Customizing driving behavior insights

Driving Behavior plug-in framework

The following diagram outlines the components of the Driving Behavior plug-in framework, which is responsible for driving behavior analysis. The framework accommodates customizations to the default Driving Behavior plug in so that you can tailor the data analysis for your business needs.

Data flow

The following diagram shows an overview of the data flow, where VDH = Vehicle Data Hub, DMM = Dynamic Map Manager, HBase = Hadoop database, and HDFS = Hadoop Distributed File System.

Logical flow and plug-in points

The logical flow of the analysis and plug-in points is noted in the following diagram.

Driving Behavior plug-in analysis core logic flow

Descriptions of each part of the logic flow

Descriptions of the plug-in points, which are shown in the diagram, are provided in the following table.

No	Name in logic diagram	Interface	Description
1	Data interpolation plug-in loader	com.ibm.iot.automotive.moma.drb. operations.plugin.IDataInterpolator	Plug-in point to interpolate customer-specific data.
2	Trajectory summary builder	com.ibm.iot.automotive.moma.drb. operations.plugin.ISummaryBuilder	Plug-in point to insert a summary of a trajectory.
3	Context splitter plug-in loader	com.ibm.iot.automotive.moma.drb. operations.plugin.IContextSplitter	Plug-in point to split a trajectory into multiple subtrajectories (context injection). This method is called multiple times if a trajectory is already split into subtrajectories.
4	Data annotation plug-in loader	com.ibm.iot.automotive.moma.drb. operations.plugin.IAnnotator	Plug-in point to insert annotations for each GPS point.
5	Behavior detector plug-in loader	com.ibm.iot.automotive.moma.drb. operations.plugin.IBehaviorDetector	Plug-in point to detect behavior.
6	Whole trip behavior detector plug-in loader	com.ibm.iot.automotive.moma.drb. operations.plugin.IWholeTripBehaviorDetector	Plug-in point to detect behavior, such as fatigued driving, for the whole trip.

Default plug-in

Descriptions of the plug-in points for the default plug-in are provided in the following table.

No	Plug-in interface	Default plug-in	Dependent plug-in	Description
1	Data interpolation plug-in loader	com.ibm.iot.automotive.moma.drb.operations. inerpolator.plugins.DefaultWeatherDataInterpolator	(none)
2	Trajectory summary builder	com.ibm.iot.automotive.moma.drb.operations. summarybuilder.plugins.DefaultTrajectorySummaryBuilder	(none)	The following metrics are calculated: time_span distance average_speed max_speed idle_time night_driving_time rush_hour_driving day_of_week day_of_month month_of_year
3	Context splitter plug-in loader	com.ibm.iot.automotive.moma.drb.operations. context.plugins.DefaultByRoadTypeSplitter	(none)	context_category= roadType,context_category_id=2
4		com.ibm.iot.automotive.moma.drb.operations. context.plugins.DefaultBySpeedPatternSplitter	(none)	context_category= roadType,context_category_id=3
5		com.ibm.iot.automotive.moma.drb.operations. context.plugins.DefaultByTimeRangeSplitter	(none)	context_category= speedPattern,context_category_id=0
6		com.ibm.iot.automotive.moma.drb.operations. context.plugins.DefaultByWeatherTypeSplitter	DefaultWeatherDataInterpolator	context_category= weatherType,context_category_id= 10
7	Data annotation plug-in loader	com.ibm.iot.automotive.moma.drb.operations. annotator.plugins.DefaultHeadingPatternAnnotator	(none)
8		com.ibm.iot.automotive.moma.drb.operations. annotator.plugins.DefaultSpeedActionAnnotator	(none)
9		com.ibm.iot.automotive.moma.drb.operations. annotator.plugins.DefaultSpeedingAnnotator	(none)
10	Behavior detector plug-in loader	com.ibm.iot.automotive.moma.drb.operations. behavior.plugins.DefaultAccelStopBehaviorDetector	DefaultSpeedActionAnnotator	behavior_id=1,behavior_name=Harsh acceleration behavior_id=2,behavior_name=Harsh braking/li> behavior_id=4,behavior_name=Frequent stops behavior_id=5,behavior_name=Frequent acceleration behavior_id=6,behavior_name=Frequent braking
11		com.ibm.iot.automotive.moma.drb.operations. behavior.plugins.DefaultSpeedingBehaviorDetector	DefaultSpeedActionAnnotator	behavior_id=3,behavior_name=Speeding
12		com.ibm.iot.automotive.moma.drb.operations. behavior.plugins.DefaultTurnBehaviorDetector	DefaultHeadingPatternAnnotator	behavior_id=7,behavior_name=Sharp turn behavior_id=8,behavior_name=Acceleration before turn ehavior_id=9,behavior_name=Over-braking before exiting turn
13	Whole trip behavior detector plug-in loader	com.ibm.iot.automotive.moma.drb.operations. behavior.plugins.DefaultFatiguedBehaviorDetector	(none)	behavior_id=10,behavior_name=Fatigued driving

Default plug-in for edge behavior

Descriptions of the plug-in points for the default plug-in for edge behavior are provided in the following table.

No	Plug-in interface	Default plug-in	Dependent plug-in	Description
1	Data interpolation plug-in Llader	com.ibm.iot.automotive.moma.drb.hybrid. inerpolator.MetaSystemMissingDataInterpolatorWithDMMRest	(none)	Car probe data interpolator uses the DMM REST API. Specify the following configuration to access the Dynamic Map Manager (DMM) as Property in DrbCustomizableConfig.xml. name=dmmUrl, value={Your DMM REST API URL} name=tenantId, value={Your tenantId (for example “public”)} name=internalTenantId, value={Your internalTenantId (for example “1”)} name=userId, value={Your REST basic authentication userId} name=password, value{Your REST basic authentication password}
2	Data interpolation plug-in loader	com.ibm.iot.automotive.moma.drb.hybrid. inerpolator.MetaSystemMissingDataInterpolatorWithDMMDataBusClient	(none)	Car probe data interpolator uses the DMM Data Bus API. Specify the following configuration to access DMM (Dynamic Map Manager) as Property in DrbCustomizableConfig.xml. name=tenantId, value={Your tenantId (for example “public”)} name=requiredAttributeList, value={Your necessary link properties by comma-separated (for example, “all”)} name=distanceThreshold, value={Your required distance by meter unit (for example 50)}
3	Trajectory summary builder	com.ibm.iot.automotive.moma.drb.hybrid. summarybuilder.ExtendedTrajectorySummaryBuilder	DefaultTrajectorySummaryBuilder	In addition to the metrics for the default plug-in, the following metrics are calculated for edge behavior. travel_time_roadtype_N travel_distance_roadtype_N
4	Data annotation plug-in loader	com.ibm.iot.automotive.moma.drb.hybrid. annotator.SpeedActionAnnotatorForEdge	DefaultSpeedActionAnnotator	If (name=ignoreInterpolated, value=true) is specified in DrbCustomizableConfig.xml as Property, which is interpolated data is ignored for annotation.
5		com.ibm.iot.automotive.moma.drb.hybrid. annotator.SpeedingAnnotatorForEdge	DefaultSpeedingAnnotator	If (name=ignoreInterpolated, value=true) is specified in DrbCustomizableConfig.xml as Property, which is interpolated data is ignored for annotation.
6		com.ibm.iot.automotive.moma.drb.hybrid. annotator.SteeringEventAnnotator	(none)	Specify name=acceleration_y_threashold and value as m/s^2 unit for setting y-axis acceleration threshold in DrbCustomizableConfig.xml.
7	Behavior detector plug-in loader	com.ibm.iot.automotive.moma.drb.hybrid. behavior.AbruptSteeringDetector		behavior_id=1000,behavior_name=Abrupt steering
8	Behavior detector plug-in loader	com.ibm.iot.automotive.moma.drb.hybrid. behavior.CorneringBehavior		behavior_id=1001,behavior_name=Cornering event Specify name=intervalThreshold,value={milliseconds order (for example) 1000} for determining whether steeringEvent comes with this interval threshold. Specify name=continuousCount,value={count (for example) 2} for determining whether steeringEvent continues greater than or equal to this threshold.
9	Whole trip behavior detector plug-in loader	com.ibm.iot.automotive.moma.drb.hybrid. behavior.GeoJsonWriterForDebug	(none)	GeoJson writer for debug purpose. Specify name= basePath and value of directory where you want to output GeoJson for debug in DrbCustomizableConfig.xml.

Developing the plug-in

For information on developing the plug-in, see Developing custom Driver Behavior plug-ins.

Data model

The data model for the Driving Behavior plug-in framework, which is comprised of a Hadoop database (HBase) model and Db2 database model, is defined in Driving Behavior plug-in data model.