Sensor-based systems form the bases of many modern computing and smart infrastructures, such as IoT, Smart Cities, Precision Agriculture, Industry 4.0, Logistics, and Data-centric Engineering etc. Established in 2000, the Adaptive Emergent Systems Engineering (AESE) group aims to embrace the many research challenges relating to the ability to get such systems to scale, remain robust when confronted with failure, maximise their usefulness while minimising or sharing resources, be energy neutral etc. Such systems bring new demands, therefore we examine the behaviours of the collective; convergence, or purposeful lack of convergence, anarchical systems, systems that understand and match context etc. Our aim is to not only expose what these behaviours are, but why they happen. In doing so, we aim to harness this understanding, bringing together models of computation and non-computing systems (e.g. social behaviours, economics etc.) to build safer, more robust and resilient sensor-based systems. This work informs our more speculative work on smart dust in terms of understanding mass densities, role compositions, resource scavenging, collaboration and self-configuration.