DE102005009604B4 - Method and device for generating a rating value for traffic data - Google Patents

Abstract

und der negative Berechnungswert gemäß der Formel

bestimmt werden, wobei Z(Q, t) einen momentanen Bewertungswert (Z(Q,t)) und t eine momentante Zeit (t) angibt,
und dass aus den Zwischenbewertungswerten (Wp, Wn) ein Wahrscheinlichkeitswert (W) für einen Normal- oder Störungszustand und/oder für die Zuverlässigkeit der Verkehrsdaten als der Bewertungswert (W) bestimmt wird.Method for generating a weighting value (W) for traffic data (d; d *), in which
Individual traffic data (p, sm, od, fcd) of at least one data source (Q) are provided by a plurality of data sources (P, SM, OD, FCD),
- is provided to the individual traffic data depending on the respective data source at least source inaccuracies considering source initial value (A (Q)) and
The evaluation value (W) for the traffic data (d; d *) is determined by means of at least one of the source initial values (A (Q)) provided,
characterized,
for each positive and negative traffic data each own intermediate evaluation values (Wp, Wn) are determined, wherein the positive intermediate evaluation value (Wp) according to the formula

and the negative calculation value according to the formula

where Z (Q, t) indicates a current evaluation value (Z (Q, t)) and t indicates a current time (t),
and determining, from the intermediate evaluation values (Wp, Wn), a normality or disturbance condition probability value (W) and / or the reliability of the traffic data as the evaluation value (W).

Description

The The invention relates to a method for assessing traffic data or to an apparatus for carrying out such a method.

traffic data increasingly come from a source, but from different sources as data sources. These data sources have different properties and also different reliability. The problem is that for an equal distance from the different data sources often contradictory Traffic information is output. As a result, traffic information or the corresponding traffic data only limited reliable. In addition, a certain arises unreliability for certain Types of data sources, which traffic data with a delay or with one too big Issue inaccuracy.

A first data source is formed by the police, which traffic information as traffic data outputs. The police deliver error messages indicating a section of the route, a type of fault, z. As traffic jam or stagnant traffic, a fault length, z. B. 3 km, and often one Fault. A Another data source are so-called traffic jams. There are traffic jams usually road users, which reports about a current traffic condition over a mobile phone to a call center, z. B. a call center of the ADAC or a broadcasting company. The messages will be Recorded in the Callcener and passed on for distribution to road users. The structure of the messages is similar to at the police station.

A Another exemplary data source are fixed detectors, which especially on highways, but increasingly also installed in the city are to measure the traffic. The stationary detectors act These are usually induction loops, infrared or radar detectors. By means of stationary detectors are usually speeds and vehicle counts determined. From the individual measurements, data about the Be derived traffic flow.

A Another source of data is vehicles, which are increasingly using telematics systems be equipped. The telematics systems provide so-called floating car data, which over Radio interfaces and a corresponding transmitter system to a central office forwarded for further evaluation. This "floating Car Data "(FCD) thus form with appropriate data processing another source for traffic information or traffic data. The determination in the vehicles can in particular by means of position detectors, z. B. GPS (Global Positioning System) or Gyro, and often done in conjunction with a navigation system. Such vehicles can determine the current position and track it continuously. In particular, it can also be determined whether the vehicle in a traffic jam located.

The different data sources have different characteristics. As a rule, traffic data from detectors and FCD arrive promptly and have a relatively high reliability. However, detectors are only available on a small part of the road network and often have too big Distances, about statement to make intermediate routes. FCD is currently hardly available. Both systems have as well the disadvantage of only detecting a momentary traffic flow. Police reports are nationwide, but often come late or, in the case of cancellation reports, often not at all. reports of congestion fields are irregular and are often by the personal Feeling of the reporting road user influenced.

From the EP 1 577 643 A1 a method for generating a rating data for traffic data is known in which individual traffic data from at least one data source is provided by a plurality of data sources, to the individual traffic data depending on the respective data source at least source inaccuracies considering source initial value is provided and by means of a provided of the source initial values the evaluation value for the traffic data is determined.

From the US 2003/0216857 A1 a method and a device for the modeling and processing of traffic data and traffic information is known in which in particular traffic data are collected from a plurality of sources, ordered according to importance, filtered and checked, based on these data average nominated travel times are calculated for different route sections, the individual traffic information is combined, the traffic situation is extrapolated into the future and the resulting full traffic information is made available to a user.

The Object of the present invention is a comparison for this state of the art more reliable evaluation of traffic data provide.

These The object is achieved by the method having the features of the patent claim 1 or the device with the features of claim 15 solved.

advantageous Embodiments are the subject of dependent claims.

Prefers Specifically, a method of generating a rating value for traffic data, in which individual traffic data from at least one data source be provided to the individual traffic data depending on the respective data source at least one source inaccuracies consider Direction Source initial value is provided and by means of at least one the provided source initial values the rating for the traffic data is determined.

Prefers in particular, a method in which the source initial value dependent from source-specified source reliability elevated or humiliated.

Prefers in particular, a method in which the source initial value by means of a time delay between a traffic event and the time of delivery the individual traffic data of the associated data source formed becomes.

Prefers in particular, a method in which for determining a current Rating is an elapsed time since a traffic event and / or since providing the individual traffic data is used.

Prefers in particular, a method in which at least one of these momentary Valuation value source-dependent is determined.

In particular, a method in which at least one such instantaneous evaluation value Z (Q, t) is determined in accordance with FIG Z (Q, t) = A (Q) - ((t-tm) * R (Q)) with A (Q) as the source individual initial value, t as the current time, tm as the time of the traffic event, or the time of providing the source traffic data and R (Q) as a reduction factor.

Prefers In particular, a method in which the reduction factor Message time and / or sources individually.

Prefers in particular, a method in which such a momentary Evaluation value is redetermined in the case of updated traffic data a single-channel data source.

Prefers in particular, a method in which such a momentary Evaluation value in case of updated traffic data of a multi-channel Data source determined per arrival of updated traffic data and the instantaneous scores thus determined become statistical be processed together.

Prefers in particular, a process in which too positive and negative Traffic data per own intermediate evaluation values be determined.

Prefers in particular, a method in which the intermediate evaluation values a probability value for a normal or fault condition and / or for the reliability the traffic data is determined as the evaluation value.

Prefers In particular, a method in which the probability value dependent from the difference of the intermediate evaluation values (W = abs (Wp - Wn)) is determined.

In particular, a method is preferred in which the probability value is determined as a function of the ratio of the intermediate evaluation values, in particular according to W = max (Wna, Wpa) with Wna = Wn · (1 - Wp) and Wpa = Wp · (1 - Wn) with Wp as a probability of disturbance and Wn as a probability of undisturbed traffic flow.

Prefers in particular, a method in which the data sources are formed are caused by traffic data of the police, traffic data of road users, Traffic data stationary detectors and / or automatic by Vehicles recorded traffic data.

Prefers In particular, a procedure in which the traffic data is put together with the evaluation value.

Prefers In particular, a procedure whereby the traffic data and the Rating value to be issued to a traffic information service.

Prefers is a device for performing such a method for generating an evaluation value for traffic data with at least one input interface, in particular automatic Inputting traffic data from various data sources, a processing device to perform such a method for determining at least one evaluation value for evaluating the traffic data and an output device for Outputting traffic data dependent from the evaluation value and / or together with the evaluation value.

This is possible device and method according to a data fusion of traffic reports to provide overall information on their evaluation to obtain. Considered In doing so, different types of sources are used as data sources for traffic data. Furthermore, advantageously the age of messages, i. H. your probable timeliness considered. By means of an initial reliability the source initial value is the reliability of the received traffic data dependent taken into account by the type of data source. As an intermediate calculation value becomes an initial reliability used as the instantaneous evaluation value, so that aging of the Message taken into account can be to the reliability reduce the assigned traffic data until this value Have reached zero or a defined threshold. For example can do that a linear aging function can be used.

to Calculation of trustworthiness at a given time, the current reliabilities of all messages or traffic data, the messages sorted according to positive and negative messages. From the positive and negative messages from different sources Probabilities for positive and negative error messages calculated. A trustworthiness of Overall information is z. B. as absolute difference of the two probabilities or calculated as the maximum of two derived probabilities.

The different types of data sources can also be distinguished after single-channel and multi-channel sources. For single-channel sources can a new information advantageously the previous information replace immediately, even if their aging process is still ongoing is. For multi-channel sources can several sources of information on the probability statement of individuals Traffic data of a single type of the corresponding data sources contribute.

One embodiment will be explained in more detail with reference to the drawing. It shows:

1 outlines a message chain of traffic data from various data sources to a central computer for the evaluation of traffic data.

How out 1 As can be seen, are found along a route of different types of data sources Q, which detect a current traffic flow on the route and generate and provide corresponding traffic data d. The data sources are, in particular, the police P, which transmits traffic disturbances or possibly no longer existing traffic disturbances as traffic data p via a central office to radio stations and the like for the transmission of traffic data d * to road users. Another data source Q are road users SM, which take part in the current traffic flow as so-called traffic jams and pass the current traffic flow via mobile devices and a center as traffic data sm.

Another data source Q is formed by fixed detectors OD, which are installed in the lane area of a route and detect the current traffic flow. The stationary detectors OD provide appropriate traffic data od, which is detected in a control center and used to control Ver traffic control systems are processed. Yet another data source Q form vehicles which are equipped with telematics systems and detect so-called "floating car data" fcd as a floating car data data source FCD and also transmit it to a traffic control center.

The such traffic data collected by different types of data sources Q d be a central computer Z via at least one interface I fed. The central computer Z comprises a central control device C for processing the received traffic data d and for generating or providing an evaluation value W for the traffic data d. Over a Output device O and optionally the interface device I be suitable data to particular road users, traffic control centers and broadcasters. For this data to be output it can only be the rating value W. Is possible also the use of the evaluation value W as a control variable for forwarding or outputting the traffic data d or revised traffic data d * dependent from the reliability assigned by the evaluation value W the traffic data d, d *, so that the traffic data only in case of sufficient reliability be issued to road users and traffic control centers. Particularly advantageous is the output of both traffic data d, d * and the associated or shared with them Rating value W.

To the Generating the evaluation value W comprises the central computer Z expediently a memory device MEM for storing data required for the data processing Programs, fixed or customizable parameters A (Q), R (Q), for processing the data and storing or caching received and processed traffic data d. It also gets through a timer or clock source CL provides the instantaneous time t, at the current reporting time tm incoming traffic data d and the elapsed since the arrival of the traffic data Time t - tm take into account in data processing to be able to. The individual components are by conventional means, for example a bus system, interconnected.

By the arrangement, in particular by the central control device C is a data processing or data fusion made at which of traffic data d of multiple data sources Q individual traffic information are derived in order to form a total information. This happens by overlay the different data. This can take into account that for an equal Route from different data sources Q not just the same Information but also contradictory information as traffic data d and messages M are provided. According to the method the different data sources Q automatically evaluated to to incoming traffic data d an automated individual To make evaluation. From the different individual reviews Ultimately, in addition, total information will be added provided as a single evaluation value W, which is a trustworthiness the overall information.

to Simplification of the description below all traffic data d from traffic data sources Q hereinafter referred to as messages M. The police P and the detectors OD each provide a single Message M to a specific stretch of road and a specific Time, whereby the arrival of the message M in the central computer Z is taken into account as the notification time tm becomes. These data sources P, OD are hereinafter referred to as single-channel Called data sources. In the case of floating car data fcd or by Traffic data sm of road users is in the Usually independent messages M, which come from the same data source FCD or SM, where it is for a single stretch of track and a single time or a time window around a final large number of individual messages M a plurality of road users SM or vehicles FCD can act. Accordingly, these data sources become FCD, SM below referred to as multi-channel data sources Q.

The different data sources Q different data source type can through an initial reliability therefore, for each data source type, a individual source initial value A (Q). This Source initial value A (Q), as well as others described below Values, are expediently determined and given on the basis of statistical surveys, whereby also an afterthought Adaptation of this or other such values advantageously allows becomes. The source initial values A (Q) can be used for positive messages M, d. H. Fault reports, and negative messages M, d. H. Undisturbed messages, differentiated so that positive source initial values are Ap (Q) and negative Source initial values An (Q) ready for further processing are. For the sake of simplicity, however, the following will be based on a simple embodiment only the general source initial value A (Q) is used.

In the data processing of the provided individual traffic data d or at the merger of Messages M are superimposed traffic data from the same or from different data sources Q, to provide a single total information d * with evaluation value W from it. To simplify the description, the simplest case is considered below, according to which only messages M or traffic data d from a data source Q are present, so that the overall information is identical to the message M or the traffic data d of this data source Q. With more than one message M or a plurality of traffic data, the various messages M can mutually confirm or contradict each other. This increases or decreases the trustworthiness or reliability of the overall information. Thus, a statement of the police P z. B. be confirmed by a traffic participant SM or questionable.

For example could the police P "2 km of traffic jam "reported have and already five Minutes later a road user "no Report congestion. "Important can in this context in particular the time course of Messages M be. Some data sources Q are known to be more explicit delay Report, such as the police P, when internal processes Slow down the transmission of information. This is about a appropriate adjustment of the initial Initial reliability, d. H. of the source initial value A (Q). The source initial value For example, A (Q) will be source-individual upon arrival Traffic data d set.

In addition, traffic data becomes less credible if it has been received a while ago. Advantageously, it is therefore defined in the data processing that traffic data d can age. This is taken into account by a gradual reduction of the initial reliability, ie the respective source initial value A (Q). Accordingly, a current reliability of a respective traffic data value is formed at a current time t, which is taken into account as a current evaluation value Z (Q, t) as a function of the source Z (Q, t) = A (Q) - ((t-tm) * R (Q)). (1)

According to the exemplary formula, the aging of the traffic data is mapped via a linear function in which the source-dependent current evaluation value Z (Q, tm), which initially corresponds to the source initial value A (Q), is gradually reduced. For example, a police report could have an initial reliability of 80% as traffic data p, which is taken into account by the corresponding source initial value A (p) = 0.8. This source initial value A (p) is then taken into account as a starting value for the determination of the linear aging of the reliability in the calculation of the instantaneous evaluation value Z (t), for example, a reduction per minute by the reduction factor R (p) of 5% , If the reliability or the instantaneous evaluation value Z (Q, t) has reached the value zero, then the message M is omitted, in the described example after 16 minutes. As the exemplary calculation field for illustrating various method steps and parameters according to 1 can be taken, the various factors, in particular the source individually determined value A (Q) as source reliability and the reduction factor R (Q) depending on the respective source type of the data source Q set. Considerable in determining or subsequently changing such factors may be regional aspects, types of roads involved, holiday periods or current time of day.

Especially preferred is a different calculation according to the type of message M, d. H. positive or negative message M when determining the Aging of the traffic data d, whereby again a source-dependent consideration is particularly preferred.

advantageously, the aging of traffic data d of a data source Q is interrupted, if a single-channel data source Q cancels its previous message M, z. For example, when the police P "2 km of traffic jam "reported has and to a later Time "no Traffic jam "reports In such a case, the new message M immediately replaces the old message M, where for the new message M or the new traffic data d a new aging process starts. For multi-channel sources, this can optionally be done accordingly carried out be, for example, if a message M or corresponding traffic data d be provided by one and the same traffic participant SM. This may be, for example, a first message "Totalsperrung" and a later message "route sharing" from the same road user be which by a mobile phone identification number or Mobile phone number is identifiable. In the case of unique Messages M such as "Totalsperrung" and following Messages M, such as "route clearance", however, can also the message M of another road user SM for interruption and immediate lifting of the previous one Message M or traffic data are used.

To calculate the trustworthiness as the evaluation value W for a total information becomes available Preferably, in a first step, both a probability for a disturbed traffic condition as intermediate evaluation value Wp and a probability for an undisturbed traffic condition are determined or calculated as intermediate evaluation value Wn. These result from a separate consideration of fault messages as positive messages M and free messages as negative messages M for a route section. In this case, only traffic data for messages M are expediently used, for which the instantaneous evaluation value Z (Q, t) is positive.

mit Xn(Q, t) als momentan empfangenem bzw. gültigem Datenwert der Verkehrsdaten d, der durch eine bestimmte Datenquelle einer der mehrkanaligen Datenquellen Q bzw. FCD, SM zum momentanen Zeitpunkt t gemeldet ist und gilt. Dabei ist der Datenwert Xn(Q, t) im Falle keiner Störung oder keiner vorliegenden Meldung M von Verkehrsdaten d bzw. fcd, sm gleich 1 gesetzt. Im Fall einer Störung werden die momentanen Bewertungswerte Z(Q, t) von 1 subtrahiert, um den momentan gültigen Datenwert Xn(Q, t) zu bestimmen.The instantaneous probability as a current positive intermediate evaluation value Wp (t) for a fault is determined, for example, by B. by

with Xn (Q, t) as currently received or valid data value of the traffic data d, which is reported by a particular data source of one of the multi-channel data sources Q and FCD, SM at the current time t and applies. In this case, the data value Xn (Q, t) is set equal to 1 in the case of no disturbance or no message M of traffic data d or fcd, sm. In the event of a fault, the current evaluation values Z (Q, t) are subtracted from 1 to determine the currently valid data value Xn (Q, t).

Transferable or, alternatively, this procedure is of Formula 2 for determining current data values Xn (Q, t) for a plurality different traffic data of different data sources Q.

mit Xp(Q, t) im Falle einer Störung oder vorliegenden Meldung M von Verkehrsdaten d bzw. fcd, sm gleich 1 gesetzt. Im Fall keiner Störung werden die momentanen Bewertungswerte Z(Q, t) von 1 subtrahiert, um den momentan gültigen Datenwert Xp(Q, t) zu bestimmen.The instantaneous probability as a negative intermediate evaluation value Wn (t) for a trouble-free state or normal state is determined, for example, by B. by

with Xp (Q, t) in the event of a fault or present message M of traffic data d or fcd, sm is set equal to 1. In the case of no disturbance, the current evaluation values Z (Q, t) are subtracted from 1 to determine the currently valid data value Xp (Q, t).

Transferable or, alternatively, this procedure is of Formula 2 for determining current data values Xn (Q, t) for a plurality different traffic data of different data sources Q.

The Value ranges for the intermediate evaluation values Wp, Wn to positive and negative, respectively Traffic data d are each in the range 0 <= Wp, Wn <= 1.

Of the larger of the both intermediate evaluation values Wp (t) and Wn (t) give an indication to the presence of a fault, if the intermediate positive evaluation value Wp is greater than the negative intermediate evaluation value Wn is. Otherwise, an indication of a normal or undisturbed Traffic condition if the negative intermediate score Wn greater than is the positive intermediate evaluation value Wp. Such a statement stands however in connection with the trustworthiness of the Data, d. H. it is relevant if the trustworthiness the traffic data d big enough is.

The trustworthiness as a common evaluation value W is advantageously derived from the two probabilities, ie the two intermediate evaluation values Wp, Wn. According to a first exemplary possibility, this can be done by calculating the absolute difference according to W = abs (Wp - Wn). (6)

Another exemplary possibility for determining the common evaluation value W as trustworthiness results from a direct inclusion of the respective other probability or the respectively other intermediate evaluation value according to FIG Wna = Wn * (1-Wp), Wpa = Wp * (1-Wn) and W = max (Wna), Wpa). (7)

The Trustworthiness achievable by such calculations gives the score W between 0 and 100%. The bigger the Rating value W for the trustworthiness is, the more reliable is the information about the traffic situation. Dependent from the evaluation value W, according to a particularly preferred embodiment the publication of traffic data d or modified traffic data d * to road users or traffic control centers or repressed become. Possible is additional or alternatively the publication of traffic data d, d * together with the rating value W or, if appropriate, other of the determined ones Intermediate values.

According to this Method or additional Procedures are by simple modifications, the considerations übergeord neter Route systems possible, so that from a locally determined traffic situation to a traffic situation at a more distant stretch or place to the same or a later date can be closed. In addition, for disturbance lengths, causes of failure, etc. appropriate Fusion operations, such as weighted averaging, carried out become.

Claims (15)

Method for generating a weighting value (W) for traffic data (d; d *), in which - individual traffic data (p, sm, od, fcd) are provided from at least one data source (Q) from a plurality of data sources (P, SM, OD, FCD) be provided - to the individual traffic data depending on the respective data source at least source inaccuracies considering source initial value (A (Q)) and - by means of at least one of the provided source initial values (A (Q)) the evaluation value (W) for the Traffic data (d; d *) is determined, characterized in that positive and negative traffic data are each determined their own intermediate evaluation values (Wp, Wn), wherein the positive intermediate evaluation value (Wp) according to the formula

and the negative calculation value according to the formula

where Z (Q, t) denotes a momentary evaluation value (Z (Q, t)) and t indicates a momentary time (t), and that from the intermediate evaluation values (Wp, Wn) a probability value (W) for a normal value or fault condition and / or reliability of the traffic data is determined as the evaluation value (W). Method according to claim 1, wherein the source initial value (A (Q)) source-individually. Method according to any preceding claim, in which the source initial value (A (Q)) by means of a time delay (R (Q)) between a traffic event and the time (tm) of providing the individual traffic data (d) of the associated data source (Q) formed becomes. Method according to any preceding claim, in which for determining the current evaluation value (Z (Q, t)) an elapsed one Time (t - tm) since a traffic event and / or since providing the individual Traffic data (d) is used. The method of claim 4, wherein at least one such instantaneous evaluation value (Z (Q, t)) is determined depending on the source becomes. Method according to Claim 4 or 5, in which at least one such instantaneous evaluation value (Z (Q, t)) is determined in accordance with Z (Q, t) = A (Q) - ((t-tm) * R (Q)) with A (Q) as the source individual initial value, t as the current time, tm as the time of the traffic event, or the time of providing the source traffic data (d) and R (Q) as a reduction factor. The method of claim 6, wherein the reduction factor (R (Q)) message and time and / or sources individually set becomes. Method according to one of claims 5 to 7, wherein such current evaluation value (Z (Q, t)) is redetermined in the case of updated Traffic data (p, od) of a single-channel data source (P, OD). Method according to one of claims 5 to 8, wherein such current evaluation value (Z (Q, t)) in the case of updated traffic data (fcd, sm) of a multi-channel data source (FCD, SM) per arrival updated traffic data is determined and so determined current evaluation values (Z (Q, t)) are statistically processed with each other become. The method of claim 1, wherein the probability value (W) dependent from the difference of the intermediate evaluation values (W = abs (Wp - Wn)) is determined. The method of claim 1, wherein the probability value (W) dependent from the relationship the intermediate rating values (Wp, Wn), in particular according to W = max (Wna, Wpa) with Wna = Wn · (1 - Wp) and Wpa = Wp · (1 - Wn) with Wp as a probability of a disturbance and Wn as a probability for one undisturbed Traffic flow. A method according to any preceding claim, wherein the data sources (Q) are formed by traffic data (p) the police (P) and / or traffic data (sm) of road users (SM) and / or traffic data (od) stationary detectors (od) and / or automatic traffic data (fcd) collected by vehicles (FCD). A method according to any preceding claim, wherein the traffic data (d *) is output together with the evaluation value (W) become. The method of claim 13, wherein the traffic data (d *) and the evaluation value (W) to a traffic information service be issued. Apparatus for carrying out a method according to any preceding claim for generating an evaluation value (W) for traffic data (d; d *) comprising: - at least one interface (I) for inputting, in particular automatically inputting traffic data (d) of different data sources (Q); a processing device (C) for carrying out such a method for determining at least one evaluation value (W) for evaluating the traffic data (w) and - an output device (I, O) for outputting traffic data (d *) depending on the evaluation value (W) and / or together with the score (W).