**8. Discussion**

As pointed out in the challenges section, communication and cooperation constitute the main challenges in the future of IoT. The importance of communication can be observed in two main areas: device-to-device and human-to-device. In the first case, we identify a research opportunity for autonomous intelligent agents to overcome the complexity of the resulting network. Regarding the second case, which involves human-to-device communication, we clearly see the need to deeply explore the understanding of human language, with emphasis on human commands. As the number of devices available to every person grows, a concise way to perform tasks involving several devices is more important. Recent advances in tools such as ChatGPT show the relevance of the convergence between the IoT and Natural Language Processing (NLP).

### **9. Conclusions**

In this chapter, we have presented the Swarm computing vision, a decentralized and self-adaptive approach to overcome the limitations of cloud-centric architecture for the IoT. We presented the principles that have driven the conception of the Swarm and summarized the main challenges to achieving its realization: communication and cooperation of devices, the inclusion of resource-constrained devices, better interfaces for human-interaction, and the complex nature of the network. We also proposed an initial architecture, whose main component is the Broker, a communication mediator that aims to solve the interoperability of devices in the Swarm network. Besides, we listed a selection of technologies that enabled our implementation and described an application example that illustrates the potential of the Swarm network. Our advances in coping with the Swarm challenges can be summarized as follows. We developed four Broker implementations, using different programming languages: C, Lua, Java, and Elixir. We developed a minimum Broker implementation. We adopted open Web semantic technologies that facilitate device communication; we implemented a mechanism of semantic service discovery as a starting point for cooperation; and we implemented a CoAP-HTTP proxy that leverages transparent communication with resource-constrained devices with minimum impact. Additional effort is needed in all fronts of Swarm challenges to concretize the vision.

*Swarm Computing: The Emergence of a Collective Artificial Intelligence at the Edge… DOI: http://dx.doi.org/10.5772/intechopen.110907*
