 It’s usually a mistake to make absolute statements about the complexity of the real world, but it’s probably safe to say that AI is affecting all areas of uncrewed systems technology and operations (writes Peter Donaldson). AI is used in a wide range of tasks, such as computer vision and sensor fusion, localisation and mapping, decisionmaking and autonomy, autonomous control and autopilots, and optimising energy usage to maximise range. It is also increasingly important in comms and connectivity, which support everything else uncrewed systems do. Here, AI algorithms are embedded in many functions. For example, they can manage comms protocols and network configurations to ensure the vehicle uses the optimal means for data transfer under the prevailing conditions. They do this by selecting the frequencies, channels and data transmission rates best suited to maintaining reliable connections. The radio frequency environment can change rapidly in unpredictable ways, and AI can help adapt the comms system in real time. It does this by assessing network conditions, signal strengths and interference levels, and can decide to adjust antenna orientations to switch to alternative channels or comms modes. Where multiple uncrewed vehicles are sharing the same channels, AI can manage traffic to help minimise interference and eliminate ‘collisions’, in which two or more vehicle radios try to transmit on the same channel or frequency at the same time; the algorithms can schedule transmission times and frequencies to minimise conflicts. In spectrum management, a form of AI known as cognitive radio can detect unused or relatively uncongested frequencies and adapt the system to use them. Uncrewed systems operating in teams or swarms often form ad hoc mesh networks to relay data among themselves and share it with a control centre. Here, AI can find the most efficient data transmission routes through the networks, especially when network topologies change frequently as the vehicles carrying the nodes move around and change their relative positions during a mission. AI can also help to secure comms by implementing encryption and authentication protocols, and by detecting and responding to potential security threats or cyber attacks. Quality of service optimisation is another task that can be eased by AI, with algorithms allocating network resources based on the needs of each vehicle, for example giving higher priority to critical mission data. Engineering fault-tolerance and self-healing into comms networks is another important approach to improving reliability, and AI can help in the event of a failure in one or more nodes by rerouting traffic through alternative paths. Predictive maintenance is another AI approach to improving reliability, in which it recommends maintenance or replacement of components ahead of predicted failures. Finally, natural language processing is a type of AI that can enable operators in command centres to issue high-level commands and receive status updates in plain language, helping mutual understanding between humans and robots. Natural language processing is a type of AI that can enable operators to issue high-level commands and receive updates in plain language December/January 2024 | Uncrewed Systems Technology PS | AI in uncrewed system comms Now, here’s a thing
RkJQdWJsaXNoZXIy MjI2Mzk4