Most Probable Path & Destination Prediction API (MPP&DP API)

Overview - What is MPP&DP

• MPP&DP predicts the route in the near future and the destination by giving a partial route of current ongoing trip and its context.
• MPP&DP is based on the unsupervised machine learning.
• MPP&DP extracts prediction model instance from historical car probe data.
• MPP&DP supports online API and state-less version, which means there's no state hand-over between the predictions.

Use-case examples with route/destination prediction

• Notification of dynamic road/traffic/weather condition information
• Route recommendation.
• Route optimization, for example based on gasoline or battery level.
• POI recommendation along the predicted route.
• Assistance information delivery around destination. For example, parking place availability, shopping mall coupon and so on.
• Arrival time estimation considering with current traffic information.

API specification

• Detailed REST interfaces are defined on Driver Behavior REST API

Considerations and recommendations

1. Only one prediction model per moving object (vehicle) or per driver.
2. If driver-based prediction is required, the customer must specify all of trips when generating the prediction model.
3. The interval to execute the prediction per moving object (vehicle) or per driver should be carefully considered and should be long enough, e.g. more than 60 seconds. The prediction itself is relatively costly operation among multiple CVI components.
4. Older prediction models remain on the server side as cache data when a new prediction model per moving object (vehicle) or per driver is created, therefore the customer must clear the cache by calling the API.

Architecture overview

How to use: Steps to enable and execute MPP&DP

Summarized steps

1. Model generation

   1.1. Gather car probe data.
   1.2. Optional step. Configure parameters of prediction model generation.
   1.3. Create prediction model per moving object (vehicle).
   1.4. Optional step. Query Origin/Destination (O/D) patterns per moving object (vehicle).
   

2. Prediction

   2.1 Execute prediction per moving object (vehicle)
   

Detailed steps by sequence diagram