METHOD AND ARRANGEMENT FOR AUTOMATED LOCAL CONTROL OF MULTI-MODAL URBAN TRAFFIC FLOWS

Description

[0001] The present invention relates generally to the field of telecommunications and specifically to urban traffic flow control. The invention relates to a method and arrangement for the automated local control of multi-modal urban traffic flows, which allows for a response to locally detected traffic flow patterns by suitable computational methods, preferably carried out within a telecommunications network.

[0002] This allows local traffic control devices, e.g. traffic signal systems (LSA), to be adapted quickly to optimize local traffic flow, and also allows the overarching traffic flow control in a non-local reference area to be adapted to optimize non-local traffic flow.

[0003] The flow of traffic on a city's road network, including private motorized transport and public transport, as well as the movement of other road users such as pedestrians and cyclists, is controlled at critical intersections by traffic lights. The aim of this control is to optimize traffic flow in order to minimize travel times for all road users while considering road safety and environmental sustainability.

[0004] Currently, the dimensioning of traffic signal systems and transport capacities in urban transport systems is based on statistically validated traffic count data. This takes into account the various modes of transport. In the following description, different road users are distinguished as traffic modes, i.e., pedestrians, cyclists, private motorized transport, public transport, motorized fleet transport, etc.

[0005] Assuming typical daily patterns, temporal patterns for flow control of an LSA are determined based on this static count data and implemented in a control device depending on the time of day and / or seasonal times.

[0006] A further refinement of traffic flow control is adaptive traffic signal systems, which can make adaptive changes to the flow control based on sensor data, e.g. from counting loop data.

[0007] The parameters of a traffic signal program, which control traffic flow, are the cycle time, clearance time (phase duration), number of phases, phase sequence, and offset time. Traffic signal control methods differ in how traffic flow control is implemented and in the degree to which the signal program elements can be influenced or modified.

[0008] The choice of control method depends primarily on predefined objectives; for example, network optimization, line optimization, or local node optimization are possible.

[0009] According to [2], the control methods are generally classified as follows (taken from [7]): 1. Fixed-time control; All intersections have a uniform, fixed cycle time. All signal groups follow a fixed schedule. The selection of the respective schedule can be time-controlled or based on logic that relies on the acquisition of traffic data via detectors. 2. Logic-based or rule-based control; This type of control usually uses a framework plan as its basis, which corresponds to fixed-time control. However, this framework plan is overlaid at the individual intersections by modifications to the signal timings. These modifications result from information about the current traffic situation at the intersection. This also includes priority signaling for public transport vehicles. 3. Model-based control; This type of control generates the signal timing plan using a mathematical optimization model based on the acquisition of the entire traffic situation in the controlled road network.Components of the system include: a system for recording the traffic situation, a mathematical model for predicting the effects (e.g., time losses, number of stops, environmental impacts) of a signal timing plan, and an optimization procedure for minimizing the disadvantages of signal control.

[0010] In the past, traditional, fixed-time-controlled "green waves" and the coordinated, traffic-dependent control systems built upon them were predominantly used to achieve a smoother traffic flow. As a further development of these control methods, fully traffic-dependent, network-wide control systems have emerged, initially in other countries, in a form previously unknown in Germany, such as SCOOT in England and SCATS in Australia.

[0011] The optimization is performed during operation based on detector counts. In Germany, so-called adaptive systems such as MOTION and BALANCE have been developed. Both are currently available on the German market. MOTION is distributed by Siemens as part of the Sietraffic product, and BALANCE is distributed by GEVAS Software GmbH and TRANSVER GmbH.

[0012] These are all model-based control systems that calculate a control response adapted to the current traffic situation during operation.

[0013] The systems focus on optimizing traffic flow in locally limited subnetworks and therefore initially disregard the optimization of the overall network. Optimization of the overall network is possible via interfaces between the subnetworks. A detailed description of the four control methods mentioned can be found in [1], [3] and [8].

[0014] The methods described above are characterized by their operation on two levels: a centrally controlled tactical or macroscopic level for line or (partial) network optimization and a decentralized local or microscopic level, i.e., the control of individual nodes, e.g., for public transport priority switching. This represents a hybrid of centralized and decentralized control.

[0015] Fig. 1The diagram schematically illustrates the architecture currently used for the control methods described above. In the local nodes 200.n, in addition to a data acquisition unit 202 using conventional sensors, a control unit 201 is installed. The systems of the local nodes are connected to the central traffic computer 100. The traffic computer 100 can thus transmit the signal programs to the nodes and receives locally acquired traffic flow data. This data is processed in a module 101 to create or adapt signal programs based on traffic models via a node control module 102. The equipment in the local nodes can vary depending on the method used, as described above. Some nodes have no data acquisition system whatsoever, and in these, the control unit, equipped without any intelligence, simply executes switching commands according to a fixed schedule specified by the traffic computer.

[0016] With reference to Figures 2 to 4 The different types of control can be described as follows: In the case of central control, see... Fig. 2 Signal programs are processed by a traffic computer 100. After processing 101a and 101b of the available traffic data 202, this computer provides each connected control unit 201 of the individual traffic nodes 200 with the necessary release and blocking time information 300 for signaling. The basis for signal program creation, in addition to the currently recorded flow data, is the underlying traffic models of the respective procedures, which require complex modeling of the traffic network.

[0017] In detail, the recorded data 202 is transmitted to a data processing unit for junction 101b and a data processing unit for the line / network 101a. Subsequently, network / line control 102a is performed, followed by the determination of a signal program 300 and / or the direct determination of the signal program 300. The signal program 300 is transmitted by the traffic computer 100 to a corresponding signal group 201a in the control unit 201 at junction 200. Additionally, the control unit 201 has a signal backup 201b.

[0018] In the case of decentralized control, see... Fig. 3 The signal programs 300 are directly available in the control unit 201 of node 200 and are processed taking into account the current measurement data 202. The central unit 100 supplies the local control units 201 with the signal programs 300 and, if necessary, synchronizes different nodes 200.

[0019] In detail, the recorded data is transmitted to the data processing unit for line / network 101a of the traffic computer 100 and / or to the data processing unit for intersection 201c of the control unit 201. Network / line control 102a takes place in the traffic computer 100. The signal program 300 is determined in the control unit and transmitted to the signal groups 201a. Additionally, the control unit 201 has a signal backup 201b.

[0020] Such decentralized approaches, which involve almost completely shifting intelligence to the control units of the nodes, have the disadvantage that each individual control unit in all relevant nodes must be equipped accordingly, and in particular, the adaptation effort, i.e., the costs for updates, can be considerable in the event of a change. Past experience has shown that, due to constantly changing demand, i.e., traffic load, adjustments are to be expected at regular intervals.

[0021] Sometimes central control is combined with decentralized control to form semi-centralized control, cf. Fig.4, in which individual control functions are divided between central unit 100 and control unit 201. At the local control unit 201, node-specific actions are executed depending on the current traffic situation. Line- or network-specific, i.e., cross-node, actions continue to be controlled by central unit 100.

[0022] In detail, a semi-centralized control system has data processing for intersection 101b and 201c in both the traffic computer 100 and the control unit 201. The signal programs 300 can also be determined in both the traffic computer 100 and the control unit 201. The other elements correspond to the respective elements of the Figures 2 and 3 .

[0023] The local control units that switch the signals are used in different qualities. A distinction can be made as follows [7]: In principle, for traffic-dependent control, the devices must provide an interface for receiving the signal parameters and be able to control the signals accordingly.

[0024] Pure switching devices merely execute the switching cycles transmitted by the central unit; control units are also capable of independently executing signal switching programs. A group control unit can coordinate several account control units.

[0025] In the event of communication disruptions between the central unit and the control unit, the signal protection system intervenes with a corresponding emergency program.

[0026] It becomes clear that once a node is equipped, it can only be converted to a different control logic at considerable expense.

[0027] In the systems described so far, traffic flow data was collected using conventional sensors, primarily induction loops and camera data. In addition, GPS probes, i.e., movement data from GPS-equipped vehicles (Floating Car Data, FCD), have been tested as an additional data source in some research projects.

[0028] However, these methods have not been tested in significant quantities and are currently not being incorporated into control systems. An overview of possible data acquisition methods is available. Figure 2-5 See page 15 in [7].

[0029] Due to the different modeling methods, each control method places different demands on the location (installation site) of the detection points, which makes a simple switch from one system to another considerably more difficult.

[0030] The systems described so far, which represent the state of the art, no longer adequately address the current traffic situation. It is important to consider that the traditional classification of transport modes is changing significantly due to the electrification of drive technologies. For example, electric bicycles can reach considerably higher speeds but are not necessarily confined to lanes. Furthermore, pedelecs and cargo bikes represent entirely new modes of transport. Consequently, the classic models and control paradigms must be adapted to these structural changes.

[0031] Furthermore, traffic data collection methods need to be adapted. Traditional induction loop counting is becoming less relevant, as it only detects moving electrical conductors of sufficient size along a predetermined and narrowly defined path.

[0032] EP 2345020 A2 relates to a traffic control system for influencing traffic flow within a limited road or traffic network with intersections, e.g., in a large city, where a number of lights corresponding to the number N of intersections transmit the currently determining signal plans to a traffic computer. Detectors are provided in connection with the traffic signal systems (LSA) through which traffic data can be recorded and transmitted to the traffic computer.

[0033] US 2012065872 A1 concerns the calculation of real-time traffic conditions based on data transmitted over a network used by mobile phones within multiple automobiles, including, but not limited to, the time required to travel a certain distance on a given route.

[0034] US 2017270413 A1 relates to a system for filtering data for a traffic control center, comprising: a multitude of data sources; one or more network computing facilities configured to process data from the multitude of data sources to predict causes associated with predicted traffic events and to select data from the multitude of data sources to be output to the traffic control center based on the predicted causes.

[0035] US 2017 / 352264 A1 concerns a traffic control device for monitoring multiple right-of-way crossings at a traffic intersection. The traffic control device includes a mobile wireless interface for communicating with an in-vehicle mobile wireless device to receive a commuter profile containing a subscribed switching time interval.

[0036] WO 2018 / 141403 A1 relates to a method and a device for recording traffic flows in a road network with one or more nodes, wherein traffic data are recorded in a calculation device near the node and adapted traffic signal profiles are selected.

[0037] The invention presented here relates in particular to the data acquisition underlying the control methods, the data sources used and the system architecture of the control system.

[0038] The problems and other tasks described above are solved by the subject matter of the attached independent claims. The dependent claims relate to further aspects of the invention.

[0039] The present invention describes a method which will make it possible to react very quickly to changes in the traffic flow of all road users, including in particular groups of road users that have so far been given little consideration, such as pedestrians and cyclists, in order to adapt the traffic flow itself to the current conditions by adjusting the switching sequences of the traffic signal systems (LSA) and in particular to avoid critical situations.

[0040] One aspect of the present invention relates to the rapid and adaptive responsiveness of an urban traffic flow control system to changes in actual traffic flows, particularly when considering all traffic modes. Unforeseen disruptions, such as a burst water pipe on a main thoroughfare or an accident, are addressed almost reflexively through rapid adaptation of local traffic signal systems.

[0041] Another aspect concerns comprehensive analytics for the traffic area under consideration. The organic integration of traffic data collection, automated traffic pattern recognition, and control algorithms via the functionalities of 5G telecommunications networks represents a significant improvement.

[0042] With the ubiquity of smartphones and the increasingly precise positioning capabilities of these devices, a data source is available that is agnostic to the mode of transport; that is, pedestrians, cyclists, motorists, and public transport users are recorded equally. While this description explicitly discusses smartphones, this term actually serves only for illustrative purposes and does not represent a limitation of the subject matter of the invention. Rather, the term encompasses any device that has a connection to a mobile network, preferably via a SIM card, and can transmit position or movement data.

[0043] The fifth generation (5G) telecommunications networks currently under development offer, in addition to reduced latency and higher data throughput rates, decentralized computing functions in the so-called edge cloud and the option of virtualized network operation for the first time.

[0044] According to the invention, this control topology is used for tactile urban traffic flow management. Instead of transporting data via latency-prone paths to a central data processing center and processing it there in globally oriented algorithms, the control application according to the invention can be distributed; suitable modules can be executed in decentralized computing units with short response times to the control unit (here: LSA). Furthermore, 5G offers the option of network slicing. Separate virtual telecommunications networks can be configured, which, according to the invention, make it possible to combine several LSAs equipped with a communication module, e.g., a SIM card, in a regionally or sovereignly defined area and thus make them accessible to a distributed control algorithm.

[0045] The invention claims a method for recording traffic flows in a traffic network according to claim 1.

[0046] In one aspect of the invention, position and / or movement data from mobile communication devices are recorded in the telecommunications network and / or processed in the node-near computing device of the telecommunications network during the acquisition step.

[0047] According to the invention, in the acquisition step, based on the geographical position of the node, at least one virtual measuring point is provided which corresponds to a conventional acquisition system.

[0048] In one aspect of the invention, in the traffic flow determination step, a modal split is determined based on machine learning (ML) techniques, whereby the detected position and / or preferably a trajectory is assigned to a specific type of road user; and wherein, in the traffic flow detection step, the traffic flow is determined based on the modal split.

[0049] According to the invention, for the control of traffic flows in a traffic network, the traffic network having nodes, lines and / or subnetworks, wherein in a local control step a local control algorithm is executed and a signal adjustment for at least one traffic control device at at least one node, line and / or subnetwork of the traffic network is determined on the basis of a traffic flow at the node, line and / or subnetwork that was detected according to one of the preceding aspects; and wherein the local control step is carried out in a node-near computing unit of the telecommunications network, namely in the same node-near computing unit in which the detection is also carried out.

[0050] According to the invention, the virtual measuring point is configured to provide input parameters for the local control algorithm that correspond to the input parameters of the conventional acquisition system.

[0051] In one aspect of the invention, the node-near computing device is an edge computing device of the telecommunications network, preferably a 5G edge unit of the telecommunications network, and / or wherein the steps of the method are carried out in a network slice of the telecommunications network.

[0052] In one aspect of the invention, the signal adaptation relates to at least one of the following parameters: cycle time, release time, number of phases, phase sequence and offset time.

[0053] In one aspect of the invention, the local control algorithm performs at least one of the following optimizations: network optimization, line optimization, or local node optimization.

[0054] In one aspect of the invention, the method involves performing a central control step in at least one central control unit, in which several signal adjustments from several local control steps are coordinated. In another aspect of the invention, a decentralized control step is performed, in which several signal adjustments from several local control steps are coordinated.

[0055] Furthermore, a computing device of a telecommunications network according to claim 9, which is configured to carry out the method according to one of the above aspects, is claimed as being according to the invention.

[0056] Furthermore, a system according to claim 10 is claimed as being based on the invention, comprising at least one computing device as described above and at least one central control device configured to carry out the method according to one of the aspects above.

[0057] Furthermore, a mobile communication device according to claim 11 is claimed as being configured to transmit its position and / or movement data to a previously described computing device and to participate in the method according to one of the aspects above.

[0058] Furthermore, a computer program product according to claim 12 is claimed as being according to the invention, which causes at least one of the following, a previously described computing device, a previously described system, or a previously described mobile communication terminal, to perform the steps of the method according to one of the above aspects. Short description of the characters

[0059] The figures show: Fig. 1 A schematic diagram of a control architecture according to the prior art, Fig. 2 A schematic diagram of a centralized control system according to the prior art, Fig. 3 A schematic diagram of a decentralized control system according to the prior art, Fig. 4 A schematic diagram of a semi-centralized control system according to the prior art, Fig. 5 Schematic diagram of a 5G edge-based node or subnetwork control system in an embodiment of the invention, Fig. 6 Control model within a subnetwork of an edge in an embodiment of the invention, Fig. 7 Distribution of the modules in a processing step of the LSA control system in an embodiment of the invention, and Fig. 8 A schematic diagram of a control architecture in an embodiment of the invention. Description of the characters

[0060] The present invention relates to a data acquisition method with which a significantly more precise representation of current traffic flows can be created while simultaneously extending it to all modalities involved in traffic. Thus, the data acquisition method according to the invention provides a basis for new control methods with correspondingly more precise modeling.

[0061] One aspect of the present invention is to map existing control methods in an optimized manner onto the computational and communication architecture of future telecommunications networks, thereby significantly improving their effectiveness.

[0062] The basis for traffic control, or rather the corresponding control of the traffic signal system (LSA) in the case of adaptive control considered here, is the detection of traffic flows at the local intersections. According to the invention, this detection is based on position determinations from mobile communication devices within a mobile communication network. Using preferably continuous position determination, traffic flows can be detected not only at specific points but also, if required, as trajectories.

[0063] According to the invention, in order to take existing control models into account, virtual measuring points are defined analogously to measuring points from the prior art, for example to permanently installed loops, in order to be able to provide compatible input variables.

[0064] In one embodiment of the invention, further virtual measuring points are defined alternatively or additionally.

[0065] In one embodiment of the invention, ML techniques are used to determine a modal split from the observed data in order to identify the different road users, i.e., car, bicycle, public transport, pedestrian, etc.

[0066] According to the invention, the processing of the position data takes place near the nodes in the computing device of the telecommunications network, preferably in edge computing devices, especially in 5G edges, instead of in the nodes or in the control units assigned to the nodes.

[0067] In the following, the terms telecommunications network computing facilities, edge computing facility, 5G edge, and edge are used synonymously. A 5G edge is a decentralized computing capacity within the mobile network (edge ​​compute). Depending on the topology of the mobile network and the associated "range" or spatial coverage of the edge, multiple nodes, up to and including subnetworks, are controlled within an edge.

[0068] According to the invention, the existing control algorithms are moved from the nodes (for the node control level) or from the traffic control center (for line or subnetwork control level) to the 5G Edge, depending on the control level, i.e. node, line or subnetwork.

[0069] Fig. 5 Figure 1 shows a schematic diagram of a 5G edge-based node or subnetwork control system in one embodiment of the invention. Decentralized and centralized tasks, or rather the microscopic and macroscopic levels, are considered here in one diagram. The strategic level remains in the traffic control center 100.

[0070] The position and movement data 501, also known as MNO data, determined in the mobile network, and the classically recorded data 202 are transmitted to the Edge 400. The Edge 400 executes the data processing for the node 401 and a local control algorithm 402. Based on the information and specifications from the tactical / subnetwork control algorithm 404 and the currently observed traffic flow data from the local data processing 401, the local control algorithm 402 determines a temporary signal program 403 for the respective node 200. For the duration of its validity, the temporary signal program 403 overrides the standard signal program 300, which is specified by the traffic computer 100. The other elements correspond to the elements of the controls in the Figures 2 to 4 .

[0071] In one embodiment of the invention, if sufficient data is available in the form of movement data 501 generated by the mobile network, the use of classical data acquisition 202, i.e., sensors, loops, etc., can be dispensed with. This embodiment offers a cost advantage, as, among other things, no sensors need to be milled into the asphalt.

[0072] The following description explains the mode of operation of various embodiments of the invention using an example and Fig. 6 clarified.

[0073] In Figure 6Figure 1 shows a control model within a subnetwork 900 of an Edge 400 in one embodiment of the invention. The various nodes 200.1 to 200.n correspond to nodes of the traffic network. These nodes 200.1 to 200.n are each connected by lines and together form a subnetwork. A distinction is made between first-order neighboring nodes, i.e., direct neighbors, which are represented as triangles, i.e., nodes 200.2, 200.4, 200.5, and 200.7, and between second-order nodes, which are represented as circles, i.e., nodes 200.1, 200.3, 200.6, and 200.8. Each node 200.n has an LSA switching device 102.n, which is connected to the Edge 400 via a cellular connection 1000.

[0074] At a junction 200.n with significant pedestrian traffic occurring in massive bursts, for example at a public transport stop, the detection of mobile communication devices 501 and a technology based on ML for the recognition of traffic modalities according to the invention determines whether there is a need to increase the clearance time for pedestrians.

[0075] In embodiments of the invention, previously observed movement patterns (historical data) from extended periods are incorporated into the demand detection calculations and processed together with current observations using machine learning to generate a forecast. The detection or forecast of such an event then leads to the calculation of input parameters for the control algorithms implemented locally at the edge, which determine the signal adjustment themselves.

[0076] The adjusted switching signals are then transmitted from the Edge 400 to the switching devices 102.n of the local nodes 200.n. No further calculations are necessary within the node itself. Thanks to the system's real-time predictive capability, switching diagrams can be adjusted in advance, i.e., at the moment a potential flow disturbance occurs, thus optimizing the entire flow.

[0077] An advantage of the invention is that early intervention based on the forecast leads to a smaller number and / or a smaller scope of signal plan adjustments, thus preventing a larger, more global network impact.

[0078] Assuming that a short-term local adjustment of the signal program can strongly influence the local traffic flow, but that the effects decrease rapidly with increasing distance from this node, according to the invention, the Edge is configured based on current observations of the movement flows of the different traffic modes and the signaling plan subsequently adapted for the neighboring nodes (1st order, see Fig. 6 The system predicts whether significant disturbances are to be expected in neighboring nodes due to the node's adaptation. If so, an adaptation also occurs there. In other words, subnetwork control takes place.

[0079] Fig. 7Figure 1 shows the modules of the processing system and their interaction with the strategic level of the traffic signal control system in the traffic control center (TC) 100. Starting with the acquisition of traffic flow data via position data from smartphones 501 within the mobile network 500, pattern recognition using machine learning techniques 801 is first performed in the Edge 400 in the new module group 800 according to the invention to create a modal split of the road users. This data is then processed on the one hand into current movement data 802 and on the other hand processed and stored for later use as historical data 803. In the next processing step, a forecast for the local node is determined in module 804, which is based on the current data 802 and historical data 803 and the machine learning techniques applied thereto.From this data, input parameters for the control algorithms are generated in the next processing module 805 and passed to the node signaling algorithm 701. According to the invention, the node signaling algorithm 701, together with the subnetwork signaling algorithm 702, is implemented as a software solution in the Edge 400. The results of module 805 are processed in parallel in module 806 to generate a forecast for the global impact on the neighboring nodes (see explanatory notes). Figure 6 , processed and entered as input parameters into the processing of the subnet control 702.

[0080] The subnetwork control module also has an interface to the strategic traffic control 103 in the traffic control center 100, in order to take into account the strategic requirements of the entire network, as explained in the preceding sections. Within the module group, an adapted signal program 703 is then determined and distributed via the communication module 807 to the relevant node switching devices 102.m in the respective nodes 200.m for execution.

[0081] According to the invention, fast telecommunications networks with integrated computing capacities are linked with a hierarchical multimodal inner-city traffic flow management system, which relates locally triggered control actions to an overarching traffic flow optimization.

[0082] Embodiments of the invention enable efficiency gains through virtualization. The computing capacities integrated into the telecommunications network make it possible to eliminate local control systems (controllers), e.g., for an LSA, as separate hardware instances. According to the invention, the software components inherent in these systems are instead executed in data processing resources that are integrated into the telecommunications network (see figure). Figure 7 and the explanations above.

[0083] This approach abstracts away specific controller characteristics, such as manufacturer-defined interfaces, thereby realizing considerable technical and economic benefits through significantly easier scaling. Furthermore, the virtualized implementation of control functions according to the invention offers considerably easier maintenance, as software updates can be performed without on-site installation.

[0084] Up to now, traffic flow sensors have been geared towards motor vehicles. However, according to the invention, all mobile devices, preferably smartphones or similar devices with inherent positioning capabilities, are used for traffic flow detection. This allows for the undifferentiated consideration of various modes of transport, such as cyclists, pedestrians, public transport users, and others, provided they are carrying mobile devices. This makes it possible, for the first time, to implement flow optimizations based on the traffic pressures of various modes of transport.

[0085] In embodiments of the invention, the assignment of individual mobile devices to a mode of transport is achieved using artificial intelligence (AI) methods, such as pattern recognition (su). The recording of traffic flows via mobile networks eliminates the rigid dependence on fixed measuring points, such as those currently provided by induction loops and other sensor technologies.

[0086] In embodiments of the invention, virtual measuring points can be defined almost arbitrarily, thus enabling a flexible selection of the models used for traffic optimization. This also generates a further technical advantage and cost savings, as construction work for installing sensors, e.g., induction loops in the asphalt, can be avoided.

[0087] The use of a 5G telecommunications network according to the invention opens up new possibilities for telecommunications network-based solutions with regard to latency, data security, and resilience. The decentralized computing options within the network enable new levels of efficiency in data storage and data flow. Furthermore, the potential chain of errors is shortened, thus reducing the risk of failure.

[0088] With regard to the local control of a traffic control system, e.g. a traffic signal system, this means that, according to the invention, a secure service can be offered from the 5G network that reacts almost instantaneously to detected changes in the traffic situation.

[0089] The efficiency gains resulting from virtualizing the control architecture in ICT (information and communication technology) implementation have already been described above. In embodiments of the invention, a robust hierarchical system consisting of overarching overall system control and local control actions is implemented. Since traffic flow is always a stochastic phenomenon, overall system control must allow for fluctuations with respect to descriptive parameters such as flow densities or vehicle following intervals. A sensitivity analysis with respect to these local fluctuations then provides a framework for local control actions. Overall system optimization is then achieved through a control loop of locally autonomously executed actions, their evaluation with respect to the overall system, and, if necessary, adjustments at the local level.

[0090] A virtualization process according to the invention and the technical parameters of 5G telecommunications networks enable new embodiments of this adaptive control method. Some of the distinguishing technical features are already implicitly included in the above list; here they are explicitly stated again: 1. Virtualization of local traffic control and traffic signal management functionalities in data processing instances of a telecommunications network. 2. Virtualization of data acquisition through location tracking in the mobile network, resulting in reduced installation and maintenance costs. 3. Automated local traffic flow control by influencing traffic signal programs using artificial intelligence methods, in contrast to currently used methods, which are typically model-based. 4. Establishment of a control loop consisting of locally autonomous traffic control actions and an overarching overall system control, and its implementation in data processing resources integrated into the telecommunications network. 5. Integration of non-autonomous traffic flows (especially pedestrian and bicycle traffic) into overall optimization via mobile telecommunications devices.

[0091] Fig. 8Figure 1 shows a schematic diagram of a control architecture in one embodiment of the invention. Compared to the prior art, the architectural solution according to the invention also shows the central traffic computer 100. This computer still has a data processing module 101, but no longer a module for controlling the local accounts. However, the traffic computer remains responsible for global, strategic network control 103 via the corresponding module.

[0092] The traffic network is divided into subnetworks 900.1 to 900.n, each served by an Edge Controller 950.1 to 950.m. Each Edge Controller then controls the local nodes 200.xy assigned to the subnetwork. Within each Edge Controller 950.x are the data processing modules 951.x and the node and subnetwork control modules 952.x, which determine the signal programs for the local nodes 200.xy.

[0093] The processing steps for signal program generation are related to Figure 7 This has been described in detail. The input data for data processing is provided by the 500 mobile network and the data acquisition module 501 located there. Additional data can also come from the local data acquisition modules 202 in the respective nodes. These include, for example, demand requests via buttons at pedestrian traffic lights or priority switching for public transport, which are recorded locally.

[0094] Within each node, there are also, as already mentioned in connection with Figure 1 A control unit 203 is described. Within the scope of this invention, this unit is implemented here as a pure switching device and has no intrinsic intelligence for processing data and generating or adapting signal programs. The switching units 202 merely execute the switching commands provided by the edge-based node controller 952.x.

[0095] Although the invention has been illustrated and described in detail by the exemplary embodiments, the invention is not limited by the disclosed examples and other variations can be derived from them by the person skilled in the art without leaving the scope of protection of the invention. Literature and sources

[0096] [1] Friedrich, B. (2002): Traffic-adaptive control of traffic signal systems - An overview. (http: / / ivh.ivs.bau.tu-bs.de / peb / de / Mitarbeiter / friedrich-Dateien / veroeffentlichungen / Verkehrsadaptive%20Steuerung%20von%20Lichtsignalanlagen.pdf) [2] FGSV (2010): Guidelines for traffic signal systems (RiLSA). Research Association for Roads and Transportation (ed.), FGSV-Verlag GmbH, Cologne, 2010 [3] BAST (2013): Traffic-adaptive network control. Reports of the Federal Highway Research Institute, Issue V 230, 2013. [4] Wagner, P. (2009): Method and device for controlling a traffic signal system. German Patent and Trademark Office, granted patent DE102009033431B4 [5] Lämmer, S. (2007): Draft rule for decentralized online control of traffic signal systems in the road traffic network [6] Oertel, R.; Ermann, J.; Hesse, T.; Mann, A.; Markowski, R.; Trumpold, J.; Wagner, P. (2017): Control of traffic signal systems based on vehicle information - implementation and field tests.HEUREKA '17 - Optimization in Traffic and Transport (http: / / elib.dlr.de / 105545 / 1 / HEUREKA%2717_VITAL online.pdf) [7] Kobbeloer, D. (2007): Decentralized control of traffic signal systems in urban traffic networks. Kassel dissertation, Traffic Series, Issue 18, July 2007. [8] AMONES (2010): Application and analysis of model-based network control methods in urban road networks. Final report on the AMONES research project, 2010. List of reference symbols

[0097] 100 Central traffic computer 101 Data processing unit 101a Line / network data processing 101b Intersection data processing 102 Intersection control unit 102.1 ... n LSA switching device 1 to n 102a Network / line control 103 Strategic traffic control 200.1...8 Nodes 1 to 8 200.m Nodes m 200.1.1...200.1.n Nodes 1.1 to 1.n 200.m.1...200.mn Nodes m.1 to mn 201 Control unit 201a Signal groups 201b Signal protection 201c Intersection data processing 202 Data acquisition 203 Switching device without program function 300 Signal program 400 5G Edge 401 Intersection data processing 402 Local control algorithm 403 Temporary signaling program 404 Tactical / subnetwork control algorithm 500 MNO 501 Position data 600 LSA switching device 600.1...8 LSA switching device at node 1 to 8 600.m LSA switching device at node m 700 existing control modules that were moved to the edge 701 Node signal control algorithm 702 Subnetwork signal control algorithm 703 adapted switching signals 800 new control modules 801 Machine learning pattern recognition 802 current movement data 803 historical data 804 forecast of local nodes using machine learning 805 calculation of input parameters for control algorithm 806 forecast global evaluation 807 communication module 900 subnetwork 900.1...m subnetwork 1 to m 950.m 5G edge controller 951.m data processing 952.m node / subnetwork control 1000 cellular connection.

Claims

1. A method for recording traffic flow in a traffic network, wherein the traffic network comprises nodes, lines and / or subnetworks with at least one traffic guidance device, wherein in a recording step carried out in a telecommunication network, which geographically overlaps with at least part of the traffic network, a position and / or a trajectory of at least one road user at at least one node is recorded; wherein in a traffic flow determination step, which is carried out in a node-proximate calculation device of the telecommunication network, a traffic flow is determined from the detected positions and / or trajectories; wherein in a local control step a local control algorithm is executed and a signal adjustment for the at least one traffic guidance device at the node, the line and / or the subnetwork of the transport network is determined on the basis of the traffic flow determination step; wherein the local control step is carried out in the node-proximate calculation device of the telecommunication network in which the recording step was also carried out; wherein in the recording step at least one virtual measurement point which corresponds to a conventional recording system is provided, based on the geographical position of the node, and wherein the virtual measuring point is configured to provide input parameters for the local control algorithm which correspond to input parameters of the conventional recording system.

2. The method according to claim 1, wherein in the recording step, position and / or movement data of mobile radio terminals are recorded in the telecommunication network and / or are recorded and processed in the node-proximate calculation device of the telecommunication network.

3. The method according to any one of claims 1 to 2, wherein in the traffic flow determination step, based on machine learning techniques, a modal split is determined in which the recorded position and / or preferably a trajectory is assigned to a specific type of road user; and wherein in the traffic flow recording step the traffic flow is determined based on the modal split.

4. The method according to any one of claims 1 to 3, wherein the decentrally positioned calculation device of the telecommunication network with short response time to the light signal system LSA is an edge computing means of the telecommunication network, preferably a 5G edge unit of the telecommunication network, and / or wherein the steps of the method are carried out in a network slice of the telecommunication network.

5. The method according to any one of claims 1 to 4, wherein the signal adjustment relates to at least one of the following parameters: circulation time, release time, phase number, phase sequence and offset time.

6. The method according to any one of claims 1 to 5, wherein the local control algorithm carries out at least one of the following optimisations: network optimisation, line optimisation or local node optimisation.

7. The method according to any one of claims 1 to 6, wherein, in the method, a central control step is carried out in at least one central control device, in which a plurality of signal adjustments from a plurality of local control steps are coordinated with each other.

8. The method according to any one of claims 1 to 7, wherein a decentralised control step is carried out in which a plurality of signal adjustments from a plurality of local control steps are coordinated with each other.

9. A calculation device of a telecommunication network which is configured to carry out the method according to any one of claims 1 to 6.

10. A system from at least one calculation device according to claim 9 and at least one central control device configured to carry out the method according to claim 7 or 8.

11. A mobile radio terminal configured to transmit its position and / or movement data to a calculation device according to claim 9 and to participate in the method according to any one of claims 1 to 8.

12. A computer program product which causes at least one of the following: a calculation device according to claim 9, a system according to claim 10 or a mobile radio terminal according to claim 11 to carry out the steps of the method according to any one of claims 1 to 8.