Providing guarantees against risk of faults for autonomous systems
We need to build our trust for autonomous and partially-autonomous machines, with the advent of their proliferation. Unlike more standard automatic machines, autonomous systems are capable of: reasoning, independent planning, decision making and learning. Systems are becoming increasingly powerful, yet this autonomy comes with the trade-off in more bugs and many possible unpredictable outputs; hence risks.
Risks compound as autonomous systems must not only safely operate on their own, yet they must also co-exist with humans, as well as other autonomous machines. We thus need to be able to guarantee their behaviour to ensure such systems are safe. A toolkit has been developed by computer scientists at Imperial College London that tackles risk in autonomous systems through verification. Verification provides guarantees in the reliability against faults as well as against cascading of faults of autonomous machines.
Download our whitepaper for this technology
+44 (0)20 7594 8599