1 / 2018-07-31 16:15:40

An IoT decision tree using a closed-loop service-oriented flow approach

IoT, Decision-Tree, Service Oriented Flow, Node, Smart Home

The representation of IoT systems as a flow of services is a popular approach as its intuitive nature is not dependent on high level programming skills. The addition and interconnection of a large number of hardware devices, represented graphically as nodes, can be easily achieved using a flow-based approach. However, this advantage can quickly make the system very complex and reduce system performance, in some cases making the overall system unstable. Flow-based tools, such as Node-Red, can be installed on low power consumption devices, such as the Raspberry Pi, and run on light-weight operating systems, such as Raspbian. Therefore, to keep performance at a high level in an IoT environment it is necessary to keep system complexity low. The solution presented here is to implement a decision tree as an independent node in the system flow. A temperature measurement system is implemented using two evaluation methods, namely, logic-based and weight-based. Experimental results show that the proposed solution reduces system complexity while maintaining system performance.