JPH10134295A - Jam discrimination method for road traffic condition - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method with which jamming can be surely discriminated on road traffic conditions with a little data processing cost and a little measuring technique cost while reducing a time loss as much as possible without being affected by traffic conditions. SOLUTION: A predictive sector terminating value is periodically calculated from traffic volume detected at the starting edge of a sector, the average value of speed, the length of sector, the timewise distribution accept value of automobile within a measuring interval and the speed over accept value of sector passing automobile, and periodically compared with the average value of traffic volume at the ending edge of sector, differential traffic volume is detected and added over cycles, the number of vehicles left in the sector to be monitored is continuously detected and when that number exceeds a critical value, the announcement of jamming is triggered.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for identifying traffic congestion in a road traffic situation in one sector to be monitored, which passes through one measurement cross-section at each of the beginning and end of a sector. The number and speed of vehicles that have been detected are continuously detected as measurement data, collected during a finite measurement interval that is continuously numbered, and periodically packed against the average value of traffic volume and speed,
An evaluation is then made, each measurement cross-section encompassing all traffic lanes that can pass in one driving direction.
The present invention relates to a method for identifying traffic congestion in road traffic conditions.

[0002]

BACKGROUND OF THE INVENTION Due to the ever increasing road traffic situation, on multi-lane roads, flow controllers with switchable traffic indicators are being used more and more frequently for precise control of the flow. This control device uses various methods for identifying traffic congestion in road traffic conditions. The purpose of the introduction of this type of equipment and the development of an associated traffic jam identification method are as follows. That is, the danger of a rear-end collision can be avoided more reliably by switching the traffic display including a warning that enables the subsequent traffic vehicle following the congested point. Of particular importance here is the identification of traffic jams as accurately and quickly as possible, ie a reduction of the critical time between the occurrence of traffic jams and the implementation of safety measures.

The following measurement methods are usually used for detecting the current traffic volume. That is, a measurement method of recording measurement data of each vehicle in a so-called measurement cross section linked to the land is used. Normally,
Monitoring for each individual sector, i.e. for each traffic line segment, delimited by the respective start and end measurement cross sections. Technically, the level of "traffic volume" is defined as, for example, the number of detected vehicles per time unit, and from this, for example, "the number of vehicles per minute" is formed as a traffic volume unit.

[0004] The method of the type described at the beginning is disclosed in the known document "Automatische Stoerfallerkennung auf Autobahnen".
mit Hilfe vonFuzzy-Logik; Simens AG, Muenchen, Bere
ich Anlagentechnik, GeschaeftsgebietStrassenverkeh
rstechnik ".
It consists of two main parts: data processing and fuzzy decision logic. 1 for each data processing part
The average value of the speed and the traffic volume averaged over the two finite measurement intervals is substantially expressed as a reference amount for congestion identification.
Two indicators (indicators) are calculated: "speed-density-difference (VK-Diff)", "trend coefficient", and "traffic trend". These three indices are used as input variables to the fuzzy decision logic, from which congestion probabilities are supplied as output parameters. The fuzzy decision logic has a three-stage structure, which consists of three modules: a "group identification" module, a "congestion preliminary investigation" module, and an original core module (so-called "congestion identification" module). This core module supplies as output parameters the so-called congestion probabilities indicated by 0-100%.

[0005] Due to the definition of the very complex but rather expedient indicator "speed-density-difference", this known method ensures that there is no delay only in completely determined traffic conditions. Can not work. In particular, when traffic congestion occurs in the area behind the sector to be monitored, the reliability of traffic congestion identification is greatly reduced. In addition, even if fuzzy decision logic is used as the decision system for the combination of indices used, the basic inadequacy of the above method cannot be compensated and only a slight improvement is made. is there.

Another generally known method is to derive the detection of congestion substantially from the stagnation situation of the measuring cross-section, that is to say that the vehicle is already stopped on the measuring cross-section or at very low speed. There is a method of identifying traffic congestion when passing. However, this so-called localized method has further disadvantages.
That is, a valuable time loss until the detection of the traffic congestion occurs depending on the distance between the traffic congestion point and the measurement cross section. In particular, when the sector to be monitored is very long (the purpose is to reduce the number of measurement cross sections for cost reasons), the point of occurrence of congestion immediately before the measurement cross section and the end of the formed stagnation A very long time is wasted between when the part reaches the next upstream measuring cross section. This is because the vehicle approaching the rear end of the stagnation from behind cannot receive any warning during the dead time caused by this system. As a result, at the rear end of the stagnation, the risk of a rear-end collision becomes extremely high as a chain phenomenon of the original traffic jam.

[0007] The known document "AV Traffic Information Technic, Inc., Aachen, Germany"
E-die Komplettloesung fuer modernes Verkersmanag
ement ", it is also known a method for identifying traffic congestion in road traffic situations based on the detection and analysis of traffic parameters for sections. In this method, based on the so-called" pattern identification "method, individual methods detected in the measurement cross section are known. The vehicle signals are evaluated, characteristic pattern characteristics are extracted, analyzed and normalized, and the measurement data (pattern characteristics) processed in this way are assigned to the next upstream measurement cross-section, respectively. Is transmitted to the current interval station, where it is compared with the measured data analyzed in the first measurement cross-section, and by using appropriate correlation techniques a so-called interval system function is formed from which Should be observed in the section between the entrance measurement cross section and the exit measurement cross section. The section between the traffic parameters "travel time" (and its average cruising speed) and the inlet-side measurement cross section and an outlet side measuring cross sections according to the section of the vehicle Groove "traffic density" can be derived.

In any case, this type of method has a very high demand for measurement technology and data transmission capacity, so that besides the data transmission between the section station corresponding to the respective measurement cross section and the central evaluation station, Direct interconnection between adjacent leg stations is required.

Currently used measuring, transmitting and evaluating devices for implementing the method according to the prior art usually do not fulfill these hardware prerequisites, thus making it possible to use the method for such sections. This requires very expensive post equipment. In addition, all techniques currently in use or still in development for traffic data detection (ie for local limited speed, traffic, traffic structure detection) include (substantially guided Loops, microwave sensors (radar), laser scanners and relatively large uncertainties (indicated by subsequent evaluation of the recorded data using a video camera and a special computer program). In this case, in particular, the detection of the private vehicle depends on the error. In practice, individual vehicles are not often detected, for example, at every lane change in the measurement cross-section area. It is also conceivable for the detection device to be defective in the traffic lane or in the measuring cross section. For these reasons, satisfactory results are not obtained with methods based on traffic volume detection for sections where the reliability of vehicle detection is important.

In a known traffic jam identification method, an average value which is supplied to a computer center and obtained by periodic packing from detected measurement data is evaluated by real-time data processing. However, this data situation is uncertain due to vehicle detection, where up to 100% reliability is absolutely undesirable.

There are various supplementary means for making decisions under uncertainty or based on uncertain data. Among them, fuzzy logic, signal processing, and a static method according to the prior art are used in addition to the methods described above in the traffic technology.

A disadvantage of signal processing in which exponential smoothing is generally performed is that parameters must be supplied and set for each measurement cross section and for all individual data. As a result, the cost associated with parameter processing increases. In contrast, the static method has the following disadvantages. In other words, individual judgments lose significance due to the fundamental lack of collectiveness.

[0013]

SUMMARY OF THE INVENTION An object of the present invention is to minimize the time loss without being affected by traffic conditions.
It is an object of the present invention to provide a method for reliably identifying traffic jams in road traffic situations with low data processing costs and low measurement technical costs.

[0014]

SUMMARY OF THE INVENTION The object of the present invention is to address the average of traffic and vehicle speed detected at the beginning of a sector, the length of a sector, and the temporal variance of vehicles within a measurement interval. A predicted value of traffic volume at the end of the sector is periodically calculated from the acceptance value and the acceptance value for the vehicle speed passage at the time of passing through the sector. A) The length of the sector and the acceptance value for the vehicle speed passage at the time of passing the sector. B) detecting, from said transit time, the value of the measurement interval at which the vehicle detected at the beginning of the sector at each measurement interval passes through the measurement cross-section at the end of the sector, c) the time of the vehicle Considering the acceptance value for the strategic allocation,
Determining an allocation factor for the distribution of the average value of the traffic volume for the measurement intervals detected according to step b); d) calculating the predicted value of the traffic volume for each measurement interval by the allocation factor and the corresponding traffic volume In this case, the sum is calculated over all preceding measurement intervals and the current measurement interval, and the predicted value of the traffic volume is calculated based on the measurement data detected at the measurement cross section at the end of the sector. , And periodically compares it with the average value of the traffic volume at the end of the sector, detects the respective difference traffic volume, performs addition over the cycle of the difference traffic value, and adds the difference traffic value to the sector to be monitored. The problem is solved by continuously detecting the number of remaining vehicles and, if this value exceeds a limit value for the number of vehicles remaining in the sector, to trigger a traffic jam notification.

According to the continuous dynamic evaluation of vehicles additionally entering the sector to be monitored, which is carried out according to the invention, a very reliable congestion independent of the respective traffic situation Identification is achieved. In non-congested traffic situations, the traffic flowing through the monitoring sector
It is detected again in consideration of the average speed, the running characteristics, the length of the sector, and the dispersion state on the exit side of the sector.

In calculating the traffic forecast for vehicles passing through the rear end of the sector, the method of the present invention determines which portion of the fleet of vehicles measured during the detection cycle currently detects the monitored sector. It is taken into account whether or not to leave only after the next detection cycle or a subsequent detection cycle, not a cycle. For this purpose, a calculation of the distribution coefficient is performed. By using this allocation coefficient, the vehicle group of one detection cycle detected at the beginning of the sector is allocated to two cycles predicted to be likely to leave the sector again.

The measurement data of the measurement cross section at the beginning of the sector and the measurement data of the measurement cross section at the rear end of the sector are observed simultaneously and independently of the sector topology, and are themselves constructed under a complicated analysis standard. The drawback of the relatively unreliable traffic jam identification method is that
It is reliably avoided by the method according to the invention. In this case, the data processing costs of the method according to the invention are kept very low, which is the level which can be compared most easily with the traffic jam identification method based on locally limited measurements. However, its major disadvantage, namely that traffic jams can only be identified based on the existing fact of the stagnant situation of the measuring cross section, is avoided in the present invention. This is because possible dynamic congestion can be detected quickly and almost independently of the position in the monitoring sector.

The calculation and analysis of traffic parameters for sections, as known from the prior art, requires extremely high data processing costs for correcting the start and end vehicle groups through sector identification. Compared to the method, the method according to the invention significantly reduces this cost. There is no need for complex indicators, such as velocity-density-difference (VK-Diff) or special tendency coefficients. This applies at the same time to the analysis of the measured data for possible vehicle groups, which in the case of section-related methods can be erroneous without their detection and appropriate consideration.

For example, three indices, "speed-density-difference",
In comparison with known methods based on “trend coefficient” and “traffic trend”, it is apparent that the method according to the present invention is
It has been found that reliable and fast traffic congestion identification can be achieved. Further, according to the present invention, it is possible to identify traffic congestion very reliably even if the sector is long, and the result is hardly affected by the position of congestion in the sector. This cannot be obtained by the known method.

The application of the method according to the invention is also very economical. This is because a direct additional connection between the section stations assigned to the measuring cross-section, which is indispensable in particular under a method based on traffic parameters for the section, can be omitted. Because of the high reliability of traffic congestion, the distance between two measurement cross sections,
That is, the length of the sector can be made very long.

[0021]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The calculation of the number of additional vehicles remaining in the monitoring sector (this is the basis for congestion notification) in the method of the present invention will be described in detail with the following equations.

1. Calculation of Transit Time The transit time _{TDiz of the} vehicle required to pass through the monitoring sector is calculated by the following equation.

TD _iz = S _i / v _iz v _iz = (1 / S _i ) * ∫ _{s = o ... siv} (s) ds In this case, the TD _iz is the transit time of the vehicle required to pass through the sector, The v (s) is the detected speed of the vehicle at the location s in the sector, the v _iz is the speed averaged over the interval of the sector S _i , and the S _i is the measurement cross section i and i + 1 The length of the inter-sector, s is a length variable (s∈0 ... S _i ).

In addition, i is the index of the measurement cross section that increases in the traveling direction, and z is the index of the detection cycle that increases each time it elapses (current cycle = 1).

For v (s), it is necessary to apply an acceptable value that takes into account the current traffic conditions and the section topology. For example, in a case where a vehicle with weak horsepower has a steep ascending section, a decrease in speed accompanying the passage must be taken into account. The same applies to slope sections and sections with sharp curves. V to traffic conditions
The dependence of (s) is determined by the measured speed vm _iz . The section topology is taken into account by the correction factor. Therefore, it is as follows.

Vm _iz = average value of the vehicle speed detected in the measurement cross section i during the detection cycle z Therefore, for example, the vehicle speed is constant, and the topology of the section does not particularly change (curve, climbing gradient, inclination) No)
In the case, v (s) = const = _vmiz is accepted. Other acceptable values are possible.

2. Distribution of Detected Vehicles During the detection cycle z, the number of vehicles in the measurement cross section i is counted. From the transit time TD _iz ,
When a vehicle that enters the sector at measurement cross section i during the current period cycle,
It is required because it can be predicted at one point. Separate observations of time detection must determine which vehicle groups leave the sector during which detection cycle. Allocation coefficient n
_iz is the current cycle z = 1 and the preceding cycle z> 1
Formed for

N _iz = Max (0, (z−TD _iz / T))
When TD _iz <z * T, n _iz = Max (0, (1-Max (0, (z-TD _iz /
T))); TD _iz ≧ z * T, where n _iz is the allocation factor belonging to the current inspection cycle (this of all vehicles entering the sector during the current inspection cycle). The _assigned vehicle leaves the sector in the same cycle z again at the measurement cross section i + 1 after the passage of the transit time _TDiz ), where T is the length of the detection cycle.

3. Calculation of predicted value of traffic volume The traffic volume is calculated from the number of counted vehicles, an acceptance value relating to the temporal distribution of vehicles detected over the detection cycle T, and the calculated vehicle distribution coefficient. Can be calculated.

The predicted value of the traffic volume is formed by the following equation.

[0031] In this case, the PQ _{i + 1.1} is the measurement cross section i +
1, the current predicted traffic volume (z = 1), Q
_iz is the number of vehicles counted in the measurement cross section i during the inspection cycle z, and P (n _iz ) is the distribution in the vehicle Q _iz for which the distribution is to be accepted. This is the vehicle exiting at measurement cross section i + 1 during inspection cycle z (eg, P (n _iz ) =
n _iz ).

4. Calculation of differential traffic volume The differential traffic volume DQ _iz is obtained from the following equation.

DQ _iz = PQ _iz -Q _{iz In} this case, the DQ _iz is the differential traffic volume at the measurement cross section i during the detection period z, and the PQ _iz is the traffic volume at the measurement cross section i during the detection period z. The predicted value, Q _iz, is the number of vehicles counted in the measurement cross section i during the detection period z.

[0034] 5. The number BDQ _i of the vehicle remaining in an additional in the detection monitoring sector residual vehicle is determined by the following equation.

[0035] Alternatively, recursively, BDQ _i = BDQ _i.alt + DQ _iz where BDQ _i is the measurement cross section i−1
The number of vehicles in the sector between i and i, said DQ _iz
The differential traffic volume of the measurement cross section i in the detection cycle z, where Z is the number of cycles in which the addition is performed.

The addition of measurement errors is avoided in a recursive way by extension to zero (negative vehicle numbers can only be due to start characteristics or the detection itself) or by damping.

If the parameter BDQ _i exceeds a predetermined limit (for example, depending on traffic conditions), a traffic jam notification is triggered. For this congestion notification for obtaining the running characteristics to be adapted to each case to a subsequent vehicle, depending on the degree of congestion, i.e. it can be different depending on the level of BDQ _i.

According to an advantageous embodiment of the method according to the invention, the calculation of the predicted value of the traffic volume is carried out separately for each driving lane in one driving direction, and for all driving lanes in one driving direction. A periodic comparison is performed between the sum of the predicted values and the value of the traffic detected in the measurement cross section at the end of the sector. The reliability of the predicted value is increased by a separate observation of this traffic lane. This is because accurate driving characteristics are accepted for each driving lane and time distribution of vehicles is performed. The individual prediction for each driving lane with the respective special driving characteristics (right driving lane; truck-only lane; left driving lane; overtaking lane, etc.) and the subsequent summation are summation before prediction (this is information loss). Leads to)
Gives better results than

According to another advantageous embodiment of the method according to the invention, fuzzy logic is applied for triggering the traffic jam notification. In addition to the number of vehicles remaining in the sector, at least one input parameter describing the traffic situation is used.

In this connection, traffic conditions are defined as vehicle speed (possibly over a plurality of cycles), traffic volume, traffic density (= traffic volume / speed), standard deviation of speed (standard for unstable traffic flow). (As a measure), topological or meteorological parameters. However, in this case, when defining the input parameters representing the traffic conditions, it is not always necessary to access all the above-mentioned parameters. The advantage of taking traffic conditions into account when triggering a traffic jam notification is that the method is significantly more reliable. According to this approach, it is very easy to adapt the method according to the invention to various application locations and use conditions. This is because time-consuming setting and adjustment of the fixed limit value can be omitted.

The same reliability improvement effect can be achieved by another embodiment of the present invention in which the above-mentioned uncertainty factors relating to the determination of the error detection point in the measurement cross section are considered by the fuzzy logic, and the congestion notification trigger is generated. It is also obtained when accepted at the time.

According to a further advantageous embodiment of the invention, an output parameter is provided which represents the type of traffic jam case by means of fuzzy logic.

According to this, various traffic congestion notifications are output according to the degree of traffic congestion (for example, "slow driving", "complete vehicle stasis", etc.). As a result, an effective warning of the amount of traffic to be moved in the congested area is issued, and a request for running suited to each is made.

Particularly preferably, the course of the vehicle speed as it passes through the sector is described using fuzzy logic.

In this way, uncertain factors which can be described when detecting the predicted value of the traffic volume at the end of the sector (for example, the vehicle driver can determine his or her own activity depending on the traffic situation in each case based on the driving style. How to use up etc.) is advantageously considered.

According to another advantageous embodiment of the invention, when the proportion of trucks in the total traffic exceeds a predetermined limit value, only the traffic and the speed of the truck are used as measurement data for the congestion identification. To evaluate.

Here, from experience, it is based on the fact that the speed of the truck is clearly less variable than the speed of the private vehicle. As a result, the reliability of traffic congestion identification by this method is further improved. EU-for trucks to be specified in the future
Speed controller (this is up to 88 km / h
The positive effect is expected to increase as described above. Furthermore, the truck itself is much more reliably detectable than a private vehicle even using current common detection techniques. At a point in time when the ratio of trucks in the total traffic volume is below the limit required for reliable traffic jam identification, detection of total vehicles, that is, switching to detection of all vehicles including private vehicles and trucks, is advantageous.

More advantageously, the measurement data detected at the end of one sector is simultaneously used as the measurement data at the start of the sector of the succeeding sector.

According to this, monitoring of one sector is extended to a complete monitoring system for all lanes composed of a plurality of sectors at a simple and very low cost. In this case, each measurement cross section represents the measurement cross section at the end of the sector of sector i and at the same time as sector i + 1.
Also represents the measurement cross section at the beginning of the sector.

Claims (8)

[Claims] 1. A method for identifying traffic congestion in a road traffic situation in one sector to be monitored, wherein the number and speed of vehicles passing through the measurement cross section at one measurement cross section at each of a sector start end and a sector end. Are continuously detected as measurement data, collected during a finite measurement interval that is continuously numbered, periodically packed against the average value of traffic volume and speed, and then evaluated, Each measurement cross section includes all the driving lanes that can pass in one driving direction. In the method for identifying traffic congestion in a road traffic situation, the average value of the traffic volume and the vehicle speed detected at the start of the sector, The traffic volume at the end of the sector from the length of the vehicle, the acceptance value for the time variance of the vehicle within the measurement interval, and the acceptance value for the vehicle speed course when passing through the sector The prediction values are calculated periodically, a) the transit time is determined from the length of the sector and the acceptance value for the passage of the vehicle speed during the transit of the sector, b) from the transit time at the beginning of the sector at the respective measurement interval Detecting the value of the measurement interval at which the detected vehicle passes through the measurement cross section at the end of the sector; c) taking into account the acceptance value for the time distribution of the vehicle,
Determining an allocation factor for the distribution of the average value of the traffic volume for the measurement intervals detected according to step b); d) calculating the predicted value of the traffic volume for each measurement interval by the allocation factor and the corresponding traffic volume In this case, the sum is calculated over all preceding measurement intervals and the current measurement interval, and the predicted value of the traffic volume is calculated based on the measurement data detected at the measurement cross section at the end of the sector. , And periodically compares the average value of the traffic volume at the end of the sector with the average value of the traffic volume at the end of the sector, detects the respective differential traffic volume, performs addition over the cycle of the differential traffic volume value, and adds Road traffic conditions characterized by continuously detecting the number of vehicles remaining in the sector and, if this value exceeds a limit value for the number of vehicles remaining in the sector, triggering a traffic jam notification. Bitterness in Delinquency identification method. 2. The calculation of the predicted value of the traffic volume is performed separately for each traveling lane in one traveling direction, and the sum of the predicted values of all traveling lanes in one traveling direction and the measurement cross section at the end of the sector. 2. The method for identifying traffic congestion in a road traffic situation according to claim 1, wherein a periodic comparison is made with a value of the traffic amount detected in the step (a). 3. Applying fuzzy logic to the triggering of the congestion notification, wherein the fuzzy logic additionally uses at least one input parameter representing the traffic condition in addition to the number of vehicles remaining in the sector. The method for identifying traffic congestion in road traffic conditions according to claim 1 or 2. 4. The method according to claim 3, wherein the uncertain factors which can be described are taken into account by means of an error detection in the measuring cross section under fuzzy logic. 5. The method for identifying traffic congestion in a road traffic situation according to claim 3, wherein an output parameter representing a type of traffic congestion is supplied by the fuzzy logic. 6. The method according to claim 1, wherein the progress of the vehicle speed when passing through the sector is described using fuzzy logic.
Congestion identification method in road traffic conditions described in section. 7. When the ratio of trucks in the total traffic volume exceeds a predetermined limit value, only the traffic volume and speed of the trucks are evaluated as congestion identification as measurement data.
The method for identifying traffic congestion in a road traffic situation according to claim 1. 8. The traffic congestion in a road traffic situation according to claim 1, wherein the measurement data detected at the sector end of one sector is simultaneously used as the measurement data at the sector start of the succeeding sector. Identification method.