|
Spatio-Temporal Analytics - Behavioral Analysis of DVA Events and Alerts
Without a second tier of analysis beyond standard digital video analytics any system will quickly become
overrun with alerts and messages. By employing Behavioral Analytics as a secondary tier the MetroTech Net
is able to digest hundreds of thousands of alerts, extrapolate the significance within the greater context
of a corridor and provide traffic managers with actionable information.
Secondary analytics act as an expert system where fuzzy logic and probability distributions are taken into
account to determine the meaning and impact of discrete events as they occur throughout the traffic system.
Based on artificial intelligence, expert systems, autonomous agent technology, and the patented MetroTech
Decision Framework; all aspects of the behavior within complex traffic systems is digested and presented to
the TMC as patterns of activity and propensity.
This set of core services employs multi-dimensional mathematics for the purpose of recognizing the complex
relationships between regions of space and time. At the heart of this mathematical approach to dimensional
analysis is Region Connection Calculus (RCC.)
RCC is a "point free geometry" and as such treats all spatial information as regions of space with relationships
that span an infinite number of regions and sub-regions. Time too, as examined through the context of RCC, can
be described as a collection of regions, each interrelated.
Only through intricate, yet elegant, mathematical calculations can such complexity be digested into substantial
information arranged to support decisions within the TMC. As objects such as automobiles and pedestrians interact
with these regions the system is able to quickly resolve the complex nature of discrete activities into a much
larger context. It is this contextual differentiation of small, seemingly insignificant, events within the
transportation system that allows for greater awareness of possible impacts and pressures on the system as a whole.
DVA events and alerts are queued by camera systems "at the edge" for processing by the secondary analytics
system. Once consumed these events are processed by a decision framework where the greater context of the traffic
network is taken into consideration. As traffic related decisions are made by the MetroTech Net larger events
and alerts are extrapolated by the secondary behavioral analytics.
Each new event is categorized and labeled based on a discrete set of predefined rules and strategies. Once an event
is established and all associated metadata is created an event bookmark is inserted into
the video content management system via an advanced video management
service interface. As events and alerts are aggregated by the MetroTech Net, all become available to geospatial maps,
web portal applications, reporting services, other technology providers attached to the ESB, and human operators in
the TMC.
Examples of secondary analytics include, but are not limited to, IntelliSection health, projected congestion sources,
impact analysis, traffic routing recommendations, and overall impact on the corridor as a whole.
The secondary analytics engine operates as a software service accessible by other ITS components within the MetroTech
Net. These services act as multi-threaded operations which use the ESB, DVA event message queue, and an enterprise
database as sources of information for analysis.
|