EP2261876B1 - Device and method for switching signal programs - Google Patents

Abstract

The method involves assigning switching variants by a temporal-local sequence of switching signal programs (SPU) for junctions (Kn). The switching variants are evaluated with respect to traffic-technical evaluation criteria, where efficiencies of the relevant signal group resulting from the switching signal programs of the switching variant are considered as evaluation criterion. The switching variant is implemented traffic-technically with respect to the evaluation. An independent claim is also included for a device for switching signal programs for light signal-controlled junctions of a road network.

Description

The invention relates to a method for switching signal programs according to the preamble of claim 1 and to an apparatus for carrying out the method according to claim 10.

It is known to control the traffic flow at junctions of a road network, such as at intersections, junctions or other road locations, by means of traffic lights. A traffic signal system has signal transmitters for emitting light signals to road users and a control device for switching on the signal transmitters according to a signal program. One or more light signals, which always indicate the same signal states, are combined into signal groups. In the signal program, the signal times of the traffic signal system are defined in terms of duration and assignment. The signal program is represented graphically by a signal time plan on a time scale, called a signal plan for short.

Different traffic conditions require signal programs with different release time distributions or orbital times and offset times. The change between two signal programs is called switching. The switching must be done at a time at which the transition makes sense in traffic terms, that is, it should be realized with the least possible disruption of the flow of traffic and be achieved with reasonable control engineering effort. The decision to switch from a current signal program to another signal program is made on the basis of preset switching times or on the basis of currently determined measurement data.

From the patent US 5,257,194 is a traffic control method known in which the control method and define the phase scheme and set request switches in each operating cycle. In daytime-dependent control methods, signal time parameters are also defined in each cycle. A common cycle length and a planned offset are calculated at the local master control unit. The offset error is measured and used along with the calculated cycle length to adjust the local signal times.

Annex F of the "Guidelines for Traffic Lights RiLSA", published in 1992 by the Road and Transportation Research Association, introduces basic switching principles described. There is a distinction between direct switching, switching with service life, switching without defined switching time and switching with Umschaltsignalprogramm. When switching over with service life, the currently switched signal program is first processed until the next changeover time. Then the signal program is deactivated in such a way that the last signaling state is maintained. During this time, the reference time of the new signal program to be switched continues to run until it has reached the next switchover point. Between the switching time of the old signal program and that of the new, a service life has thus formed. In order to avoid device-related and traffic-technical problems, when converting to concrete applications, it should be noted that the switching times within or outside the release times should be in coordination direction, but not within a transitional period. To minimize the planning effort and the computer-related time required for switching in larger networks switchover signal programs should be provided only in exceptional cases.

The layout DE 25 12 912 discloses a method for changing the signal plan in road traffic signal systems, which starts from predetermined switching points with the same signal phase, which may differ greatly from each other due to the different lengths of the individual round trip times. In order to ensure a smooth course of the signal plan change and the road traffic is to be avoided over 10% beyond stretching the signal phase phases with certainty. This is achieved by checking the temporal assignment of the switching points and deriving therefrom an instruction for shortening or lengthening the signal phase in question. The timeline of a signal schedule is shifted at a switching instant abruptly relative to the timeline of the other signal plan, causing at least one of the time axes to be asynchronized prior to switching. The incoming switching request causes the switchover from a signal plan in the next switching point the other. After switching, the switching points synchronize the time axes again.

From the manual for basic supply of a control unit SITRAFFIC C800V, published on 15.12.2003 by Siemens AG, from chapter 4.6.5 a switching and synchronizing method called stretch is known, after which it is first to check in the new signal plan whether there is a time in the program circulation where the signal image matches the current signal image. If this is the case, then this time is set immediately and you are in the new signal plan. Synchronism with the reference time is established by upsetting or stretching the signal plan over several round trips if necessary. If no identical signal image is found, the old signal plan is executed until an identical signal image is found or until the signal plan reaches the most favorable switching time. There will be changed in any case in the cheapest switching time of the new signal plan.

In the known switching method, however, only a single node is considered isolated. There is thus no vote on the switching process at the adjacent nodes instead.

The invention is therefore based on the object to provide a method of the type mentioned, with which can minimize disturbing effects on the flow of traffic in the signal program change at several adjacent nodes.

The object is achieved by a method having the features of patent claim 1. Several switching variants are specified, wherein a switching variant is determined by a time-local sequence of switching signal programs for the nodes. The predetermined switching variants are evaluated on the basis of one or more traffic-technical evaluation criteria, with regard to the evaluation traffic-favorable switching variant is executed. Regardless of what the majority of switching variants is given, can be determined by taking into account quantitative and / or qualitative traffic assessment criteria, the cheapest switching variant. If coordinated nodes are involved, the mutual location of the nodal points and a progression speed on a link train connecting the nodal points can be taken into account in the specification of the switchover variants.

In an advantageous embodiment of the switching method according to the invention, switchover signal programs with service life are determined for each node, which correspond to the respectively possible mutual positions of the switchover times of the currently switched signal program and the signal program to be switched, wherein switchover variants of temporal-local orders combine switchover signal programs of the node points. The switching signal programs are composed of portions of the respective currently switched signal program and the signal to be newly switched. The ratio of the two components is dependent on the mutual position of the switching times in the considered Umschaltumlauf, whereby the respective life - which can also be zero - is determined. For the length of the service life there is a periodicity which corresponds to the smallest common multiple of the circulation times of the currently switched signal program and the signal circuit to be newly switched. Each of these positions or each of the service lives corresponds to a possible switching signal program. Thus, there is a defined, finite number of switching signal programs per node. If the analysis is extended to several nodes whose signal programs form, for example, a green wave, the statement remains valid. Therefore, it is possible to examine a finite number of switching variants and to select the most favorable ones.

In an advantageous embodiment of the switching method according to the invention is for traffic engineering evaluation of a Switchover variant determines the extent to which the evaluation criteria or the criteria for the switchover signal programs of the switchover variant deviate or deviate from the signal program to be newly switched. Advantageously, it is possible to dispense with traffic-related evaluation of the switching variants to traffic data determined online by detectors. The entire relevant information about the traffic situation is extracted here from the new signal program to be switched. This is assumed to be optimal for the current and expected traffic situation. It is assumed that the creation of the new signal program to be switched according to certain traffic engineering or planning criteria or rules has been done. These then allow conclusions to be drawn regarding the expected traffic volume and the expected traffic flows at the individual junctions. Therefore, the traffic-technical evaluation of the switching variants on the determination of the respective deviations of the Umschaltsignalprogramme of the newly switched signal programs.

In a preferred embodiment of the switching method according to the invention, the evaluation criteria taken into account are the power capabilities of the relevant signal groups resulting from the transition signal programs of a switching variant. With green waves, the signal groups in the main load and their opposite directions are relevant. From the considered signal program the green time distribution, i. the ratio of green time to round trip time, determines and the saturation traffic intensity, i. the maximum possible traffic volume in a cross-section of a lane during one hour release time, estimated. The latter is a constant value, which depends only on road construction conditions. The efficiency results from this as a product of the share of the release time in the total circulation and the saturation traffic strength.

In an advantageous embodiment of the method according to the invention, the left-turning phases of the switching signal programs of a switching variant are taken into account as the evaluation criterion. For this purpose, the offset of the release times of the mentioned signal groups is considered, which has been used for the consideration of left turn in the creation of the signal to be switched signal program.

In a further advantageous embodiment of the method according to the invention, the effective capacities of pulse bands generated by a switching signal program of a switching variant at a node are taken into account as an evaluation criterion. The release time at a considered node creates a burst of force that hits the next node in the direction of progression. It depends on its signaling state, how many vehicles from the incoming pulse band may possibly pass through the node in the next signal circulation. In order to be able to quantify this effect, the temporal overlap of the pulse band with the release time of this node is measured.

In a preferred embodiment of the method according to the invention, lead and lag times for the pulse band in the switching signal programs of a switching variant are taken into account as the evaluation criterion. A waiting at a node vehicle queue should be reduced if possible before the arrival of the vehicle body in the direction of progression. This is achieved by possible lead times and lead times to the pulk band and thus included in the evaluation.

In a preferred embodiment of the switching method according to the invention, conditional waiting times and / or numbers of holding are taken into account as the evaluation criterion by the switching signal programs of a switching variant. As a result, the traffic assessment is supplemented by a qualitative measure, namely how often a vehicle is forced to stop and / or wait in a vehicle cluster moving toward the signalized node.

Advantageously, the evaluation criteria in the switching method according to the invention for the implementation of a predetermined traffic strategic goal, in particular a green wave, a user optimum or a system optimum, weighted. The different weighting of the evaluation criteria makes it possible to set different traffic engineering priorities when choosing the cheapest switchover variant. With green waves, the focus would be on the signal groups in the direction of progression. But also the lowest possible blocking time for vehicles in the transverse direction can be taken into account. In addition, system or user optima can be determined. At a user optimum, the maximum deviation of a signal program per round is considered, for example the maximum waiting times per vehicle at all considered nodes. This roughly corresponds to the experience of a driver in the generated pulse band. We are looking for the switchover variant in which these deviations all have the lowest possible upper limit. In contrast to this, the switchover variant with the lowest possible average deviation is searched for at a system optimum. This can be much lower than the user optimum. Depending on the distribution of the deviations, however, this can be done at the expense of a few pulse bands within the switching variant. In return, a collective property such as the vehicle throughput, optimized.

The object is further achieved by a device for switching signal programs for light-signal-controlled nodes of a road network, wherein at each node a traffic signal with a signal generator and a control device for executing a signal program is arranged, and wherein the control devices and a traffic computer are operatively connected to each other and configured in such a way in that a switching method according to one of Claims 1 to 9 can be carried out.

• FIG 1 die sich periodisch wiederholende Lage der Umschaltzeitpunkte aus einem aktuell geschalteten und einem neu zu schaltenden Signalprogramm und
• FIG 2 ein Zeit-Weg-Diagramm einer Umschaltvariante für die Hauptlastrichtung

schematisch veranschaulicht sind.Further details and advantages of the switching method according to the invention will become apparent from an embodiment shown in the drawings, in which

• FIG. 1 the periodically repeating position of the switching times from a currently switched and a new switching signal program and
• FIG. 2 a time-distance diagram of a switching variant for the main load direction

are illustrated schematically.

In the diagrams according to FIG. 1 are to the right a reference or reference time t in seconds and upwards each of the time steps Tx in the signal circulation of a currently switched signal program SP1 (upper diagram) and a newly switched signal program SP2 (in the lower diagram) plotted. The sawtooth patterns illustrate the periodically repeating signal circulations after a circulation time TU1 of 140 seconds in the currently switched signal program SP1 and with a circulation time TU2 of 120 seconds for the signal program SP2 to be newly switched. When Tx = 40 is the cheapest switching time TS1 for the currently switched signal program SP1, while the most favorable switching time TS2 for the new-switching signal program SP2 is Tx = 80. According to the switchover method with service life, switchover signal programs with different lengths of service life t1,..., T6, which are repeated periodically, result for the various signal circulations. The period length Tp of this switching operation corresponds to the smallest common multiple of the two round trip times TU1 and TU2 and in the illustrated embodiment is 840 seconds. The number N of switching signal programs at a node is therefore determined from the quotient of the period length Tp of the switching process and the circulation time TU1 of the currently switched signal program SP1. In the case of L nodes in the line segment with common cycle times, there are a total of N ^L switching variants, each of which is determined by a time-distance diagram according to FIG FIG. 2 can be represented. The finiteness of the number of switching variants makes it possible in accordance with the invention to evaluate the individual switching variants in terms of traffic and to select the most favorable one for execution.

In the traffic assessment, the switching signal program of each round is evaluated at each node. Each switchover variant L x N contains switchover signal programs to be evaluated. Since the switching can take place at different nodes in different rounds, the entire round trip period Tp of a switching variant is always evaluated. The performance of the relevant signal groups, the consideration of left-turn phases, the effective capacities of the pulse band and the lead times and lead times to the pulse band can be evaluated. These four quantitative evaluation criteria can still be determined by a qualitative evaluation criterion, such as Waiting times or stops, to be supplemented. Generally, an attempt is made to record the characteristics of the signal programs relevant to vehicles in the direction of progression. The consideration can be extended analogously to the secondary directions depending on the objective.

As an essential characteristic, the switching method according to the invention does not use currently determined traffic detector data. Instead, the following approach is used: The rewritten signal program is known for all nodes and is assumed to be optimal for the current and expected traffic demand. Therefore, it serves as a reference for the evaluation of the most technically favorable switching variant. The characteristic properties of all switching variants are compared with those of the respective new signal plan to be switched. This happens for each switching signal program at each node in each revolution of the switching period Tp. That switching variant with the smallest deviation is considered to be the one best adapted to the traffic situation and selected according to the invention as the most favorable switching variant. To track different traffic strategic goals, such as a green wave on a line, or a user or system optimum, the individual rating criteria can be weighted accordingly.

As a result, the switching method according to the invention provides the traffic-technically most favorable switching variant, which according to FIG FIG. 2 is shown as a time-distance diagram. The switching process was optimized here for a road that connects six nodes K1 to K6. In this case, the distance s in meters from a reference point 0 to the nodes K1 to K6 is plotted to the right. At the nodes Kn (n = 1,..., 6), t pairs of signal time diagrams are plotted upwards in the direction of increasing time, of which the left one is assigned to the signal groups in the main load direction of the line train and the respectively right signal plan is assigned to the signal groups of the opposite direction. In a signal plan, release times are symbolized by an empty bar and blocking times by a solid line. In the illustrated embodiment takes place at t0 = 273 seconds, the switchover, characterized by the horizontally extending dashed line. Until t = 360 seconds, the currently switched signal programs SP1 are processed at the nodes Kn, which in the exemplary embodiment FIG. 2 have a circulation time TU1 of 120 seconds; the associated circulation limits are indicated by horizontal, dotted lines. Then, the switching operation starts with a period length Tp of 840 seconds. The beginning and end of this switching period are indicated by solid lines. From t = 1200 seconds, the switching process is completed at each node Kn and the new signal programs SP2 to be switched are processed. The new signal programs SP2 to be switched have a circulation time TU2 of 120 seconds; their circulation limits are in FIG. 2 drawn by dotted lines. The local switching process can be recognized by the changed width of the pulse band P. In the illustrated embodiment, the first three nodes K1 to K3 are switched in the fourth round and the last three nodes K4 to K6 in the fifth round of the switching period Tp. The switching variant therefore receives the signature (4, 4, 4, 5, 5, 5). In the illustrated embodiment, only the pulse bands P in the main load direction, in FIG. 2 from left to right, shown. The inverse slope of the The center line of a pulselet P corresponds to the rate of progression Vp with which a vehicle must actually travel in order to be able to pass the coordinated nodes K1 to K6 without stopping.

Overall, all switching signal programs SPU and thus also the service lives ti (i = 1,..., 6) for the individual signal groups are calculated explicitly. The finiteness of the number of available switching signal programs and thus also of the switching variants is worked out according to the invention. This creates the basis of the traffic-technical evaluation of the individual switching variants. Furthermore, by exploiting the qualitative characteristics of the variant space, it is suitably reduced. This part of the procedure requires a switchover with service life. However, similar approaches for determining the possible switching variants are also conceivable for other methods.

According to the structure of the described switching method, the selection of the switching variant is an essential part. In particular, the evaluation part is independent of the provision of the possible switching signal programs. With a finite number of possible switching variants, the described evaluation method would still apply; also with an infinite number of switching signal programs generated for example by genetic algorithms.

Without detector data for determining the current traffic situation, the evaluation of the switchover variants by a traffic model appears to have little prospect. Without suitable detector data, only a stochastic parameterization of the model would be possible. In the case of pulp models, this results in a smooth intensity profile and thus produces a continuous vehicle flow in the ensemble means. This is also implicitly assumed in the switching method according to the invention. However, in the case of pulp models, absolute values are explicitly derived from this. But these are then based on stochastically averaged Assumptions and do not take into account the current traffic situation. Therefore, an absolute model-based evaluation only makes sense on the basis of real detector values and therefore represents a useful extension of the invention.

In the method considered here, therefore, the entire relevant information about the traffic situation is extracted from the signal program to be newly switched. This approach is the innovative core of the evaluation process. The signal program to be switched is assumed to be optimal for the current and expected traffic situation. The selection of the signal program to be newly switched is not part of the described method. It is assumed that the creation of the new signal program to be switched has been carried out according to certain traffic engineering or planning criteria or rules. These then allow conclusions to be drawn regarding the expected traffic volume and the expected traffic flows at the individual junctions. The evaluation of the switching variants takes place according to the invention via the determination of the respective deviations from the signal program to be newly switched.

This quantifies in particular two basic processes of the switching process: the delayed switching and the effects of the switching itself. The signal program change does not take place locally at the earliest possible time. As a result, a signal program not adapted to the traffic situation runs until the changeover cycle. The resulting effects on the traffic flow are taken into account in particular by the comparison with the new signal program to be switched. The switching signal program generally has a different signal from the currently switched and re-switching signal program signal image. The pulse band P thus generated hits at a node on a variant-dependent signaling state, e.g. on a blocking time over the entire width of the Pulkbandes. Thus, the direct effects of the switching itself are taken into account at a node. The desired optimum results from a compromise between the two evaluations.

Claims (10)

1. Method for changing over signal programmes for light-signal-controlled nodes (Kn) in a road network,

   - wherein for each node (Kn) a currently connected signal programme (SP1), a new signal programme to be connected (SP2) and at least one changeover signal programme (SPU) are prescribed,
   characterized

   - in that a plurality of changeover variants, each determined by a local-time order of changeover signal programmes (SPU) for the nodes (Kn), are prescribed,

   - in that the changeover variants are rated using one or more traffic-related rating criteria, wherein the respective discrepancies in the changeover variants from the new signal programme to be connected are determined,

   - in that the changeover variant having the smallest discrepancy is selected as the most favourable changeover variant,

   - and in that the changeover variant which is most favourable for traffic in terms of the rating is executed.
2. Method according to Claim 1,

   - wherein, for each node (Kn), changeover signal programmes (SPUi) with idle times (ti) are ascertained which correspond to the respective possible reciprocal positions of the changeover times (TS1, TS2) for the currently connected and the new signal programme to be connected (SP1, SP2),

   - and wherein the changeover variants formed are changeover signal programmes (SPUi) - combined into local-time orders

   - for the nodes (Kn).
3. Changeover method according to Claim 1 or 2,

   - wherein the traffic-related rating of a changeover variant involves ascertaining the extent to which the rating criterion or rating criteria for the changeover signal programmes (SPUi) of the changeover variant differs or differ from the respective new signal programme to be connected (SP2).
4. Changeover method according to one of Claims 1 to 3,

   - wherein the rating criterion taken into account is the performance capabilities of the relevant signal groups, as arise from the changeover signal programmes (SPUi) of a changeover variant.
5. Changeover method according to one of Claims 1 to 4,

   - wherein the rating criterion taken into account is the left-turn phases of the changeover signal programmes (SPUi) of a changeover variant.
6. Changeover method according to one of Claims 1 to 5,

   - wherein the rating criterion taken into account is the effective capacities of bunching bands (P) produced by a changeover signal programme (SPUi) of a changeover variant at a node (Kn).
7. Changeover method according to one of Claims 1 to 6,

   - wherein the rating criterion taken into account is lead and lag times for the bunching band (P) in the changeover signal programmes (SPUi) of a changeover variant.
8. Changeover method according to one of Claims 1 to 7,

   - wherein the rating criterion taken into account is waiting times and/or numbers of stops which are caused by the changeover signal programmes (SPUi) of a changeover variant.
9. Changeover method according to one of Claims 1 to 8,

   - wherein the rating criteria are weighted for the implementation of a prescribed traffic-strategy aim, particularly a green wave, a user optimum or a system optimum.
10. Apparatus for changing over signal programmes for light-signal-controlled nodes (Kn) in a road network,

    - wherein at each node (Kn) there is arranged a set of traffic lights having signal generators and a controller for executing a signal programme (SP),

    - wherein the controllers and a traffic computer are operatively connected to one another and are designed such that it is possible to execute a changeover method according to one of Claims 1 to 9.

Images