Mondrian Hierarchical Clustering for Identifying the Optimal Location of Charging Infrastructure Based on Telemetry Data

摘要

The present invention presents a method by which one may use vehicle telemetry data to make determinations about where electric vehicle charging infrastructure should be placed. The proposed strategy stems from the idea that areas with high-concentrations of vehicle activity, including trip endpoints, represent attractive locations for placement of charging infrastructure. This invention uses an unsupervised method for generating random forests of hierarchical clusters called the Mondrian Process. It generalizes the Mondrian by defining a density form constructed from the counting measure on the manifold defined by the input telemetry data.