(since the last release of Connected Vehicles Acceleration 1.2.1. on May 9, 2016)
July 25 2016, release of Connected Vehicles Accelerator 1.3.0. Full release notes are included in the download, but some highlights:
- Added new shift and crew functionality. Now trips can be added to shifts and crews assigned to these shifts for filling various roles during trip execution. Shift and Crew have been added to the 2990/2991 services in the railway case for testing purposes.
- Removed BusinessWorks from the product stack for the accelerator. Replaced all BW functionality with new StreamBase Event Flow.
- Added a new bus scenario to show shift and crew functionality. Added new datasets and UI skin to display the bus functionality.
- Added a new Rostering page to the bus context to allow for changing trip assignments to blocks and shifts, as well as cancelling entire trips. Also allow for assigning vehicles to blocks and crews to shifts.
- Fixed an issue with block delays not being handled correctly when other delay types are present.
Traditionally, transportation companies relied on routes, schedules, work assignments and other isolated systems to model their business. Much of the data is historical, making it difficult or impossible to predict future state. Plus, with the data in silos there is no overall holistic view of what's going on across the entire network. Stale, batch-oriented feeds mean that the data is in the wrong place at the wrong time, degrading its value. Getting the data to the right people is also a challenge. Backwards-facing data means that exceptions are always surprises and handling them is always a reactive process often resulting in sub-optimal outcomes.
In the modern world of Internet of Things (IoT), vehicles have become mobile devices, leading to the Internet of Trains, Boats, or Airplanes. These new information sources provide an opportunity to increase the available operational intelligence, both quantity and quality. Of course the data volume increase can be both a benefit and a hindrance if you can't find the signal in the noise. But the clever use of smart event processing technology and predictive analytics allows you to cut through the clutter to find the events that matter. Now with forward-looking data, exceptions can be proactively handled with the best possible outcome, for the company, and its customers and partners, improving their experience. Plus it opens up new avenues to monetize the value of the data through real-time APIs that can be exposed and marketed to third parties.
Benefits and Business Value
Connected Vehicles platform acts as a single source of truth for all trip and vehicle data. Using an in-memory model exposed using integration services the data is available to any system that needs it, reducing the need for data silos. As a real-time data repository, it is fed directly by data streams from vehicles and systems, so the information is guaranteed to be timely and accurate.
The business rules are primarily configuration-driven which allows decision table changes to be deployed in hours rather than weeks. This means a more agile system, able to adapt to business needs quicker and more effectively. By detecting anomalies and sending alerts, the accelerator acts as an efficient and fast first check on network health. It decides when something needs operations input and alerts them quickly and effectively. Using the real-time operation dashboard, operations staff has visual confirmation of network health at a glance, helping them quickly identify critical business moments.
Better operational intelligence with predictive capability means a single view of resources, all updated in real-time, with more timely and more accurate data. The net result is a more agile business, able to react on both the micro and macro scale more effectively.
The platform deployment is naturally scalable giving better data distribution and the ability to meet growth targets and beyond. The event-based architecture and in-memory network model support large scale deployments both on premise and in the cloud. Exposing this data using APIs empowers employees, customers, and partners.
At the heart of the Connected Vehicles Accelerator is the Trip. This is a journey consisting of several stops operating on a schedule. There are three resources that a trip depends on: Vehicle, Crew, and Passengers/Cargo. Plus the Trip also has a dependency on the Processes that make them happen.
The Connected Vehicles Accelerator captures data from existing systems, and combines it with real-time feeds from these resources and processes. In addition, it can capture real-time feeds from third party data providers such as weather and traffic. Accelerator rules analyse this data and produce automated actions, advisories to operations staff, and alerts to outside parties. The current state of the network is displayed in true real-time on an operations dashboard, and near real-time using analytics tools.
By aggregating all this information in one place, the accelerator gives unique insight into network operations that just is not available in any other single system.
The accelerator includes demos showing several scenarios. The first is for a passenger railway operating in the Netherlands called Virtual Train. The second is for a port operator called Virtual Port, tracking arrivals of ships into the Port of Rotterdam. Finally, the third is for a regional airline with a hub at Manchester in the UK called Virtual Air.
A simulator is used in place of actual vehicles, publishing data directly into the accelerator environment. This includes information about vehicle speed, direction, distance, and position, as well as occupancy.
At the heart of the accelerator is the Network Model which is implemented as an in-memory model in ActiveSpaces, exposed using services in BusinessWorks. It is based on the GTFS (General Transit Feed Specification) for trips and routing, and extended further with Reference data for schedules and vehicles. This is the base static reference data for the system, which is used by the Event Manager to model the transportation network.
The Event Manager is implemented using BusinessEvents. It uses the static GTFS and Reference data stored in the Network Model, combined with dynamic data feeds that arrive as events. These events are called Reports. Combining the static and dynamic data, it builds a core data model in memory that tracks the existing state of the network.
Through rules and decision tables it analyses the core model and produces outbound Notification events. These rules and decision tables do things like classify occupancy, determine proximity to a defined point, determine if a vehicle has not departed a stop when it should have, or determine if a vehicle is moving slower than it should be. It also calculates and estimates arrival and departure times at subsequent stops based on this information, making a best guess as to when the vehicle will arrive and depart using the information at hand.
Additionally, the Event Manager responds to external Alerts indicating service alterations that cause delays and cancellations. It evaluates the impact of these on operating and planned trips, and can produce additional Alerts for impacted trips in response to the original Alert.
The Operational Dashboard is implemented using Live Datamart and it captures the current state of the network from the Event Manager notifications. It displays this information on a fully-interactive, HTML5 application. This displays trips on a map, plus gives information about stops, progress, occupancy, block sequencing, and alerts.
Underlying all the components is a service bus, implemented using Enterprise Message Service and BusinessWorks. This provides the connectivity between components, and with other systems and the vehicles.
TIBCO software products and versions used:
TIBCO Enterprise Message Service
TIBCO BusinessEvents Standard
|TIBCO BusinessEvents Data Modeler||5.3.0|
|TIBCO BusinessEvents Decision Manager||5.3.0|
|TIBCO BusinessEvents Event Stream Processor||5.3.0|
|TIBCO Live Datamart||7.6.4|
|TIBCO LiveView Desktop||2.1.4|
|TIBCO LiveView Web||1.1.1|