To address the challenges of slow response times and less than ideal energy saving improvements in neighborhood energy systems, this research develops an IoT edge smart node featuring built-in AI functions and introduces a combined energy saving method that uses collaborative learning and real-time updating techniques. The hardware part of the node, to put it simply, uses various kinds of sensor devices and a simplified design approach to gather and process power usage information locally. For the algorithm side, a kind of collaborative learning arrangement helps train models across different locations, working together with an online updating machine to predict short-term energy needs and make automatic adjustments. Testing results indicate several findings: 1) the edge node responds in about 25.1 milliseconds on average, which is much faster than cloud-based alternatives; 2) the overall system helps save around 14 percent of energy, with even better savings during high-demand periods; 3) the prediction accuracy for energy loads shows moderate error levels; and 4) importantly, the collaborative approach reduces data transmission needs significantly. This work offers practical methods and evidence for creating neighborhood energy systems that work efficiently, operate securely, and respond quickly to changes.
Li et al. (Thu,) studied this question.