Traffic Data Model

From Navit's Wiki
Jump to: navigation, search

The data model Navit uses for traffic messages is based on the TraFF format, a universal traffic feed format.

Not every consideration that went into TraFF may apply to Navit—for example, one use case is conversion of TMC messages into the feed format without access to map data, which is not a scenario one would encounter with Navit (map data is available, and reports outside the map coverage area are safe to ignore).

Some parts of the specification may mention “data users”, “consumers” or “applications”—all of these refer to Navit in this context.

As traffic support is still under development, some aspects may not have been implemented yet.

Traffic Data Model.png

Message[edit]

A message is the atomic element of traffic information, referring to a particular condition at a given location. Examples:

  • 1 km of stationary traffic on the A8/A9 before Barriera Milano Nord in northwestbound direction.
  • L198 closed between Lech and Warth for at least the next day.
  • Heavy snowfall in the area of Vilnius.

Every message is identified by an alphanumeric identifier which incorporates the identifier of the source from which the message originated. No two messages with the same identifier can exist at the same time.

Field Description
id An identifier, which remains stable over the entire lifecycle of the message. The colon (:) is a reserved character to separate different levels of source identifiers from each other and from the local message identifier.
receiveTime A timestamp indicating at what time the message was first received by the source. This attribute should be kept stable across all updates.
updateTime A timestamp indicating at what time the last update to this message was received by the source.
expirationTime A timestamp indicating when the message will expire if not updated. An expired message should be deleted or no longer be considered current.

If end_time is specified and is longer than expiration_time, end_time should be used to govern expiration and this attribute ignored.

The expiration time governs how consumers will treat a message if they do not receive any further updates, e.g. because they lose the data connection. Sources should therefore choose the expiration time carefully: if a situation is not expected to resolve in a certain period of time, the expiration time should not be shorter. Conversely, if a situation is likely to have resolved within a certain period of time, the expiration time should not be longer than that.

startTime A timestamp indicating when the condition is expected to begin.
endTime A timestamp indicating how long the condition is expected to last.
isCancellation If true, this message is a cancellation message, indicating that existing messages with the same ID should be deleted or no longer considered current. All other attributes of a cancellation message should be ignored.
isForecast If false, the message describes a current situation. If true, it describes an expected situation in the future.
replaces Identifiers of other messages which are replaced by this message, e.g. if multiple messages are being merged into one. The value of the current message’s id attribute is never included here, as any existing messages with the same identifier will be replaced anyway.

Event[edit]

An event refers to a condition, its cause or its effect. Examples:

  • Queuing traffic
  • Road closure
  • Accident

A message may contain multiple events. For example, if a road is closed due to an accident, the message would report events for an accident and a road closure.

Similar events may be grouped into event classes, for example:

  • Congestions: This class comprises various events for different types of congestion, such as heavy traffic, slow traffic or stationary traffic.
  • Hazards: This class includes events such as objects on the road, nearby fires, unprotected accident areas or black ice.
  • Restrictions: This includes closed lanes, complete road closures as well as temporary weight or size limits.

For a list of event classes and events, see Traffic Event List#Events.

Field Description
class The event class (generic category).
type The event type, which uniquely identifies the event and can be mapped to a string to be displayed to the user.
length The length of the affected route in meters.
speed The speed at which vehicles can expect to pass through the affected stretch of road. This can either be a temporary speed limit or the average speed in practice, as implied by the event or whichever is less.
quantifier Additional quantifier for events allowing this. Permissible data types and their meanings depend on the event type.

Supplementary Information[edit]

Supplementary information is extra information which can be added to an event. Multiple supplementary information items can be added to the same event. They are qualifiers, not events in their own right. Examples:

  • Vehicle types to which a restriction applies
  • Refinements to the position (e.g. in a tunnel, in the left lane)
  • Instructions to drivers (e.g. avoid the area, approach with care)

For a list of permitted values, see Traffic Event List#Supplementary Information.

Field Description
class The supplementary information class (generic category).
type The supplementary information type, which can be mapped to a string to be displayed to the user.
quantifier Additional quantifier for supplementary information types allowing this. Permissible data types and their meanings depend on the type.

Quantifier[edit]

Quantifiers can be used with events and with supplementary information. The syntax is the same in both cases.

DurationQuantifier[edit]

Specifies a duration, e.g. a delay.

IntQuantifier[edit]

An integer, e.g. number of distinct events, number of available parking spaces.

SpeedQuantifier[edit]

A speed that does not refer to moving traffic (e.g. wind, queue length increase/decrease). Note that the speed of moving traffic is expressed through the speed field.

TimeQuantifier[edit]

A timestamp, e.g. expected end of a condition.

Location[edit]

Encapsulates the location data for the message.

Locations can occur in two different shapes:

  • Point location: These indicate that the message refers to a single point on the road, indicated by the at field. Values for from and/or to fields can also be specified (one of the two must be non-NULL if the location is unidirectional) and serve just as an indication of direction if at is specified.
  • Linear location: These indicate that the message refers to a stretch of road, starting at the location indicated by from and ending at the location indicated by to (if the location is bidirectional, the same goes for the opposite direction). The at field must be NULL.

Locations are thus represented by one or more coordinates. However, coordinates come with several constraints:

  • Low accuracy: Some formats use a single point to represent multiple adjacent points in reality. For example, TMC generally represents motorway junctions as a single point, whereas physically they are really four distinct points (entry and exit in both directions).
  • Different map data: The encoding and decoding end may work with different maps, on which coordinates for the same feature may differ slightly. The same issue arises when the two sides use different versions of the same map: changed road layouts may be reflected in one map but not in the other.

TMC presents an additional challenge for ramps and connecting carriageways at motorway junctions: these are represented by a single coordinate pair (that of the junction). The fact that the message refers to the ramps rather than the motorway, as well as the ramps affected, must be inferred from event and/or supplementary information codes. The direction is expressed through the direction of the motorway.

For this reason, extra information should be included where applicable and available:

  • Road class
  • Road names and numbers
  • Junction names and numbers
  • Carriageways affected, if the message does not refer to the main carriageway

When matching a location to a map, data users should treat the above information as hints, not as hard requirements, as different map data on both ends may cause differences:

  • Maps may disagree on the road class, or the road class may have changed
  • Road names may be spelled differently (Via Rossi vs. Via Marco Rossi vs. Via Rossi Marco vs. Via privata Marco Rossi; Kowalskiego vs. ulica Kowalskiego; Jono Jonaičio gatvė vs. J. Jonaičio g.)
  • Road numbers may differ in delimiters (A4 vs. A 4 vs. A-4), case differences (SS342DIR vs. SS342dir), use of prefixes (DK 8 vs. 8; FFB 11 vs. K FFB 11) or even different prefixes (A 8 vs. BAB 8); some sources even invent their own road numbers (TMC for local roads in Germany).

These possibilities need to be taken into account. Rather than opting for a “match or no match” decision, data users should establish criteria for ranking matches based on their quality.

Compass directions such as “northbound” are ambiguous for some stretches of this road
Linear features, as well as points located on linear features, additionally have a directionality indicating the direction of travel to which the report applies. Directionality is expressed by an attribute marking the location as unidirectional or bidirectional. For unidirectional locations, the direction can be expressed in one of the following ways:
  • Linear features have an inherent directionality, expressed by ordering the points in the order in which they are passed (i.e. from the end of the queue to the obstruction).
  • For points located on linear features, the same can be achieved by two auxiliary points on the linear feature, one situated before, the other after the target location (as seen in the direction of travel). One of these points may coincide with the target location. If map data is available, this can be achieved by picking two points adjacent to the target location on the line feature. When translating TMC events, an offset in either direction can be used.
  • A compass direction, such as “northbound”. However, this may be ambiguous in situations as the one shown in the illustration: the point lies on a stretch of the road whose direction is contrary to the general direction of the road. What is southbound with respect to the whole road is northbound with respect to the stretch between the two adjacent turns, and vice versa. On ring roads, the concept of a compass direction has only local relevance. Global designators would be “clockwise” and “counterclockwise”, although this may be similarly difficult to establish in practice.
  • A destination on the road, such as “towards Chur” or “towards Bellinzona”. While this is easy to turn into a human-readable indication, it is difficult to translate into a direction on a map. This is only possible with map data to translate the destinations into coordinates, and unless the road in question is tagged with a matching destination, the destination indication suffers from similar ambiguities as a compass direction.

Therefore, it is advisable to rely on coordinates for directionality wherever possible. A destination or a compass direction may be added for easier representation in prose. In any case, compass directions should be avoided in conflict situations where a road briefly turns against the main direction of travel.

A location currently cannot span multiple roads, i.e. at least one of the following must be true:

  • the road number remains unchanged throughout the location,
  • the road name remains unchanged throughout the location,
  • the stretch of road to which the location refers must not connect to any other road of the same or a higher category.

Some sections are shared by multiple roads and bear multiple road numbers. They are considered as belonging to either road and may be referenced by either number. A location may at the same time span a shared section and no more than one of the roads it belongs to, and reference them with the respective road number.

Field Description
at If non-NULL, indicates a point location, i.e. the message refers to a single point on the road, indicated by this field.
from The starting point of the location, i.e. the coordinates at which the driver would first encounter the condition reported.
to The end point of the location, i.e. the coordinates at which the driver would encounter the end of the condition reported.
destination A destination, preferably the one given on road signs, indicating that the message applies only to traffic going in that direction (e.g. “Milano”, “München”, “Vilnius”). Illegal for bidirectional locations.
direction A compass direction indicating the direction of travel to which this message applies (e.g. “N”, “SE”). Discouraged for ring roads (including partial ring roads) or for sections which significantly deviate from the principal direction of the main road; illegal for bidirectional locations.
directionality Indicated whether the message is unidirectional (valid only for the indicated direction) of bidirectional (valid for both).
fuzziness
  • LOW_RES: Locations are constrained to a predefined table; the actual extent of the condition may be shorter than indicated.
  • END_UNKNOWN: The end of the condition (where drivers leave the affected stretch) is unknown, as is typical for a report by a driver who has just encountered the end of a traffic jam.
  • START_UNKNOWN: The start of the condition (where drivers would first encounter it) is unknown, as is typical for a report by a driver who has just passed an obstruction.
  • EXTENT_UNKNOWN: It is unknown where the condition begins and ends, as is typical for a driver who is in the middle of a traffic jam.
ramps
  • ALL_RAMPS: Affects the entry and exit ramps only.
  • ENTRY_RAMP: Affects the entry ramp only.
  • EXIT_RAMP: Affects the exit ramp only.
  • NONE: No ramps are affected, i.e. the message refers to the carriageways of the main road.

Any value other than NONE implies that only the specified ramps are affected while the main road is not. In that case, the road* fields refer to the main road served by the ramp, not the ramp itself. This is mainly intended for compatibility with TMC, where junctions with all their ramps are represented by a single point. Other sources should use coordinate pairs instead.

roadClass The importance of the road within the road network. Represented as a Navit item type; values for motorway, trunk, primary, secondary, tertiary and a generic value for other types are allowed.
roadName A road name (e.g. “Savanorių prospektas”, “Viale Certosa”, “Wasserburger Landstraße”). Do not use this attribute if the road name is not consistent throughout the entire location.
roadRef A road number (e.g. SP526). Do not use this attribute if the road number is not consistent throughout the entire location.
tmcTable For messages received via TMC, the country identifier (CID) and location table number (LTN or TABCD) for the location table to be used for location lookup. The CID is the decimal number stored in the COUNTRIES and LOCATIONDATASETS tables, not the hexadecimal code from the PI (referred to as CCD in TMC).
tmcDirection For messages received via TMC, the direction of the road to which this message applies (positive or negative). Ignored for bidirectional messages.

Point[edit]

A point represents a point on the road, which can either be a point location or an endpoint of a linear location. It specifies a coordinate pair and can optionally be supplemented with a junction name and/or number where applicable.

Field Description
lat The latitude of this point.
lon The longitude of this point.
junctionName The name of a motorway junction (e.g. “Lainate”).
junctionRef The number of a motorway junction (e.g. 42).
tmcId The TMC identifier of the point, for locations obtained via TMC. The following formats are allowed:
  • LCID (e.g. 12345): Refers to the point identified by the LCID in the corresponding location table, typically used with the primary location.
  • LCID with offset (e.g. 12345+2, 12345-1): Refers to a point that is the given number of “steps” from the point with the indicated LCID, in the direction specified. This is typically used with the secondary location.