From surveillance applications in controlled environments to navigating and mapping unknown hostile environments, autonomous, multi-sensor platforms are increasingly important to improve mission efficacy and reduce human risks.

While autonomy in controlled environments has been heavily investigated, it remains difficult to ensure reliability of autonomous systems in dynamic environments and unknown, or out-of-distribution, sensing scenarios.

Such challenging yet common case studies require dynamic resource allocation across sensors, task prioritization and scheduling, and reliable data fusion between autonomous platforms.

This theme develops reliable computational solutions for autonomous localisation and mapping in adverse environments. The aim is to find solutions that are adaptable without a HITL and resilient to adversarial attack.

Exemplar projects include:

• “underwater inspection with multimodal imagery”;
• “mapping from aerial platforms”;
• “ground-based mapping for search and rescue”.

This theme interacts with Sensor Signal Processing for fusion challenges, and with Multi-agent Systems and Data Intelligence for multi-agent systems.