At the heart of the  architecture is the Core Engine, surrounded by a Control Hub composed of multiple Cognitive Modules. This control hub acts as the integration center where insights from different cognitive streams are fused.

 

The cognitive modules are designed for specific input streams that are analysed using data specific reasoning capabilities. For example, the telemetry data module receives data streams from IoT devices that are analysed for drift or anomalies whereas event log streams are analysed for complex event triggers, temporal causal chains or compliance violations. Other cognitive modules are designed for visual perception, spatial awareness, and health diagnostics.

​

All these cognitive modules operate semi-independently in their domain but are orchestrated by the Core Engine. The modules feed their findings into the Core Engine, which can fuse them to produce comprehensive situational awareness. Likewise, the Core Engine can distribute context information back to specific modules. For example, if the geospatial module identifies a location of interest, that geospatial context can be provided to visual perception module to refine vision processing (such as focusing attention on a particular camera near that location).

 

This interrelationship between core and modules ensures that each cognitive task is informed by others, achieving a holistic intelligence greater than the sum of its parts. The modular separation also means each module can be developed or updated independently (e.g., improving the computer vision algorithms in Visual COGs won’t directly affect how Sensor COGs process IoT data, as long as their interfaces to the Core remain consistent).