EP4111434A1 - Verfahren und einrichtung zur restzeitprognose einer signalphase - Google Patents
Verfahren und einrichtung zur restzeitprognose einer signalphaseInfo
- Publication number
- EP4111434A1 EP4111434A1 EP21704446.0A EP21704446A EP4111434A1 EP 4111434 A1 EP4111434 A1 EP 4111434A1 EP 21704446 A EP21704446 A EP 21704446A EP 4111434 A1 EP4111434 A1 EP 4111434A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- traffic
- network
- signal
- data
- tna
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims description 30
- 238000013528 artificial neural network Methods 0.000 claims abstract description 73
- 230000000295 complement effect Effects 0.000 claims abstract description 12
- 238000012549 training Methods 0.000 claims description 27
- 230000007935 neutral effect Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 4
- 238000004393 prognosis Methods 0.000 claims description 2
- 230000006399 behavior Effects 0.000 description 9
- 230000006870 function Effects 0.000 description 8
- 210000002569 neuron Anatomy 0.000 description 8
- 230000001537 neural effect Effects 0.000 description 7
- 101000739577 Homo sapiens Selenocysteine-specific elongation factor Proteins 0.000 description 6
- 102100037498 Selenocysteine-specific elongation factor Human genes 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 230000001133 acceleration Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 3
- 239000003344 environmental pollutant Substances 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 231100000719 pollutant Toxicity 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 239000000446 fuel Substances 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 210000004205 output neuron Anatomy 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 229940000425 combination drug Drugs 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
- G08G1/0116—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from roadside infrastructure, e.g. beacons
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
- G08G1/0141—Measuring and analyzing of parameters relative to traffic conditions for specific applications for traffic information dissemination
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/09—Arrangements for giving variable traffic instructions
- G08G1/096—Arrangements for giving variable traffic instructions provided with indicators in which a mark progresses showing the time elapsed, e.g. of green phase
Definitions
- a method is known from the publication EP3438942A2 in which switching times of a signal system or remaining times until the signal system is switched over are predicted by means of artificial intelligence.
- By means of this information when approaching a signal system, road users can be prompted to adapt their speed, initiate braking processes early on or avoid unnecessary acceleration processes. In this way, fuel consumption, pollutant emissions, noise emissions and waiting times can often be effectively reduced.
- the described artificial intelligence usually requires very computationally intensive training and a large amount of training data.
- the object of the present invention is to create a method for influencing traffic, a traffic influencing device and a method for training them, which allow a better remaining time forecast or require less training.
- traffic data in an environment of a traffic signal generator and a signal phase specification differing from different signal phases of the traffic signal generator are recorded for influencing traffic.
- the traffic data are fed as input data to an artificial neural network, which comprises a first subnetwork and a different second subnetwork as well as a combination network for combining output data from both subnetworks.
- the artificial neural network is trained to use traffic data to reproduce a remaining time until a phase change of the traffic signal generator.
- an output of the output data of the first sub-network and an output of the output data of the second sub-network are controlled in a manner complementary to one another as a function of the signal phase specification.
- output data of the combination network or forecast data derived therefrom are transmitted as a remaining time forecast for influencing traffic to a means of transport or to a road user.
- traffic data of the surroundings of a traffic signal transmitter and a signal phase specification that distinguishes different signal phases of the traffic signal transmitter are recorded.
- the traffic data is fed as input data to the artificial neural network, which comprises a first subnetwork and a different second subnetwork as well as a combination network for combining output data from both subnetworks.
- the artificial neural network is trained to use the traffic data to reproduce a remaining time until a phase change of the traffic signal generator.
- an output of the output data of the first sub-network and an output of the output data of the second sub-network are controlled in a manner that is complementary to one another as a function of the signal phase specification.
- the method according to the invention and the traffic control device according to the invention can be executed or implemented, for example, by means of one or more processors, computers, application-specific integrated circuits (ASIC), digital signal processors (DSP) and / or so-called "Field Programmable Gate Arrays" (FPGA) .
- ASIC application-specific integrated circuits
- DSP digital signal processors
- FPGA Field Programmable Gate Arrays
- the quality of a remaining time forecast can generally be improved and / or the required training effort can be reduced.
- noise emissions, waiting times, pollutant emissions and / or traffic delays can in turn be reduced.
- only minor or no modifications to existing traffic engineering systems are required to use the invention.
- the output of the output data of the two subnetworks can be controlled as a function of the signal phase specification in such a way that in a first signal phase the first subnetwork value a prognosis and the second subnetwork output a neutral value to the combina tion network, and that in a second signal phase the second sub-network output a forecast value and the first sub-network output a neutral value for the combination network.
- a numerical or logical zero or a very small value can be output as the neutral value which, when combined with a quantified forecast value, does not change it or changes it only slightly.
- the combination network can preferably combine output data from the first sub-network with output data from the second sub-network by means of preferably numerical addition.
- a neutral value is added to a quantified forecast value, the latter does not change or changes only slightly.
- a signal phase-specific forecast value of the first subnetwork and in a second signal phase a signal phase-specific forecast value from the second subnetwork can each be output or selected as the result of the addition.
- a signal-generator-specific signal phase information can be recorded for several traffic signal generators.
- signal transmitter-specific forecast values for traffic signal transmitters in a first signal phase and signal transmitter-specific forecast values for traffic signal transmitters in a second signal phase can be output to the combination network by the first subnetwork. In this way, signal transmitter-specific forecast values for several or all traffic signal transmitters involved can be determined and output in parallel.
- the signal phase information and / or other data influencing a switching behavior of the traffic signal generator can be fed to the artificial neural network as input data, the artificial neural network being trained to use signal phase information and / or other data that influence the switching behavior of the traffic signal generator to reproduce a respective phase change.
- data influencing a switching behavior of the traffic signal transmitter can in particular give time, date information, environmental data, e.g. about weather conditions, lighting conditions, smoothness or pollution, information about peak times and / or event data about events, about the approach of prioritized vehicles. information, accidents or other traffic incidents.
- an automatic start-stop system, a brake, a recuperation device, an autonomous vehicle, a navigation device and / or a route planner can be controlled or a message sent to the Road users are issued.
- means of transport or road users can be prompted to adapt their speed, initiate braking processes early, avoid unnecessary acceleration processes or otherwise react to the expected phase change.
- the artificial neural network can be trained to assign a mean value, a median, a quantile, a probable value, a minimum, a maximum and / or a statistical fluctuation range of the remaining time or a probability for a phase change in a given time interval as a remaining time forecast reproduce.
- the neural network can be trained to output a probability of a change to green in the next second as a remaining time forecast.
- Figure 1 shows a street intersection with a traffic signal system
- FIG. 2 shows a neural network according to the invention
- FIG. 3 shows a traffic control device according to the invention.
- FIG. 1 shows, in a schematic representation, an example of a road junction KR with a traffic signal system which comprises several traffic lights S1,..., S4 as traffic signal generators. Traffic lights that are activated in phase each form a signal group.
- S1 and S3 form a first signal group and S2 and S4 form a second signal group.
- the traffic signal generators S1, ..., S4 are each used to control or regulate the local traffic depending on current traffic data or other influencing factors.
- a respective traffic signal generator Sl, ... or. S4 differentiated two mutually complementary signal phases. That is, if a first signal phase is currently not present, then the second and vice versa. A first signal phase “green” for free travel and a complementary second signal phase “not green” for blocked can be provided.
- the traffic signal generators S1, ..., S4 and their signal phases and phase changes are each controlled by a connected Am pelêtung CTL.
- the latter can be implemented locally or as part of a cross-signaling traffic control system.
- the traffic signal generators S1, ..., S4 are controlled by the traffic light control CTL as a function of locally recorded traffic data or other influencing factors.
- the traffic lights at the KR intersection control which vehicle has free travel and which vehicle has to wait.
- the traffic light S1 is decisive for the vehicle F1 and the traffic light S2 for the vehicle F2.
- a traffic influencing device RSU according to the invention is also arranged in the vicinity of the intersection KR.
- the latter is used to influence the means of transport, here in particular the vehicles Fl and F2 or other road users.
- internal control of the means of transport or a smartphone can be influenced by cyclists or pedestrians.
- the traffic influencing device RSU should determine a remaining time forecast, i.e. predict a remaining time until a subsequent phase change of a respective traffic signal generator.
- a probability of a phase change to green or non-green in the next second or in another predetermined time interval can also be determined as a remaining time forecast.
- means of transport here F1 and F2 or other road users, when approaching the traffic signal system, can be prompted to adapt their speed, initiate braking processes early on or avoid unnecessary acceleration processes.
- an automatic start / stop system, a brake, a recuperation device, a navigation device, or an autonomous vehicle controller can be controlled in a vehicle or a message can be given to a road user.
- the traffic influencing device RSU determines signal transmitter-specific or signal group-specific remaining time forecasts RP1 and RP2 and transmits them to vehicles F1 and F2 to influence them.
- RP1 quantifies a predicted remaining time until the next phase change of the first signal group S1, S3 and RP2 a predicted remaining time until the next phase change of the second signal group S2, S4.
- the traffic influencing device RSU can preferably be implemented as a so-called roadside unit.
- the traffic control device RSU can be wholly or partially integrated or coupled to a so-called SPaT box (SPaT: Signal Phase and Timing) be.
- SPaT box Signal Phase and Timing
- the remaining time forecasts RP1 and RP2 of the traffic control device RSU can be transmitted to a so-called onboard unit or another control device in a means of transport or to a smartphone of a road user. Remaining time forecasts can also be sent to pedestrians or cyclists in particular via a smartphone.
- the remaining time forecasts RP1 and RP2 are determined on the basis of traffic data VD, signal phase information SPA and possibly other data influencing a switching behavior of the traffic signal generators S1,.
- traffic data VD the number and speeds of vehicles, their waiting times, the approach of prioritized vehicles, traffic events such as accidents or other traffic disruptions, or other information about a current traffic load can be recorded as traffic data VD.
- the traffic influencing device RSU has a sensor system S, which can include, for example, vehicle sensors, cameras, speed sensors or other sensors. Alternatively or additionally, the traffic influencing device RSU can have a receiving device for otherwise recorded data influencing the switching behavior.
- the signal phase information SPA indicate for a respective traffic signal generator or for a respective signal group in which signal phase the respective traffic signal generator or the respective signal group is currently located. Alternatively or additionally, a respective point in time of a last phase change can also be quantified in a signal transmitter-specific or signal group-specific manner by means of a respective signal phase specification.
- the current signal phase information SPA is available in the traffic light control CTL and is transmitted continuously or in the event of phase changes from the traffic light control CTL to the traffic control device RSU.
- the other data influencing the switching behavior of the traffic signal system can in particular include time information, date information, environmental data, e.g. on weather conditions, light conditions, smoothness or pollution, information on peak times and / or event data about events, approaches of prioritized vehicles, accidents or others Traffic incidents include.
- the traffic control device RSU has an artificial neural network NN that is coupled to the sensor system S.
- the neural network NN is trained to use the traffic data VD, the signal phase information SPA and possibly the other data influencing the switching behavior of the traffic signal generators S1, ..., S4 to determine the remaining times up to a respective phase change of the traffic signal generators S1, .. ., S4 to forecast.
- the traffic control device RSU also has a transmitting device TX coupled to the neural network NN.
- the remaining time predictions RP1 and RP2, which are preferably recalculated every second, are transmitted to the vehicles F1 and F2, preferably on a radio basis, by the transmitting device TX.
- the invention can also advantageously be used for influencing or controlling rail traffic, robots, ship traffic or air traffic in order to determine remaining time forecasts for traffic signal generators used there, for example radio-based traffic signal generators, in an analogous manner, and to communicate them to the relevant traffic tel or road users.
- FIG. 2 illustrates a neural network NN according to the invention.
- the same or corresponding reference symbols are used in FIG. 2 as in FIG. 1, these reference symbols denote the same or corresponding entities activities that can be implemented or designed in particular as described above.
- the neural network is formed by several neural layers A to M. Data propagation between the neural layers A to M is illustrated in FIG. 2 by arrows.
- the neural network NN comprises, in particular, two neural subnetworks TNA and TNB, the first neural subnetwork TNA being formed by the layers A, E, G, I and K and the second neural subnetwork TNB being formed by the layers C, F, H, J and L will.
- the two subnetworks TNA and TNB are separated from one another in the present exemplary embodiment and have separate, mutually independent data propagation paths. In particular, no neurons are shared. Alternatively, the subnetworks TNA and TNB can also be only partially separated, so that the two subnetworks TNA and TNB use individual neurons together.
- the neural network NN also has an input network EN. This is used to receive input data from the neural network NN and to preprocess them and to pass on the preprocessed input data to both subnetworks TNA and TNB.
- the input network EN is formed by the input layer B and the hidden layer D.
- the input layer B is supplied with traffic data VD, signal phase information SPA and other data (not shown) influencing a switching behavior of the traffic signal generator.
- the input data supplied can be configured as described in connection with FIG.
- the preprocessed input data, ie the output data of the input network EN are fed from layer D to both layer E of the first sub-network TNA and to layer F of the second sub-network TNB.
- the neural network NN furthermore comprises a combination network which, in the present exemplary embodiment, is formed by the individual layer M, which at the same time functions as an output layer of the neural network NN.
- the combination network M is used to combine output data from the subnetworks TNA and TNB.
- one output neuron of the layer M is provided for a respective signal transmitter or a respective signal group.
- Each of these signal transmitter-specific or signal group-specific output neurons outputs a quantified remaining time value, here RP1 or RP2 for a respective traffic signal transmitter or a respective signal group.
- the layers A to M can each comprise one or more neurons.
- a connection between every two layers is indicated in FIG. 2 by arrows. With such a connection between two layers, each neuron of an output layer can potentially be connected to each neuron of a target layer.
- the subnetwork TNA with the hidden layers E, G, I and K is used in the present exemplary embodiment for remaining time forecasts of signal groups currently in the "green” signal phase, while the subnetwork TNB with the hidden layers F, H, J and L is used for remaining time forecasts of signal groups currently in the "non-green” phase.
- both layer I and layer J generate residual time forecasts for all signal groups, but with layer I only those residual time forecasts that are assigned to those signal groups have an influence on the end result that are currently in the "green” signal phase.
- layer J only those remaining time forecasts that are assigned to the signal groups in the "non-green” signal phase have an influence on the end result.
- This different, more precisely complementary consideration of the different signal phases by the subnetworks TNA and TNB is carried out or enforced by the layers K and L in conjunction with the layers A and C.
- Layer K takes over the remaining time forecasts from layer I only for those signal groups that are currently in the "green” signal phase from layer I, while for those signal groups that are currently in the "non-green” signal phase each have a neutral value, e.g. a numeric zero is entered.
- the information required for this distinction is provided by layer A.
- the latter outputs, depending on the supplied signal phase information SPA for all signal groups that are currently in the "green” signal phase, a numerical or logical one to the layer K and a logical or numerical zero for those signal groups that are currently in the Signal phase "non-green".
- the remaining time forecasts of layer J are only adopted by layer L for those signal groups that are currently in the "non-green” phase, while for those signal groups that are currently in the "green” phase, a neutral value is entered.
- the information required for this distinction is supplied in an analogous manner by the layer C as a function of the supplied signal phase information SPA.
- the combination network M summarizes the forecast results of the layers K and L, preferably by simply adding them together, so that the combination network M contains the remaining time forecasts of all signal groups, regardless of whether a respective signal group is currently in the signal phase "green” or "non-green". is located.
- the combination network M outputs a remaining time forecast RP1 for the signal group consisting of the traffic lights S1 and S3 and a remaining time forecast RP2 for the signal group consisting of the traffic lights S2 and S4 as output data of the neural network NN.
- the normal learning process of the neural network has an effect NN to the effect that the subnetwork TNA is specifically optimized for remaining time forecasts for the green phase and the subnetwork TNB specifically for remaining time forecasts for the non-green phase.
- FIG. 3 shows a learning-based traffic control device RSU according to the invention in a schematic representation. If the same or corresponding reference symbols are used in FIG. 3 as in the previous figures, these reference symbols denote the same or corresponding entities which, in particular, can be implemented or configured as described above.
- the traffic influencing device RSU has a processor PROC for executing the method according to the invention as well as a memory MEM coupled to the processor PROC for storing data arising during execution.
- the traffic influencing device RSU also has a sensor system S.
- the sensor system S is used to measure or otherwise record traffic data VD or other data influencing a switching behavior of a traffic signal system.
- traffic data VD traffic data
- the latter are interpreted below as traffic data and added to the traffic data VD.
- the traffic influencing device RSU has an artificial neural network NN coupled to the sensor system S.
- the neural network NN is supplied with the traffic data VD from the sensor system S and signal phase information SPA as input data.
- the neural network NN is to be trained to use the supplied traffic data VD and signal phase information SPA signal group-specific and signal phase-specific remaining time forecasts, here RP1 and RP2 to be determined and output as output data.
- the traffic influencing device RSU comprises a transmitting device TX coupled to the neural network NN for preferably radio-based transmission of determined remaining time forecasts, here RP1 and RP2 to means of transport, here vehicles F1 and F2 and / or to other road users.
- the neural network NN has an input network EN, neural sub-networks TNA and TNB and a combination network M.
- the traffic data VD and the signal phase information SPA are fed to the input network EN as input data.
- the input network EN preprocesses the input data VD and SPA and transmits the preprocessed input data to both neural subnetworks TNA and TNB.
- the subnetworks TNA and TNB are to be trained implicitly to determine signal phase-specific forecast values that each quantify a remaining time until the next phase change.
- forecast values PIA and P2A are output from the subnetwork TNA and forecast values P1B and P2B from the subnetwork TNB as output data.
- the forecast values PIA, P2A, P1B and P2B are selected and output by the subnetworks TNA and TNB in each case for a specific signal phase.
- the sub-network TNA has a selection network SELA and the sub-network TNB has a selection network SELB.
- the selection network SELA can include the layers K and A described in FIG. 2 and the selection network SELB the layers L and C.
- the selection networks SELA and SELB are each supplied with the signal phase information SPA in order to select and output the forecast values PIA, P2A, P1B and P2B in a signal phase-specific manner as a function thereof.
- the selection network SELA outputs a numerical zero as a non-green phase-specific forecast value P2A.
- the selection network SELB outputs the non-green phase-specific forecast value P2B of the sub-network TNB for all signal groups located in the non-green phase according to the signal phase information SPA.
- the selection network SELB outputs a numerical zero as a green phase-specific forecast value P1B.
- the output data of the subnetworks TNA and TNB are therefore set to zero by the selection networks SELA and SELB depending on the signal phase information SPA signal phase-specific in a mutually complementary manner. That is, in the green phase of a respective signal group, a quantified forecast value is output by the sub-network TNA and - complementary to this - a forecast value from the sub-network TNB is suppressed to a certain extent. Analogously to this, in the non-green phase, a quantified forecast value of the sub-network TNB is output and - complementary to this - a forecast value of the sub-network TNA is, as it were, suppressed.
- the forecast values PIA, P2A, P1B and P2B are output from the subnetworks TNA and TNB to the combination network M, a quantified forecast value being passed on to the combination network M in each signal phase for all signal groups, either from the subnetwork TNA or from the subnetwork TNB. A zero is output from the other subnetwork.
- the combination network M serves to combine output data from the subnetworks TNA and TNB and to output the remaining time forecasts, here RP1 and RP2.
- the combination network M adds the forecast values PIA and P2A of the subnetwork TNA to the forecast values P1B and P2B of the subnetwork TNB in a signal phase-specific and signal group-specific manner.
- the forecast values PIA and P1B are added for the green phase and the forecast values P2A and P2B for the non-green phase.
- the complementary combination of the output data of the subnetworks TNA and TNB by the combination network M has the effect that quantified forecast values of the subnetwork TNA in the green phase and in the non-green phase quantified forecast values of the sub-network TNB are output as respective quantified remaining time forecast RP1 or RP2. In this way, a quantified forecast value is always available for all signal groups.
- the neural network NN should first be trained in a training phase to use the input data VD and SPA supplied as training data to reproduce remaining times until the next phase change of the traffic signal generators, i.e. to output remaining time forecasts RP1 and RP2 as precisely as possible.
- reproducing means that a respective remaining time, which is predicted on the basis of input data available up to a respective point in time, should correspond as closely as possible to a phase change that actually occurs later.
- Training is generally understood to mean an optimization of a mapping of input data, here VD and SPA of a parameterized system model, here the neural network NN, to output data, here the remaining time forecasts RP1 and RP2.
- This mapping is optimized according to predetermined criteria that have been learned or are to be learned during a training phase.
- a minimization of a prediction error can be used as a criterion for prediction models.
- a network structure of neurons of the neural network NN and / or weights of connections between the neurons can be set or optimized in such a way that the specified criteria are met as well as possible.
- the training can thus be viewed as an optimization problem for which a large number of efficient optimization methods are available.
- the parameters optimized in this way, in particular the optimized connection weights between neurons can be stored as a training structure, output or transferred to another neural network in order to configure this in an optimized manner.
- the neural network NN to train the neural network NN, its output data RP1 and RP2 are compared with times of later actually occurring phase changes and time deviations At between prognosti ed phase changes and actually occurring phase changes are determined.
- the time discrepancies At can be formed by an amount or a square of a respective time difference.
- the times of the actual phase changes occurring can be determined using the signal phase information ben SPA.
- the time discrepancies At represent a prediction error of the neural network NN and are fed back to this.
- the neural network NN is trained - as indicated in FIG. 3 by a dotted arrow - to minimize the time deviations At on average. This training enables the neural network NN to determine relatively accurate forecasts of the remaining times to be expected until the next respective phase change.
- the training can also be carried out in whole or in part in a cloud or by another external computer.
- the training data here VD and SPA
- VD and SPA are transmitted there in order to train an external neural network according to the invention as described above.
- the training structure of the trained external neural network obtained in this way can then be obtained from the external neural network are transmitted to the traffic control device RSU in order to occupy their neural network NN with this training structure.
- the external neural network is, as it were, copied from the external computer to this traffic control device RSU.
- the signal phase-specific selection of the forecast values of the subnetworks TNA and TNB usually has the effect that changes in the parameterization of the respective subnetwork TNA or TNB for the forecast of non-green phases or green phases do not affect the output of the neural network NN and therefore not optimized in the course of training.
- the training of the sub-network TNA is not disturbed by data from the non-green phase and the training of the sub-network TNB is not disturbed by data from the green phase.
- This separation of the training paths is advantageous because the learning tasks in the different signal phases usually differ significantly from one another.
- the red phase is essentially fixed-time controlled and therefore only varies relatively slightly, while the green phase is often controlled as a function of the traffic and thus variably.
- the training paths in the neural network NN are so to speak switched on and off in signal phases or switched between the subnetworks TNA and TNB.
- the normal learning process of the neural network NN then ensures the signal phase-specific learning of the subnetworks TNA and TNB. It turns out that the described neural network NN in many cases delivers considerably better remaining time forecasts or requires considerably less training than conventional neural networks used for remaining time forecasts.
- the trained neural network NN can be used to influence traffic.
- current traffic data VD is recorded by the sensor system S and transmitted to the trained neural network NN together with current traffic data from the Am- pel control CTL deriving signal phase information SPA supplied.
- the trained neural network NN uses this to determine, as described above, remaining time forecasts RP1 and RP2, which are transmitted to the transmitter TX and from there to the means of transport F1 and F2 or to other road users.
- the neural network NN can preferably be further trained during operation using the currently recorded traffic data VD and the signal phase information SPA.
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Traffic Control Systems (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE102020202380.8A DE102020202380A1 (de) | 2020-02-25 | 2020-02-25 | Verfahren zur Verkehrsbeeinflussung und Verkehrsbeeinflussungseinrichtung |
PCT/EP2021/052624 WO2021170360A1 (de) | 2020-02-25 | 2021-02-04 | Verfahren und einrichtung zur restzeitprognose einer signalphase |
Publications (1)
Publication Number | Publication Date |
---|---|
EP4111434A1 true EP4111434A1 (de) | 2023-01-04 |
Family
ID=74586993
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP21704446.0A Pending EP4111434A1 (de) | 2020-02-25 | 2021-02-04 | Verfahren und einrichtung zur restzeitprognose einer signalphase |
Country Status (4)
Country | Link |
---|---|
US (1) | US20230080193A1 (de) |
EP (1) | EP4111434A1 (de) |
DE (1) | DE102020202380A1 (de) |
WO (1) | WO2021170360A1 (de) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102019219748A1 (de) * | 2019-12-16 | 2021-06-17 | Siemens Mobility GmbH | Verfahren zum Bestimmen mindestens eines zu bestimmenden Restzeitwerts für eine Anlage |
TW202409639A (zh) * | 2022-08-25 | 2024-03-01 | 緯創資通股份有限公司 | 用於配置抬頭顯示器的電子裝置和方法 |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102013204241A1 (de) | 2013-03-12 | 2014-09-18 | Bayerische Motoren Werke Aktiengesellschaft | Verfahren und Vorrichtung zur Ermittlung eines erwarteten Umschaltzeitpunkts einer Signalgruppe |
PL3333823T3 (pl) * | 2016-12-07 | 2024-06-24 | Yunex Gmbh | Przewidywanie sygnalizacji urządzenia sygnalizacji świetlnej za pomocą sztucznej inteligencji |
US10652464B2 (en) | 2017-07-31 | 2020-05-12 | Honeywell International Inc. | Systems and methods for automatically switching a surveillance camera into an auto corridor mode |
DE102017213350A1 (de) | 2017-08-02 | 2019-02-07 | Siemens Aktiengesellschaft | Verfahren zur Vorhersage eines Schaltzeitpunktes einer Signalgruppe einer Signalanlage |
DE102017223579A1 (de) | 2017-12-21 | 2019-06-27 | Siemens Aktiengesellschaft | System und Verfahren zum Unterstützen einer Prognose einer zukünftigen Signalisierung eines Verkehrsinfrastrukturelements |
-
2020
- 2020-02-25 DE DE102020202380.8A patent/DE102020202380A1/de not_active Withdrawn
-
2021
- 2021-02-04 WO PCT/EP2021/052624 patent/WO2021170360A1/de unknown
- 2021-02-04 EP EP21704446.0A patent/EP4111434A1/de active Pending
- 2021-02-04 US US17/802,224 patent/US20230080193A1/en active Pending
Also Published As
Publication number | Publication date |
---|---|
WO2021170360A1 (de) | 2021-09-02 |
US20230080193A1 (en) | 2023-03-16 |
DE102020202380A1 (de) | 2021-08-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3365741B1 (de) | Verfahren zur vollautomatischen führung eines fahrzeugsystems und kraftfahrzeug | |
DE102019104974A1 (de) | Verfahren sowie System zum Bestimmen eines Fahrmanövers | |
DE102016205152A1 (de) | Fahrerassistenzsystem zum Unterstützen eines Fahrers beim Führen eines Fahrzeugs | |
DE102016205153A1 (de) | Verfahren zum Erzeugen von Steuerdaten für ein regelbasiertes Unterstützen eines Fahrers | |
WO2009037164A1 (de) | Erstellung von qualitätsparametern einer digitalen karte | |
DE102019108644A1 (de) | Verfahren und vorrichtung zum automatischen lernen von regeln für autonomes fahren | |
DE102008023972A1 (de) | Verfahren und Vorrichtung zur Erkennung von verkehrsrelevanten Informationen | |
DE10354322A1 (de) | Verfahren und System zur Ermittlung der Fahrsituation | |
WO2021170360A1 (de) | Verfahren und einrichtung zur restzeitprognose einer signalphase | |
DE102019002790B4 (de) | Verfahren zur Prädiktion einer Verkehrssituation für ein Fahrzeug | |
DE102018203583B4 (de) | Verfahren, Fahrerassistenzsystem sowie Kraftfahrzeug zur Prädiktion einer Position oder einer Trajektorie mittels eines graphbasierten Umgebungsmodells | |
EP3688538A1 (de) | Verfahren und system zum aktualisieren eines steuerungsmodells für eine automatische steuerung zumindest einer mobilen einheit | |
DE102017213350A1 (de) | Verfahren zur Vorhersage eines Schaltzeitpunktes einer Signalgruppe einer Signalanlage | |
DE102015212027A1 (de) | Verfahren und Vorrichtung zum automatischen Ermitteln von Verkehrsregelungen an Straßenkreuzungen | |
DE102015223656A1 (de) | Fahrerassistenzsystem und -Verfahren zur Fahrspurempfehlung | |
DE102017209258A1 (de) | Verfahren und Vorrichtung zum Überwachen einer Fahrstabilität eines Fahrzeugs auf einer vorausliegenden Fahrroute | |
EP2662846A1 (de) | Verfahren zum Reduzieren einer Staugefahr | |
WO2021110370A1 (de) | Verfahren und anordnung zur vorhersage von schaltzeitpunkten einer signalgruppe einer signalanlage zur steuerung eines verkehrsflusses | |
DE102018221044A1 (de) | Verfahren und Vorrichtung zur Prognose eines Schaltzustands und eines Schaltzeitpunkts einer Signalanlage zur Verkehrssteuerung | |
DE102018202498A1 (de) | Verfahren und Steuergerät zum Betreiben eines Antriebsstrangs | |
EP3402126A1 (de) | Verfahren und system zum verteilten erkennen von objekten und gefahren | |
DE102023201567A1 (de) | Verkehrssteuerung für Mischverkehr | |
DE112022002031T5 (de) | Bordinformationsverarbeitungsvorrichtung, system für autonomes fahren und bordsystem | |
DE102021107796A1 (de) | Verfahren zur Ermittlung eines wahrscheinlichsten Fahrpfades eines Kraftfahrzeug | |
DE102021002004A1 (de) | Verfahren zum Betrieb von Fahrzeugen innerhalb eines Platooning Konvois |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
17P | Request for examination filed |
Effective date: 20220916 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
DAV | Request for validation of the european patent (deleted) | ||
DAX | Request for extension of the european patent (deleted) | ||
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: EXAMINATION IS IN PROGRESS |
|
17Q | First examination report despatched |
Effective date: 20240814 |