DE102019215930A1 - Method of operating a traffic control system and traffic control system - Google Patents

Abstract

The present invention relates to a computer-implemented method for operating a traffic control system (1), comprising the following steps: Providing (S1) time series data from several operating variables (OV), one or more traffic variables (TV), one or more environmental variables (EV ) and one or more emission values; - generating (S4) a diagram of a causal structure based on the time series data; - determining (S5) the one or more operating variables (OV) among the plurality of operating variables (OV) as a function of their causal effects on one Emission level (EL), the emission level (EL) being determined by the one or more emission values; - Optimizing the determined one or more operating variables (OV) by minimizing the emission level (EL); - Operating (S6) the traffic control system (1 ) as a function of the one or more optimized operating variables (OV).

Description

Technical area

The present invention relates to a traffic control system which enables multiple interventions to influence the traffic so that the air pollution caused by the traffic is reduced. The present invention also relates to methods of optimizing interventions in traffic control systems to reduce air pollution.

Technical background

Traffic control systems are used to control the flow of traffic, the amount of traffic and the pollution caused by the traffic. In urban areas in particular, air pollution from traffic can be high and can cause serious damage to the health of those nearby. It is previously known to provide interventions in the flow of traffic through traffic control systems in order to reduce or distribute air pollution. In order to reduce air pollution emanating from traffic, a traffic reduction based on speed limits up to a complete traffic ban by external factors such as e.g. B. Weather conditions are initiated.

However, the effects on traffic flow and air pollution can depend on many external factors that are not well known. In addition, previous behavior of the system is not taken into account, so that the dynamics of the system are not taken into account for the selection of the type of intervention carried out to reduce air pollution.

Summary of the invention

According to the invention there is provided a method of operating a traffic control system for reducing air pollution according to claim 1 and the traffic control system according to the further independent claim.

Further embodiments are specified in the dependent subclaims.

• - Bereitstellen von Zeitreihendaten mehrerer Betriebsvariablen, einer oder mehrerer Verkehrsvariablen, einer oder mehrerer Umweltvariablen und eines Emissionspegels;
• - Erzeugen eines Diagramms der Kausalstruktur basierend auf den Zeitreihendaten;
• - Bestimmen der einen oder der mehreren Betriebsvariablen unter den mehreren Betriebsvariablen in Abhängigkeit von ihren kausalen Auswirkungen auf den Emissionspegel,
• - Optimieren der bestimmten einen oder mehreren Betriebsvariablen durch das Minimieren des Emissionspegels;
• - Betreiben des Verkehrssteuersystems in Abhängigkeit von der einen oder den mehreren optimierten Betriebsvariablen.

According to a first aspect there is provided a method for operating a traffic control system comprising the following steps:

• - providing time series data of several operational variables, one or more traffic variables, one or more environmental variables and an emission level;
• Generating a diagram of the causal structure based on the time series data;
• - determining the one or more operating variables among the several operating variables depending on their causal effects on the emission level,
• - optimizing the determined one or more operating variables by minimizing the emission level;
• - Operating the traffic control system as a function of the one or more optimized operating variables.

The transport system and its impact on the environment is a complex system that is partly influenced by traffic control interventions, such as e.g. B. setting variable speed limits, controlling the status of the light signals, controlling the directional signs and the like can be controlled. For a specific area, such as B. one or more road junctions, the facts about the air quality depend on various factors, such as e.g. The number of vehicles passing through the specific area, the time for the vehicles to stop at traffic lights, slow traffic and the like. Therefore, deciding which intervention will produce the greatest air pollution reduction is not a straightforward approach. The above method intends to analyze the time series data concerning the traffic, the control operation and the environmental conditions in the specific area and the resulting pollution conditions, aiming to find the strongest causal relationships for an assessment of which of the interventions should be carried out to the highest Effect reduction in pollution.

According to the above method, the time series data including the emission amounts, the operational variables, the traffic variables, and optionally the environmental variables, are collected. While the amount of emissions is related to an amount of the type of emissions caused by traffic, such as B. the emissions of CO ₂ , NO _x , particulate matter, such as. PM10, PM2.5 and the like, and others, the traffic variables are used to assess the traffic situation, e.g. B. the traffic density (the number of cars per route), the traffic flow (the average speed of the cars in a part of the specific area or the full specific area), the uniformity of the traffic (deviations from the traffic density and / or the traffic flow) and the like , to describe. The environmental variables contain the information about the weather conditions, such as B. the temperature, the wind speed, the wind direction, the rain condition, the air pressure and the like, and can also take external measures, such as E.g. a discount on public transport, and other external parameters that may affect the amount of traffic and / or the emissions caused by vehicle traffic. In addition, the operating variables can change the control settings for the light signal control, e.g. B. a time of the red light phases, a time of the green light phases, the time between the red / green phases, the control settings for an electronic speed limit sign, such. B. the switching on / off and the exact speed limit to be displayed contain.

The resulting time-series data of each of the parameters and variables mentioned above can be analyzed to extract the time-series data of the relevant features. The time series data and the optionally extracted feature time series data are input into a determination of a diagram of a causal structure which shows the dependencies between the variables and the parameters between the various time steps. The causal structure diagram can be used to select the relevant operating variables to be controlled for the highest effect on the emission level. This makes it possible to select the one or more operating variables to be controlled in order to obtain the highest effect on air pollution.

Applying an optimization problem to the particular one or more operating variables. The operating variables in the result of the Granger analysis are then optimized for reducing emissions, and the optimized operating variables are used to control the traffic control system to reduce emissions from traffic.

According to a further embodiment, the time-series data further contain feature time-series data, the at least one of a mean, a variance, an autocorrelation and a cross-correlation of the plurality of operational variables, the one or more traffic variables, the one or more environmental variables and the one or more Emission values included.

Moreover, the diagram of the causal structure can be generated by a Granger analysis.

It can be provided that the environmental variables contain at least one of the metrological data, in particular cloud cover, wind speed, wind direction, temperature, relative humidity, precipitation, and an external parameter, in particular a price reduction for public transport, and / or that the traffic variables include at least one of a traffic density, a traffic flow and a traffic uniformity, and / or that the operating variables include at least one of a traffic signal control parameter and a speed limit sign control parameter.

It can be provided that the emission level is determined by one or more emission values, wherein in particular the emission values _{contain at least one of a concentration of NO x} , a concentration of CO ₂ and a concentration of fine dust.

Furthermore, the optimization can be based on an objective function which is based on the diagram of the causal structure between the variables.

Moreover, the steps of optimizing the determined one or more operating variables and operating the traffic control system can be carried out cyclically.

It can be provided that the step of generating a diagram of the causal structure based on the time series data and the step of determining the one or more operating variables among the plurality of operating variables are only performed if a change point of a sudden change in a property of the traffic control system has been detected .

As a result, change point detection analysis (fluctuation analysis) can be carried out to keep track of the changes in the traffic control system. If dramatic changes occur, the procedure of generating a diagram of the causal structure and performing optimization on the operating variables must be followed.

• - Zeitreihendaten mehrerer Betriebsvariablen, einer oder mehrerer Verkehrsvariablen, einer oder mehrerer Umweltvariablen und eines Emissionspegels bereitzustellen;
• - ein Diagramm der Kausalstruktur basierend auf den Zeitreihendaten zu erzeugen;
• - die eine oder die mehreren Betriebsvariablen unter den mehreren Betriebsvariablen in Abhängigkeit von ihren kausalen Auswirkungen auf den Emissionspegel zu bestimmen;
• - die bestimmte eine oder die bestimmten mehreren Betriebsvariablen durch das Minimieren des Emissionspegels zu optimieren;
• - das Verkehrssteuersystem in Abhängigkeit von der einen oder den mehreren optimierten Betriebsvariablen zu betreiben.

According to a further aspect, a traffic control unit for operating a traffic control system, wherein the traffic control unit is configured:

• - provide time series data of a plurality of operational variables, one or more traffic variables, one or more environmental variables and an emission level;
• generate a diagram of the causal structure based on the time series data;
• - determine the one or more operating variables among the plurality of operating variables depending on their causal effects on the emission level;
• - optimize the determined one or more determined operational variables by minimizing the emission level;
• - operate the traffic control system in dependence on the one or more optimized operating variables.

Figure list

• 1 eine beispielhafte Verkehrssituation zeigt, in der ein Verkehrssteuersystem verwendet werden kann;
• 2 einen Ablaufplan zum Veranschaulichen des Verfahrens zum Ausführen eines Verkehrssteuersystems zeigt;
• 3 ein Diagramm der Kausalstruktur für das Emissionsverringerungsproblem beispielhaft zeigt; und
• 4 die kausalen Abhängigkeiten zwischen den Zeitreihen der Betriebsvariablen, der Umweltvariablen, der Verkehrsvariablen und der Emissionswerte beispielhaft zeigt.

The embodiments are described in more detail in connection with the accompanying drawings, in which:

• 1 shows an exemplary traffic situation in which a traffic control system may be used;
• 2 Figure 12 shows a flow chart illustrating the method of executing a traffic control system;
• 3 shows a diagram of the causal structure for the emission reduction problem by way of example; and
• 4th shows the causal dependencies between the time series of the operational variables, the environmental variables, the traffic variables and the emission values.

Description of the embodiments

1 shows schematically a traffic control system 1 for controlling traffic in a traffic controlled area, e.g. B. at an intersection, by means of a traffic control unit 2 . The traffic control unit 2 is configured to control traffic intervention units. The light signals 3 be provided to by means of the operating variables, the variables, such as. B. the red times, the green times, the switching modes or the like specify to be controlled. Furthermore, the operating variables can be used to control the electronically controlled speed limit signs 4th (Traffic Intervention Unit) can be used to control the speed limits of vehicles passing the intersection 5 to control.

Any of the light signals 3 and each of the speed limit signs 4th can be controlled separately by one or more operating variables. In order to avoid impermissible conditions, a number of light signals 3 such as B. two opposite light signals 3 , be controlled synchronously so that only one of the pair of the two opposite light signals shows a green light. The operating variables for the light signals 3 can also be set as a mode of operation by which the green and red times for each of the light signals 3 are specified in the traffic-controlled area. The operating variables for setting the speed limit signs 4th can be a power off condition or any speed limit such as 50, 60, 70, 80 or the like.

Furthermore is the traffic control unit 2 with one or more environmental sensors 6th connected to the environmental variables, such as B. the temperature, the rain condition, the ambient light situation or the like to detect. One or more of the environmental variables can also be received by an external source, in particular via a communication system, wherein the environmental variables also include factors such as e.g. B. a discount on public transport, or other factors that may have an impact on the traffic situation in the traffic-controlled area.

Furthermore, by means of a traffic sensor system 7th the traffic variables that indicate the traffic situation are detected. The traffic situation can be determined by the traffic variables, such as B. a traffic flow (the number of cars per route), a traffic flow (the average speed of the cars in a part of the specific area or in the entire specific area), the traffic evenness (an indicator of the temporal deviations of the traffic density and / or the traffic flow) or the like, can be specified that the traffic situation for the entire traffic-controlled area or for parts of the traffic-controlled area, such. For each of the streets that intersect, separately or in general by the number of positions in the traffic-controlled area.

There are also one or more emission sensors 8th provided to measure the emission values, such as B. the concentration of CO ₂ , the concentration of NO _x and the concentration of particulate matter, such as PM10, PM2.5, PM1, and the like to detect. While the operating variables through the control unit 2 can be controlled directly, the environmental variables and the emission values cannot be controlled directly, but are a result of the development of the traffic situation in an earlier period.

The traffic control unit 2 carries out a procedure that is related to a schedule according to 2 is described in more detail. The method is available as software and / or hardware, e.g. B. by means of a microprocessor of the traffic control unit 2 , implemented.

In step S1 the above-mentioned operating variables, traffic variables, environmental variables and emission values are recorded continuously. Therefore, a time series of the operating variables, the Traffic variables, the environmental variables and the emission values, which are sampled at predetermined times, preferably equidistant times, provided. As a result, historical time series data up to the current point in time can be obtained. There are z. B. the time series data of the meteorological data (weather condition) WC (such as the cloud cover, the wind speed, the wind direction, the temperature, the relative humidity, the precipitation and the like), the traffic density TD (such as the traffic density , the traffic flow), the operating variable TL (which e.g. indicate a red time of a light signal), an environmental variable, such as e.g. B. a discount PD on public transport, and the emission level E. Each of the time series can be multivariate.

An emission level E at a specific point in time can be calculated from the various emission values at this point in time by means of a given emission function which combines the emission values into an emission value which indicates the harmfulness of the emissions caused by traffic. The emission value can e.g. B. be a sum of the weighted emission values at the specific point in time.

In step S2 each of the time series obtained can be preprocessed, e.g. By determining a moving average, a variance or a correlation exponent which determines the autocorrelation or the cross-correlation, or other scaling parameters, such as e.g. B. a multifractal spectrum or the like. The characteristic time series can be obtained in this way. In principle, several characteristic time series of the operational variables, the traffic variables, the environmental variables and the emission values can be obtained.

In step S3 change point detection is carried out. Change point detection observes features such as A mean, a variance, a long term correlation of the emissivity to detect sudden property changes that are behind the time series data and to detect a point in time at which such a change in the operation of the traffic control system 1 can be detected. A change point can e.g. B. be detected if the deviation of the current emission value from an average emission value, such as. B. an average amount of emitted NO _x , is higher than a given threshold value. If a change point has been detected (alternative: yes), the method is further carried out by step S4 processed. Otherwise (alternative: no) the procedure for step S6 processed.

In step S4 the causal structure of the system is determined. The causal structure can be indicated by a causal structure diagram that can be generated in order to select the relevant operating variables that can be varied to intervene to achieve the defined goal, which is the reduction of air pollution.

The causal structure diagrams, also known as causal Bayesian networks, are graphical models that are used to encode the assumptions about a data generation process. The diagram of the causal structure enables the effect quantities to be estimated from the time series data. To build a causal structure diagram, each variable in the model has a corresponding vertex or node, with an arrow drawn from a first variable to a second variable whenever it is judged to be responsive to changes in the first variable when all other variables are kept constant. The causal models often contain “error terms” or “omitted factors” that represent all unmeasured factors that influence a second variable when a predecessor of the second variable is held constant.

This can be done using Granger analysis. The Granger analysis is based on the extracted features of the obtained time series data. The causal structure diagram is output and can be illustrated as it is in 3 for the example given above.

In 3 it is shown that for the example given above, the red time of the light signal TL causes the traffic density TD as shown by the arrow. Furthermore, the weather condition WC also causes the traffic density TD, but also the emission level E.

In 4th the dependencies between the various time steps are shown by way of example, the temporal relevance of the causal effects being indicated, the red time of the light signal TL influencing z. B. the traffic density TD with a delay of one time step. In the matrix shown, the variables N _{IT represent} an exogenous noise that influences the parameter I at time t. The causal structure diagram can be used to select the operating variables relevant to the traffic control system 1 to control. In the example given, it appears that the discount PD, which at first glance is correlated with the emission level, is a relevant variable that has to be controlled in order to reduce the emission. With regard to the diagram of the causal structure, however, no influence on the emission level can be detected on the basis of the diagram of the causal structure obtained.

One result of the Granger analysis is a linear model that describes the causality of the time series data. In the example given, the model shows coefficients, the weights a, b, c, d, e, of the diagram of the causal structure 3 on. The Granger analysis is based on an assumption that the causal structure of the system does not change during the analysis. Therefore must, as with the step S3 has been introduced, change point detection can be performed to identify and predict a sudden change in the properties of the system so that Granger analysis is repeated to identify the new causal structure of the system.

In step S5 is the graphical representation of the influence after 4th together with a correlation structure, which is the connection distribution of the time series data, is used in an optimization process, the objective function applied being the reduction in the emission level EL. The optimization process is carried out in order to obtain a decision as to which of the operating variables should be used to intervene in the traffic control system 1 to execute.

As a result, a decision is made as a function of the optimization of the objective function as to whether the light signal or the speed limit sign is varied.

The optimization process is therefore used to decide on the objective function for each of the operating variables.

For the present example, it has been decided that the red time of the light signal TL should be controlled. The output of the last time step t-1 is used here to obtain the solution, which is the time of the red phase of the light signal TL. In step S6 the selected operating variable OV, which is the red time of the light signal TL, is controlled.

The red time of the light signal TL is determined by solving an optimization problem. For the given example, the optimization problem is written in a form that conditions the observed weather condition WC _T-2 : $$T{\mathrm{L.}}_{t-2}^{*}=\underset{T{\mathrm{L.}}_{t-2}}{{\displaystyle arGmin}}\mathbb{E.}\left[\mathrm{E.}{\mathrm{L.}}_t\mid \mathrm{W.}{\mathrm{C.}}_{t-2}\right]$$

wobei $T{L}_{min}^{constr.}\text{ und }T{L}_{max}^{constr.}$

vorgegebene Werte sind. Dies sind Angaben, dass es eine maximale Zeit zwischen zwei Grünphasen des Lichtsignals geben sollte, andernfalls müssen die Autos auf dieser Straße übermäßig lang warten, und dass es eine minimale Zeit zwischen zwei Grünphasen geben sollte, so, dass die Autos auf der Kreuzungsstraße nicht zu lange warten müssen.Furthermore, additional restrictions can be applied, $$T{\mathrm{L.}}_{min}^{cOnstr.}\le T{\mathrm{L.}}_{t-2}\le T{\mathrm{L.}}_{max}^{cOnstr.},$$

in which $T{\mathrm{L.}}_{min}^{cOnstr.}\text{and}T{\mathrm{L.}}_{max}^{cOnstr.}$

are given values. These are indications that there should be a maximum time between two green phases of the light signal, otherwise the cars will have to wait excessively long on this road, and that there should be a minimum time between two green phases so that the cars do not close on the intersection road have to wait a long time.

The argmin equation along with these constraints gives the final general constraint optimization problem.

The optimization problem is solved based on the specific linear causal model as obtained by the causal process diagram. For the specific example, the linear optimization problem is: $$\begin{array}{c}\begin{array}{l}T{\mathrm{L.}}_{t-2}^{*}=\underset{T{\mathrm{L.}}_{t-2}}{{\displaystyle arGmin}}\mathbb{E.}\left[\mathrm{E.}{\mathrm{L.}}_t\mid \mathrm{W.}{\mathrm{C.}}_{t-2},\mathrm{E.}{\mathrm{L.}}_{t-1}\right]\\ \end{array}\\ =\underset{T{\mathrm{L.}}_{t-2}}{{\displaystyle arGmin}}\mathbb{E.}[c\mathrm{W.}{\mathrm{C.}}_{t-1}+dT{\mathrm{D.}}_{t-1}\\ \begin{array}{l}+x\mathrm{E.}{\mathrm{L.}}_{t-1}\mid \mathrm{W.}{\mathrm{C.}}_{t-2},\mathrm{E.}{\mathrm{L.}}_{t-1}]\\ \end{array}\\ =\underset{T{\mathrm{L.}}_{t-2}}{{\displaystyle arGmin}}\mathbb{E.}[cu\mathrm{W.}{\mathrm{C.}}_{t-2}+daT{\mathrm{D.}}_{t-2}\\ \begin{array}{l}+x\mathrm{E.}{\mathrm{L.}}_{t-1}\mid \mathrm{W.}{\mathrm{C.}}_{t-2},\mathrm{E.}{\mathrm{L.}}_{t-1}]\\ \end{array}\\ \begin{array}{l}=\underset{T{\mathrm{L.}}_{t-2}}{{\displaystyle arGmin}}cu\mathrm{W.}{\mathrm{C.}}_{t-2}+daT{\mathrm{D.}}_{t-2}+x\mathrm{E.}{\mathrm{L.}}_{t-1},\\ \end{array}\\ \text{so that}T{\mathrm{L.}}_{min}^{cOnstr.}\le T{\mathrm{L.}}_{t-2}\le T{\mathrm{L.}}_{max}^{cOnstr.}.\end{array}$$

This applies to the example of the time of the red phase of the light signal as a solution depending on the weather condition in time step t - 2. Obviously, the solution to this specific problem is $$T{\mathrm{L.}}_{t-2}^{*}=T{\mathrm{L.}}_{min}^{cOnstr.}.$$

wobei $\mathrm{\Phi }\left({\overrightarrow{OV}}_{t-2}\right)$

einen Satz von Einschränkungen der Betriebsvariablen definiert. Das nächste Beispiel zeigt einen Aufbau, in dem die ausgewählten Betriebsvariablen die Dimension zwei aufweisen $\overrightarrow{OV}=\left(TL,SL\right)$

(TL: Lichtsignal-Steuerparameter, SL: Geschwindigkeitsbegrenzungszeichen-Steuerparameter). Das kausale Diagramm und seine entsprechende Matrixgleichung sind in den 3 und 4 gezeigt.In the general case where the selected operating variables OV, environmental variables EV, the traffic variables and the emission level EV can be written as follows, so that the optimized setting can be obtained for the set of the operating variables OV. $$\begin{array}{l}{\overrightarrow{OV}}_{t-2}^{*}=\underset{{\overrightarrow{OV}}_{t-2}}{{\displaystyle arGmin}}\mathbb{E.}\left[\mathrm{E.}{\mathrm{L.}}_t\mid \mathrm{E.}{V}_{t-2},\mathrm{E.}{\mathrm{L.}}_{t-1}\right],\\ \text{so that}\mathrm{\Phi }\left({\overrightarrow{OV}}_{t-2}\right)\ge \mathrm{0,}\end{array}$$

in which $\mathrm{\Phi }\left({\overrightarrow{OV}}_{t-2}\right)$

defines a set of constraints on the operational variables. The next example shows a setup in which the selected operating variables have dimension two $\overrightarrow{OV}=\left(T\mathrm{L.},\mathrm{S.}\mathrm{L.}\right)$

(TL: light signal control parameters, SL: speed limit sign control parameters). The causal diagram and its corresponding matrix equation are in the 3 and 4th shown.

The optimization for this example is $$\begin{array}{l}\left(T{\mathrm{L.}}_{t-2}^{*},\mathrm{S.}{\mathrm{L.}}_{t-2}^{*}\right)=\underset{T{\mathrm{L.}}_{t-2},\mathrm{S.}{\mathrm{L.}}_{t-2}}{{\displaystyle arGmin}}\mathbb{E.}\left[\mathrm{E.}{\mathrm{L.}}_t\mid \mathrm{W.}{\mathrm{C.}}_{t-2},\mathrm{E.}{\mathrm{L.}}_{t-1}\right]\\ \begin{array}{l}=\underset{T{\mathrm{L.}}_{t-2},\mathrm{S.}{\mathrm{L.}}_{t-2}}{{\displaystyle arGmin}}\mathbb{E.}\left[c\mathrm{W.}{\mathrm{C.}}_{t-1}+dT{\mathrm{D.}}_{t-1}+x\mathrm{E.}{\mathrm{L.}}_{t-1}\mid \mathrm{W.}{\mathrm{C.}}_{t-2},\mathrm{E.}{\mathrm{L.}}_{t-1}\right]\hfill \\ =\underset{T{\mathrm{L.}}_{t-2},\mathrm{S.}{\mathrm{L.}}_{t-2}}{{\displaystyle arGmin}}\mathbb{E.}[\left(cu+db\right)\mathrm{W.}{\mathrm{C.}}_{t-2}+daT{\mathrm{L.}}_{t-2}\hfill \\ +df\mathrm{S.}{\mathrm{L.}}_{t-2}\mid \mathrm{W.}{\mathrm{C.}}_{t-2},\mathrm{E.}{\mathrm{L.}}_{t-1}]=\underset{T{\mathrm{L.}}_{t-2},\mathrm{S.}{\mathrm{L.}}_{t-2}}{{\displaystyle arGmin}}daT{\mathrm{L.}}_{t-2}+df\mathrm{S.}{\mathrm{L.}}_{t-2},\hfill \\ \hfill \\ sOdassT{\mathrm{L.}}_{min}^{cOnstr.}\le T{\mathrm{L.}}_{t-2}\le T{\mathrm{L.}}_{max}^{cOnstr.}\hfill \\ \mathrm{S.}{\mathrm{L.}}_{min}^{cOnstr.}\le \mathrm{S.}{\mathrm{L.}}_{t-2}\le \mathrm{S.}{\mathrm{L.}}_{max}^{cOnstr.}.\hfill \end{array}\end{array}$$

It is stated that because the constant terms are not taken into account in the optimization, they can be removed from the objective function. xEL _t-1 is e.g. B. removed from the above set of equations.

In step S7 the values found are placed in the traffic control unit, the traffic control system 1 is operated accordingly. In particular, the optimized operating variables OV are used to control the corresponding traffic intervention unit, such as e.g. B. a light signal 3 or a speed limit sign 4th to control.

The process is cyclical and is stepped S1 continued further.

Claims (11)

Computer-implemented method for operating a traffic control system (1), comprising the following steps: - Provision (S1) of time series data of several operating variables (OV), one or more traffic variables (TV), one or more environmental variables (EV) and an emission level (EL); - generating (S4) a diagram of the causal structure based on the time series data; - Determination (S5) of the one or more operating variables (OV) among the several operating variables (OV) as a function of their causal effects on the emission level (EL), - Optimizing the determined one or more determined operational variables (OV) by minimizing the emission level (EL); - Operation (S6) of the traffic control system (1) as a function of the one or more optimized operating variables (OV). Procedure according to Claim 1 wherein the time series data further includes feature time series data representing at least one of a mean, a variance, an autocorrelation and a cross correlation of the plurality of operational variables (OV), the one or more traffic variables (TV), the one or more environmental variables (EV ) and one or more emission values. Procedure according to Claim 1 or 2 , wherein the diagram of the causal structure is generated by a Granger analysis. Method according to one of the Claims 1 to 3 , wherein the environmental variables (EV) contain at least one of meteorological data, in particular a cloud cover, a wind speed, a wind direction, a temperature, a relative humidity, a precipitation, and an external parameter, in particular a price reduction for public transport, and / or wherein the traffic variables (TV) contain at least one of a traffic density, a traffic flow and a traffic uniformity, and / or wherein the operating variables (OV) contain at least one of a light signal control parameter (TL) and a speed limit sign control parameter (SL). Method according to one of the Claims 1 to 4th where the optimization is based on an objective function based on the causal structure between the variables. Method according to one of the Claims 1 to 5 , wherein the steps of optimizing the determined one or more operating variables (OV) and of operating the traffic control system (1) are carried out cyclically, in particular the step of generating a diagram of a causal structure based on the time series data and the step of determining the one or the a plurality of operating variables (OV) among the plurality of operating variables (OV) are only executed if a change point of a sudden change in a property of the traffic control system (1) has been detected. Method according to one of the Claims 1 to 6th , the optimized operating variables (OV) being used to control a traffic intervention unit (3, 4). Method according to one of the Claims 1 to 7th , where the emission level (EL) is indicated by an or a plurality of emission values is determined, the emission values in particular containing at least one of a concentration of NO _x , a concentration of CO ₂ and a concentration of fine dust. Traffic control unit for operating a traffic control system (1), wherein the traffic control unit (2) is configured: - to provide time series data of several operational variables (OV), one or more traffic variables (TV), one or more environmental variables (EV) and an emission level (EL); generate a diagram of the causal structure based on the time series data; - to determine the one or more operating variables (OV) among the several operating variables (OV) depending on their causal effects on the emission level (EL); - to optimize the determined one or more determined operational variables (OV) by minimizing the emission level (EL); - to operate the traffic control system (1) as a function of the one or more optimized operating variables (OV). Computer program product comprising instructions which, when the program is executed by a data processing unit, cause the data processing unit to execute the method according to one of the Claims 1 to 8th to execute. A machine readable medium having stored therein a computer program comprising a routine of predetermined instructions for causing the machine to perform the method of any one of Claims 1 to 8th to execute.

Images