WO1997034274A1 - Fuzzy logic-assisted traffic-responsive control system for traffic light systems - Google Patents

Abstract

The disclosure relates to a traffic-responsive control system for traffic light systems using a multistage modular fuzzy control (FS) comprising an observation level (BE) and a control level (SE). In the observation level (BE) the processed detector values and data on signal group states (DSD) are processed, e.g. every second, in a first stage with the aid of a data-bank module (DBM) to form compressed signal group-related data. From the latter, in a second stage using a green time requirement module (GBM) and a signal groups weighting module (SGGM), the required green times of individual signal groups and the signal group weightings are calculated. In a third stage in the control level (SE), using a signal/frame plan adaptation module (SRAM) and based on the green time requirement for the subsequent signal cycle, the signal/frame plan is adapted for the subsequent signal cycle, and switching recommendations (SED) for signal groups are generated in a signal group updating module (SGAM) based on the signal/frame plan and the signal group weightings, or alternatively, in a phase-updating module (PAM), the signal group weightings are converted to phase weightings for switching recommendations for the current and/or subsequent phase.

Description

description

_V erkehrsabhängige control of traffic light systems using fuzzy logic

_T he invention relates to a traffic-dependent Steue¬ tion of traffic light systems with the aid of fuzzy logic in accordance with the preamble of claim 1.

The control of traffic lights for road traffic depending on traffic volume has long been known and common. In this case, the green time is often measured, that is to say a need-based adjustment of the release time is determined. In doing so, time gaps and occupancy levels in the access routes to the nodes were predominantly used for adapting the release time to traffic. For this purpose, fixed threshold values have been defined. Exceeding or falling below these threshold values, or combinations of these input variables, leads to the switching off or extension of the current traffic flow. When the release time is adjusted by means of time gap measurement, the time intervals between successive vehicles of a vehicle stream are measured as a time gap by means of a detector in the access point. The release time is adjusted after the selected minimum release time has elapsed or after an earliest point in time in circulation has been adapted to the current requirements of the inflowing vehicles. The release time can be extended until the measured time gap is at least as large as a predefined time gap threshold or until the longest defined release time or the latest extension time in the signal circulation is reached. When the release time is adjusted by measuring the degree of occupancy, threshold values for the termination of the release time are also defined. The measurement values are evaluated separately for each lane. The original occupancy rate is smoothed using a compensation procedure. If necessary, a compensation factor is used for increasing and decreasing tendencies of the original values.

In the green time measurement method SAEWA, as described in Grünlight, issue 25, pp. 4-9, the degree of utilization of a current signal group is compared with the degree of utilization of an enemy signal group standing on red.

The volume density method compares the number of vehicles in front of red with the time gaps in traffic. A limit value curve is defined, which assigns a fixed time gap value to a fixed number of waiting vehicles, up to which the current release time is extended.

Green time assessment according to urgency is an extension of the volume density method. The number of vehicles in front of red is compared with the time gaps and also with the occupancy in the current traffic. Threshold values are also used here.

It is the object of the invention to enable the release time, ie a green time assessment, to be adjusted as required and the next phase to be selected in the case of traffic-dependent light signal controls, with the aim of improving the quality of the _k e _h rsa _bl to increase. In the case of a signal group control of the _L ic _h tsignalanlage to switching recommendations determined who _d s, in order to influence the signal groups, ie those

_S ignalgruppen to determine which off or are on.

This object is achieved according to the invention with the features of claim 1.

In the traffic-dependent control system according to the invention, the current traffic condition is assessed using fuzzy logic. After weighing up the different interests of the competing traffic flows, a decision is made about the termination or extension of the current phase and a selection of the next phase, but depending on the possibilities given by the current signal / master plan. So far, hard, ie clearly defined threshold values have been used as a basis for all known methods. The use of fuzzy logic in the area of node control now makes it possible to replace these hard threshold values with smooth transitions. This enables a much more differentiated consideration of the input values. Another advantage is that it is now possible to link several input variables in a clear manner. For this purpose, the fuzzy control is constructed in several stages with a series of processing modules. In a first stage, modules for processing the input values (database module) and for classifying traffic conditions are used in an observation level. With the values prepared in this way, the required green times are in a green time requirement module for the individual Signal groups and a signal group weighting are calculated in a signal group weighting module. On the basis of the signal group weights, phase weights are calculated in a third stage for use in phase controls, and recommendations for extending, stopping or switching on phases are generated. For this purpose, the phase update module is provided in the control level. Recommendations for extending and canceling or switching on signal groups are generated for use in signal group control. A signal group update module is provided for this. This takes place within the framework of the signal plan, which is adapted accordingly by means of a signal / master plan adaptation module.

Further advantages and details of the invention emerge from the subclaims and from the following description with reference to the drawing. 1 schematically shows the principle according to the invention, FIG. 2 shows a structure of the green time requirement module, FIG. 3 shows a section of the rule base 1 and FIG. 4 fuzzy sets of the input variable assignment.

The traffic-dependent light signal control according to the invention with fuzzy logic has a modular structure and consists of two levels, as shown in FIG. 1. The data processing of the control unit DVSG with the operating system BS and with the control method SV, which can be a phase control or a signal group control, is indicated schematically on the left. The fuzzy control FS with the individual modules is shown schematically on the right. With the operating system BS of the control unit, the raw detector data are recorded and preprocessed in (1). Furthermore, in (2) information _t ion he _üb signal group states detected Furthermore belongs in _{(3) d} azu to make security checks _V om operating system will input data ED to the Steuerver¬ drive SV, the suggestively with (4) Monitoring of of- _f ent _l cozy transport network (OPNV) and the "green wave" _b he _ü taken into account in (5) is indicated the monitoring of the signal / outline plan and the call of the modules. with (6) is still expressed in that the control method also circuits of special phases, phase inversion or possibly Additional modules as they are used in a local control according to a certain procedure ("MOTION").

The fuzzy control FS receives both detector and signal status data DSD and module call data MAD from the control unit. In the observation level BE of the fuzzy control FS there are modules for data acquisition, data processing and information compression. Detector values prepared by the control unit and information about the system status (DSD) are used as input values. From this, compressed, signal group-related information VEI is generated, which is processed in the control level . With the help of this information, an adjustment of the signal / frame plan for the next cycle is carried out in the control level SE. In a further module, a recommendation for extending or switching off the current phase / signal groups is generated and the number of the next phase transition or the signal groups to be switched on is output to the higher-level control method SV (eg PDM or VS-PLUS). In addition to the actual switching of the signals (SB), these take over the monitoring of the OPNV and the "Green Waves" (MAD). It checks the conditions in the signal / master plan and calls if necessary, the modules m of the control level SΞ of the fuzzy control FS. These modules activate the required modules m of the observation level BE. The database module DBM is, for example, updated automatically every second. Depending on which control method is used, either Phase update module PAM or the signal group update module SGAM can be called. The switching recommendations SED of the fuzzy modules (FS) are converted into switching commands SB by the control method (SV). The operating system BS in the control unit continues to take over the safety - Relevant checks (3) It receives the raw detector data and processes them in time gaps, occupancy and numerical values, etc.

The modules are designed in such a way that they can be used independently of a defined intersection geometry and independently of the installation of the detectors. In the case of traffic-dependent control, of course a certain basic equipment of detectors is required. The modules are based on a manageable number of parameters set. Normally, all signal groups are treated equally. However, individual signal groups can also be prioritized by changing the corresponding parameters. The use of the output values of the detectors depends on the position of the detectors. This will be explained in more detail. When using phase controls, a switching recommendation is given in a subsequent phase if the weighting of the current phase is less than or equal to the weighting of a subsequent phase. When using signal group control, the recommendations for switching on or off are given Signal groups generated As can be seen from FIG. 1, the observation plane BE essentially has three modules. In the database module DBM, input values (DSD) such as degree of occupancy, time gap, count values during the blocking time and count values during the release time are processed and stored in relation to the signal group or traffic flow. With the target values ZW and the occupancy values BW from this database DBM and additional information about the last completed time gap LZ before the start of the blocking time and the time period AR until the arrival of the first vehicle after the beginning of the blocking time, the green time requirement module GBM provides the required Green times of the signal groups calculated for the next round. A more detailed explanation follows with reference to FIG. 2. This calculation is carried out once, at the end of the cycle. With the information about the green time requirement GBA of the signal groups for the current circulation, values ZW, BW from the database DBM, signal group states and waiting times, an evaluation of the urgency of the release request is made every second in the signal group weighting module SGGM. For this purpose, a weighting is calculated for each signal group. The higher the weight of a signal group, the greater their desire to get release times. When generating the weighting, a distinction is made between signal groups that are enabled and signal groups that are currently locked.

The green time requirement module GBM has four control bases RB1 to RB4, as shown in FIG. 2. The linking of the input values depends on the signal status and the position of the detectors. During the green time of the signal groups

Counting and occupancy values ZW, BW determined. These values are normalized, fuzzyfied and processed with the help of the rule base RBl to a green time factor GFl.

Figures 3 and 4 are considered below. The rules of the rule base RB1 shown in FIG. 3 can be read in the following form:

Rule 1: if the occupancy is small and the count is small, the green time factor is very small. Rule 2: if the occupancy is medium and the count is small, the green time factor is medium.

Rule 3. If the occupancy is small and the count is medium, the green time factor is small.

The terms "very small (VL)", "small (LOW)", "medium (MED)" and "large (HIGH)" are referred to as linguistic descriptions of the input / output variables and are presented by fuzzy sets (see Figure 4). In this form, the expert knowledge can be represented relatively easily and comprehensibly in a knowledge base (rule base).

With the aid of the fuzzy sets (LOW, ED, HIGH), see FIG. 4, the standardized, "sharp" occupancy value for further processing in the fuzzy control (FS) is converted into a fuzzy description form. It is converted this is described by its degree of belonging (belief) to the fuzzy sets. This process is referred to as fuzzyfication. The conversion of the "unsharp" value at the output of the fuzzy control to a sharp value is accordingly referred to as defuzzification.

The green time factor GF1 thus determined is determined in a subsequent step via the rule base RB3 with the green time _d ar _f GBA linked for the current circulation. The resultieren¬ _d e _A usgangsgröße is called green-time factor GF2. During the red time of the signal groups, the input variables used are determined as a function of the detector position. If there are only detectors on the stop line, the last time gap LZ before the green termination is linked via the rule base RB2 with the time until the first arrival of a vehicle in "red" AR. This red time factor RF1 determined in this way is also linked in the subsequent stage via the rule base RB3 with the green time requirement GBA for the current circulation. The resulting output variable is called the red time factor RF2. If detectors are available from about 35 m in front of the stop line for dimensioning, the vehicle arrival rate in "red" ARR is linked to the green time requirement GBA for the current circulation with the rule base RB4. The resulting factor is called the red time factor RF2 multi-lane carriageways, the maximum green time or red time factor is used.

The control level SE of the fuzzy control FS has three modules. In the signal / master plan adaptation module SRAM, an adaptation of the signal / master plan for the next round is carried out based on the green time requirement of the signal groups for the next round. This happens once at the end of a cycle and taking into account coordination points and restrictions by the planner, the higher-level control method and taking into account the structure of the given signal / framework plan. A list of the signal groups to be switched on and off is generated in the signal group update module SGAM on the basis of the signal / master plan and the signal group weights. in the Phase update module PAM are the weights of the

Signal groups processed into phase weights and the number of the next phase transition output. If there are several possible subsequent phases, the optimal subsequent phase for the traffic flow is determined.

Modules of the observation level BE will be discussed again. The following input values are prepared and processed in the database module DBM: the sum of the occupancy times, vehicle count values, time gaps, duration of the green times of the signal groups, duration of the red time of the signal groups and signal state. The following data are then available at the output of the database module: vehicle arrival rate in "red" (ARR), time until the first registration of a signal group in "red" (AR), last time gap LZ before the green termination, count values ZW in red , Occupancy rate (BW) in green and red and time gaps in green.

The green time requirement module GBM, which determines the green time requirement of the individual signal groups, is provided with the following input values: occupancy value BW (in green), count value ZW (in green), time gap ZL before the green termination, time until the first one arrives Vehicle in "Red" (AR), green time requirement old and arrival rate in "Red". In addition to the calculated green time requirement (GBA), the duration of the red time is calculated and made available at the module output.

With the invention described here, it is possible to achieve an improvement in conventional technology in traffic-oriented light signal systems in an application-oriented and useful manner. The modules of the _F uzzy- _ontrol in the common technique of the traffic signal control system ^¬ possible.

Claims

claims 1. Traffic-dependent control of traffic light signal systems with the aid of fuzzy logic, vehicle detector values such as count values, time gaps, occupancy values, etc., being recorded in the individual node access points, and processed in the control unit and made available for further processing, characterized in that the vehicle detector values of the intersection of approaching traffic flows are related to each other by means of fuzzy logic and an adaptation of a signal / master plan for a subsequent signal circulation takes place therefrom. 2. Traffic-dependent control according to claim 1, g e k e n n z e i c h n e t d u r c h the following features: - The control method has a multi-stage fuzzy control (FS), which is modular and consists of an observation level (BE) and a control level (SE), - in the observation level (BE) are regularly at short intervals (e.g. each Second) in a first stage, by means of a database module (DBM), the processed detector values and information about signal group states (DSD) are processed into compressed signal group-related or traffic flow-related data, in a second stage the required green times of the individual signal groups and the signal group weights are calculated from this data using a green time requirement module (GBM) and a signal group weighting module (SGGM), - in _t he control level (SE) are used for the following Si¬ gnalumlauf in a third stage by means of a signal / _{R h} a menplan adaptation module (SRAM) based on the _d Grünzeitbe¬ ARFs for the following signal circulating an adjustment of the Si gnal- / master plan and in the case of the signal group control of the traffic signal system in a signal group update module (SGAM) based on the signal / master plan and the signal group weighting switching recommendations (SED) for the signal groups to be switched or generated in the case of phase control of the traffic signal system in a phase update module (PAM) the signal group weights are processed into phase weights for a switching recommendation (SED) to extend the current phase or to switch to a subsequent phase. 3. Traffic-dependent control according to claim 2, characterized in that in the database module (DBM) the following input values: occupancy times, vehicle count values, time gaps, duration of the green times of the signal groups, duration of the red time of the signal groups and signal state are processed and processed in order to obtain the following initial values: vehicle arrival rate in red, time until a signal group is first registered in red, last time gap before the termination of green, count value in red, occupancy level in green and red and time gap in green. 4. Traffic-dependent control according to claim 2 or 3, characterized in that in the green time requirement module (GBM) from the following input values: occupancy value in green, count value in green, green time requirement old, arrival rate in "red", time gap before the green termination, the time until a vehicle arrives in red, the following values are calculated: green time requirement of the signal groups for the next round and duration of the red time. 5. Traffic-dependent control according to claim 4, characterized in that the green time requirement module (GBM) has four control bases (RBl to RB4), the input values being normalized and fuzzyfied and linked as a function of the signal state and the position of the detectors, that during the green time of the signal groups from the count values (ZW) and the occupancy values (BW), a first green time factor (GFl) is determined with a first rule base (RBl), that during the red time of the signal groups with an arrangement of the vehicle detectors before Stop line from the values of the last time gap (LZ) and the first arrival in "red" AR with a second rule base (RB2) a first red time factor (RF1) is determined that from the first green time factor (GFl), the first red time factor ( RFl) and the value of green time requirement old (GBA) in a third rule base (RB3) a second green time factor (GF2) or a second red time factor (RF2) The new green time requirement (GB1) is determined from the second green time factor (GF2), the second red time factor (RF2) and / or the third red time factor (RF3) in a manner known per se. 6. Traffic-dependent control according to claim 5, characterized in that for the calculation of the green time requirement of the individual signal groups by means of the green time requirement module (GBM) different input values are taken into account as a function of the arrangement of the vehicle detectors, with one detector before the stop line the last gap (LZ) and the first Arrival in "red" (AR) and, in the case of a detector at a greater distance (e.g. more than 35 m) from the stop line, the old green time requirement (GB1) and the arrival rate in "red" (ARR) are taken as a basis 7 Traffic-dependent control according to one of the preceding claims, characterized in that in the signal / master plan adaptation module (SRAM) at the end of a round using the The basic time requirement of the signal groups for the next round is adapted to the signal / master plan for the next round taking into account the higher-level control process and the structure of the predetermined signal / master plan. 8 Traffic-dependent control according to one of the preceding claims, d a d u r c h g e k e n n z e i c h n e t that between the first and the second stage in the observation vein (BE) of the fuzzy controller (FS), a further stage is provided, which has a forecast module for processing the data from the first stage