Inrix acquires sensor data—occupancy and speed measurements for road segments—from Department of Transportation sensor networks and other public and private sources. Additionally, our predictive model uses “metadata”—attributes that we would expect to influence the observed sensor data both now and some time in the future—including traffic incidents, road construction schedules, current and forecast weather, school schedules, sporting and entertainment event schedules, and other locale specific variables, such as the legislative calendar in Washington, D.C.
To make a prediction of an event we need to determine all factors that combine to result in that event occurring, as well as how those factors both individually and collectively influence the outcome of the event. In the context of traffic, obvious factors which influence the nature of flow at a given time include the day of the week, the prevailing weather, the occurrence of an accident or some other incident, sporting or other types of event, road construction and more surprisingly school schedules. Inrix views predictive and causal relationships in the context of the effect that one variable has on the probability distribution of outcomes of the other. The natural way to evaluate and combine such probabilistic relationships is in the framework of Bayesian statistics, which Inrix uses as the basis for our models.
To control complexity in the modeling process, Inrix aggregates sensor data to the proprietary Inrix Sensor Group level—a unique and highly logical grouping of sensors which comprises an independent stretch of road of interest, independent of underlying choice of map data vendor. For each sensor group Inrix maintains a large number of local attributes which we use as inputs to the predictive modeling process, in conjunction with a series of metadata variables.
Bayesian Network example of predicting future traffic flow in the San Francisco Bay Area.
Highlighted in blue are the Inrix Sensor Groups in which predicted traffic speed is causally dependent upon current traffic.
We provide easy integration of Inrix Traffic Services through an XML interface that enables flexible delivery to your customers via SMS text alerts, e-mails, Web, cell phones, satellite and terrestrial radio, TV, personal and in-car navigation devices, and other devices. Our data structure is designed for applications such as speed maps, expected congestion clear times, traffic alerts, and highly personalized traffic and navigation services.
