Predictive hierarchical human augmented map generation for itinerary perception

Abstract

A new class of augmented map application is introduced which can provide detailed knowledge about any area, to a user. This brief particularly focuses on obtaining itinerary perception subject to different environmental conditions. This refers to extraction of traffic related information from an augmented map. The problem is modelled as a machine learning technique where the traffic distribution at different times (including same days, different days and different weather) are observed continuously using a service robot. This data is posed as a Gaussian process for post-estimation. Our system consists of a vision sensor which will acquire the region of interest input, queried to a database of traffic density distributions, learned from the scenes at different points of time. The user interacting with the system will obtain an information pertaining to the region conditioned on environmental and timing events.