How do AI and 'collaborative caching' assist in disaster response?

AI and collaborative caching (also known as cooperative caching) assist in disaster response by maintaining the flow of critical real-time information when traditional communication infrastructure, such as telecom towers and power lines, is damaged or destroyed.

The Mechanism of Collaborative Caching

In a collaborative caching system, diverse network components—including satellites, drones (UAVs), ground-based stations, and emergency vehicles—work together as nodes to store and share data. This approach addresses the limitations of individual communication channels:

• Overcoming Latency: While satellites can cover wide areas, they often suffer from data latency (delay); collaborative caching allows rescue teams to retrieve data from the nearest available node rather than a distant original source, significantly reducing delays.

The Role of AI in Optimization

Because nodes like drones and rescue vehicles have limited storage capacity and are constantly moving, researchers developed AI models to automate and optimise caching decisions:

• Contextual Multi-Armed Bandit (CMAB): This AI model reviews data recency, current demand, and available memory to make split-second storage decisions. For instance, it can automatically determine that a photo taken 10 minutes ago is more valuable to rescuers than one taken an hour ago, or that caching a short text alert is more efficient than storing high-resolution 4K video that would exhaust limited cache space.

Space Air Ground Integrated Network (SAGIN)

These technologies operate within a three-tier framework called SAGIN, which integrates space, air, and ground layers. By using AI caching to bridge the gaps between these layers, the system efficiently delivers "actionable data"—such as updated maps showing usable bridges or live videos identifying urgent rescue needs—directly to responders in the field.