IoT Autonomous Agents Powered by Blockchain Technology

Traditional Internet of Things (IoT) systems rely on Cloud or centralised systems for decision making and storage, resulting in an additional layer in the threat model [1]. Having a single point of failure, denial of service (DoS) attacks and third-party data manipulation are just some of the issues that are present in such systems. Blockchain is offered as a solution to these issues in IoT infrastructures focussing on data immutability, data access permission and device authentication [2,3]. Current implementations normally rely on a centralised entity to make state changes to IoT devices such as consuming a RESTful service to instruct IoT devices to perform another action [4,5], which comes with the same downsides as using any centralised architecture.
Autonomous Agents are pieces of software that act and can function without any human intervention by reacting to states and events in their respective environment [6]. The rise in blockchain technology has stimulated renewed interest in Autonomous Agent-oriented systems [7,8]. An example of an autonomous agent is a computer virus [9]. Viruses almost take on a 'life' of their own as once released, they can continue to infect machines other than the original. Autonomous Agents can also work in conjunction with each other [10], often called a multi-agent system. Centralised architecture is hostile towards autonomous agents as downtime and cybersecurity threats disrupt autonomy of systems. A Blockchain network can act as a communication layer between IoT devices and has the potential to promote total autonomy between devices enabling true peer-to-peer communication.
This project aims to determine the viability of decentralised decision-making in this context, and discover the challenges arising in the areas of ethics and technological limitations. Our use case is Agri-tech; connected devices used in farms at various scales. We will use NUFarms' state of the art Agri-Tech test-bed to gather data from relevant IoT devices e.g. location, optical, electro-chemical, mechanical, dielectric soil moisture, air flow, mobile apps, crop management systems (e.g., semios & arable) and a range of digital farm management programmes. A decentralised system will be designed and developed utilising the Autonomous Agents comments to address the security and privacy issues of such systems. The testbed will simulate a full smart farming cycle including Pre-planting, Cultivation, Growing, Harvesting, Storage, Processing, Wholescale marketing, Retail marketing, and Consumption.

References
1. Fernandez-Carames, et al., 2018. A Review on the Use of Blockchain for the Internet of Things. IEEE Access
2. Kshetri, 2017. Can Blockchain Strengthen the Internet of Things?. IT Professional
3. Conoscenti, et al., 2016. Blockchain for the Internet of Things: A systematic literature review. IEEE AICCSA
4. Hang, et al., 2019. Design and Implementation of an Integrated IoT Blockchain Platform for Sensing Data Integrity. Sensors
5. Cui, et al., 2019. IoT Data Management and Lineage Traceability: A Blockchain-based Solution. IEEE/ICCC
6. Lieberman, 1997, March. Autonomous interface agents. ACM CHI
7. Afanasyev, et al., 2019. Towards Blockchain-based Multi-Agent Robotic Systems: Analysis, Classification and Applications
8. Calvaresi, et al., 2018. Multi-Agent Systems & Blockchain: Results from a Systematic Literature Review. Advances in Practical App of Agents
9. Buterin, V., 2014. DAOs, DACs, DAs and More: An Incomplete Terminology Guide. [online] Blog.ethereum.org.
10. Singh et al., 2015. Autonomous Agent Based Load Balancing Algorithm in Cloud Computing. Procedia Computer Science