The US Defense Advanced Research Projects Agency's (DARPA's) Assured Autonomy programme is scaling up in its second phase to incorporate larger neural networks and more challenging environments for autonomous systems, the head of the programme told Jane's .
The programme aims to develop formal safety assurances for the autonomous domain, ensuring the systems operate safely and correctly. According to DARPA, while there have been huge advances in the technologies needed for autonomous systems to operate - such as modelling and sensing - there has not been similar progress in assuring their safe and correct operation.
It also aims to develop mathematically verifiable approaches and tools that can be applied to machine learning (ML) algorithms, enhancing their autonomy and assuring they are achieving an acceptable level of safety.
Want to read more? For analysis on this article and access to Jane's unrivalled data and insight, learn more about our subscription options at janes.com/products