Improving Yield of Energy Resources - AI Driven Approach

Опубликовано 1 августа 2018, 17:16

As we are building smart energy infrastructure that includes various forms of hybrid energy resources including electric vehicles, solar, battery storage etc., it firstly requires data to be collected from these disparate energy sources and then analysed for solving interesting problems such as demand response, preventive maintenance for the equipments, peer-to-peer energy sharing etc. Most of the existing systems operate by collecting data at the edge gateway and transmitting and analysing it at the cloud. But, due to intermittency of network infrastructure and latency in decision making, such an approach is not viable for low latency control actions. This talk presents machine learning (ML) approach and results for solving some of these problems at the edge constrained devices.

See more at microsoft.com/en-us/research/v...