Emissions estimation and anomaly

Abstract

To relate to emissions estimation and anomalies.SOLUTION: A computer-implemented method comprises: clustering sets of traffic data on the basis of similarity to each other to generate a plurality of clusters of the sets of traffic data; selecting, as a representative set of traffic data, the set of traffic data of the cluster which is most similar to an average; and on the basis of, a representative set of traffic data corresponding to a target geographical region, performing mobility analysis to determine parameters for a periodic traffic forecast and, on the basis of at least one representative set of traffic data corresponding to at least one other geographical region, performing mobility analysis to determine parameters for at least one other periodic traffic forecast for the at least one other geographical region; performing a traffic forecasting process for the target geographical region; and predicting emissions produced by traffic in the target geographical region on the basis of the combined traffic forecast.SELECTED DRAWING: Figure 19

Description

[Technical field]

[0001] The present invention relates to emission estimation, and in particular to a computer implemented method, a computer program and an information programming apparatus. [Background technology]

[0002] There is a tendency to represent the real world as a digital world. Concepts like Smart Cities, Digital Twins (DT) and the Metaverse are gaining more attention, at least in part, due to new / improved technologies that enable new steps in these concepts, especially in data, security, the Internet of Things (IoT), the fifth generation technology standard for telecommunications (5G), Artificial Intelligence (AI) and Quantum Computing.

[0003] CO2 emissions from transportation modes, i.e., the release of carbon dioxide into the atmosphere as a result of human transportation activities, contribute significantly to the overall greenhouse gas emissions that drive climate change. According to the International Energy Agency, transportation accounted for about 24% of the world's energy-related CO2 emissions in 2019. Traffic forecasting, the process of predicting future traffic conditions in the near-term or short-term future based on current and past traffic observations, is important in understanding and dealing with traffic in order to reduce CO2 emissions from transportation modes. Short-term traffic flow forecasting, which involves the prediction of traffic volume in the next time interval, usually ranging from 5 minutes to 1 hour, is one of the important research problems in the field of traffic congestion and has been addressed by many researchers in the past two decades (e.g., Kumar, SV, Vanajakshi, L. Short-term traffic flow prediction using seasonal ARIMA model with limited input data. Eur. Transp. Res. Rev. 7, 21 (2015). https: / / doi.org / 10.1007 / s12544-015-0170-8).

[0004] In light of the above, improved estimation of emissions is desirable. Summary of the Invention

[0005] According to an embodiment of a seventh aspect, the present disclosure provides a computer-implemented method, comprising: executing a parameter determination process based on a plurality of (historical) traffic data sets for a plurality of geographical areas, said parameter determination process comprising: clustering the traffic datasets based on their similarity to each other to generate a plurality of clusters of traffic datasets, and for each cluster including a plurality of traffic datasets, selecting a traffic dataset of the cluster that is most similar to an average of the traffic datasets of the cluster as a representative traffic dataset; Based on a representative traffic data set corresponding to a target geographical area, performing a mobility analysis to determine (optimal) parameters for a periodic traffic forecast for said target geographical area, and based on at least one representative traffic data set corresponding to at least one other geographical area, performing a mobility analysis to determine (optimal) parameters for at least one other periodic traffic forecast for at least one other geographical area; and performing a traffic prediction process for the target geographic region using a traffic dataset for the target geographic region and at least one traffic dataset for at least one other geographic region, the traffic prediction process comprising: generating the nodal traffic forecast for the target geographical area (by estimating the periodic component of traffic data for the target geographical area) based on the traffic dataset for the target geographical area and using the determined parameters for periodic traffic forecast for the target geographical area (predicting traffic data for a future period); generating a periodic traffic forecast for the at least one other geographical area (by estimating a periodic component of traffic data for the at least one other geographical area) based on the at least one traffic data set for the at least one other geographical area and using the determined parameters for the at least one other periodic traffic forecast for the at least one other geographical area (predicting traffic data for a future period); analyzing mobility flows in the geographical area based on a traffic dataset of the geographical area and at least one traffic dataset of the at least one other geographical area to determine at least one correlation between the (traffic data trend of) the target geographical area and the at least one other geographical area; generating a combined traffic forecast (or final forecast or cumulative forecast or overall forecast) by adjusting the periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; and forecasting emissions generated by traffic in the target geographic area based on the final traffic forecast; A computer-implemented method is disclosed, comprising:

[0006] Features associated with any aspect / embodiment may be applied to any other aspect / embodiment. [Brief description of the drawings]

[0007] By way of example, reference is made to the accompanying drawings in which: [Figure 1] This is a diagram that is useful for understanding predictions. [Diagram 2] FIG. 1 illustrates a system and architecture. [Diagram 3] FIG. 1 illustrates a system. [Figure 4] FIG. 1 illustrates a system. [Diagram 5] FIG. [Figure 6]FIG. [Figure 7] FIG. [Figure 8] 1 is a diagram useful for understanding an embodiment. [Figure 9] FIG. [Figure 10] 1 is a diagram useful for understanding an embodiment. [Figure 11] 1 is a diagram useful for understanding an embodiment. [Figure 12] 1 is a diagram useful for understanding an embodiment. [Figure 13] FIG. [Figure 14] 1 is a diagram useful for understanding an embodiment. [Figure 15] 1 is a diagram useful for understanding an embodiment. [Figure 16] 1 is a diagram useful for understanding an embodiment. [Figure 17] This is a diagram that is useful for understanding predictions. [Figure 18] FIG. 1 illustrates a system and architecture. [Figure 19] FIG. 1 illustrates a system. [Figure 20] FIG. 1 illustrates a system. [Figure 21] FIG. [Figure 22] FIG. [Diagram 23] 1 is a diagram useful for understanding an embodiment. [Figure 24] FIG. 1 illustrates graphs that are useful for understanding embodiments. [Diagram 25] FIG. [Figure 26] FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Embodiments disclosed herein can address the problem of estimating emissions, e.g., CO2 emissions, i.e., how to monitor past, present, and future emissions. Embodiments can detect anomalous peaks in emissions and, e.g., apply countermeasures to assist in achieving Net Zero Targets (NZT).

[0009] Conventional techniques for monitoring traffic and / or emissions have drawbacks, including the following: Traditional approaches do not provide periodic, coarse-grained forecasts in high-demand systems, such as real-time environments that require zero-latency responses. Traditional forecasts don’t take into account context and don’t account for complex patterns (e.g., correlated regions, citizen movement flows, future events, etc.). Traditional approaches treat CO2 anomalies as CO2 levels, but do not analyze them in detail as a mobility component, nor do they consider cycle- and trend-independent aspects.

[0010] Aspects disclosed herein enable large scalability for monitoring past, present, and future emissions. These can assist in achieving Net Zero Targets (NZT) by detecting anomalous peaks in emissions and applying countermeasures, for example, in digital twin (DT) solutions that require managing multiple areas simultaneously with very short response times.

[0011] Further disadvantages of conventional techniques for monitoring traffic and / or emissions include the following: Traditional approaches to DT solutions require processing 100 to 1000 domains simultaneously, which means managing thousands of complex models in near real-time.

[0012] Aspects disclosed herein can achieve advantages including the following. · Scalable periodic prediction for DT systems. · Context-aware prediction and anomaly detection at multiple levels, the key to simulation and digital rehearsal. More accurate detection of abnormal behavior by not relying on cycles or trends. · Large scale implementation of DT systems. Reduced complexity, which can be considered important to manage the required number of models, such as in a real-time DT environment, where many domains need to be processed simultaneously with short response times.

[0013] The following terms may be used in the description. The definitions are not exhaustive. 5G: The fifth generation technology standard for telecommunications ACF: Auto Correlation Function AI:Artificial Intelligence AR: Auto Regressive ARIMA: Auto Regressive Integrated Moving Average BN: Batch Normalisation CNN: Convolutional Neural Network DL: Deep Learning DT: Digital Twin DX: Digital Transformation FC: Fully Connected Network IoT: Internet of Things. ITS: Intelligent Transportation System MA: Moving Average LSTM:Long-Short Term Memory NZT: Net Zero Targets PACF: Partial Auto Correlation Function SARIMA:Seasonal Auto Regressive Integrated Moving Average SARIMAX: Seasonal Auto Regressive Integrated Moving Average with exogenous features SDLTFP: Supervised Deep Learning Based Traffic Flow Prediction SVM: Support Vector Machine SVR: Support Vector Regression

[0014] Methods based on time series analysis, such as Autoregressive Integrated Moving Average (ARIMA), have been reported to be one of the most accurate methods for forecasting traffic flows (e.g. Mai T, Ghosh B, Wilson S (2012) Multivariate short-term traffic flow forecasting using Bayesian vector autoregressive moving average model. Proceedings of the 91 st Transportation Research Board Annual Meeting, Washington, DC).

[0015] Time series models attempt to identify patterns from historical data by decomposing long-term trends and periodic patterns, and then extrapolate the patterns into the future. Since traffic flow patterns exhibit strong periodic patterns due to peak and off-peak traffic conditions that repeat at more or less the same times each day, the Periodic ARIMA (SARIMA) model is said to be particularly suitable for modeling traffic flow behavior (e.g., Ghosh B, Basu B, Mahony MO (2007) Bayesian time-series model for short-term traffic flow forecasting. J Transp Eng 133(3):180-189). In many studies, SARIMA models have been found to perform better than models based on random walks, linear regression, support vector regression (SVR), historical averages, and simple ARIMA (e.g., Williams BM, Hoel LA (2003) Modeling and forecasting vehicular traffic flow as a seasonal ARIMA process: theoretical basis and empirical results. J Transp Eng 129(6):664-672, in Ghosh B, Basu B, Mahony MO (2005) Time-series modelling for forecasting vehicular traffic flow in Dublin. Proceedings of the 85th Transportation Research Board Annual Meeting, Washington, DC, and in Lippi M, Bertini M, Frasconi P (2013) Short-term traffic flow forecasting: an experimental comparison of time-series analysis and supervised learning. IEEE Trans intell Transp Syst 14(2):871-882).

[0016] ARIMA models are characterized by three terms: p, d, and q. p is the order of the AR term, q is the order of the MA term, and d is the number of differences to make the time series stationary. In SARIMA models, the s term refers to a periodic parameter, usually representing the number of observations in a period. Furthermore, SARIMA can also be implemented with exogenous features, resulting in a SARIMAX model (for example, as described in Khandelwal, R., 2020: Time series prediction using SARIMAX). These exogenous features have a different origin than the original time series. This means that the features describe characteristics of the observations that are different from the time series. The addition of these features can lead to more accurate forecasts of the time series. An example of these added features is a list of weather or events, each with their own impact, occurring in a particular region of interest.

[0017] The standard method in literature and research to assign correct values ​​to the parameters p, d, and q is to use Autocorrelation Function (ACF) and Partial Autocorrelation (PACF) plots. An ACF plot is a bar plot of the correlation coefficient between a time series and its lags. This helps to determine the value of the p or AR term. A PACF plot is a plot of the partial correlation coefficient between a time series and its own lags. This helps to determine the value of the q or MA term.

[0018] Some aspects disclosed herein can be considered as additional extensions of the SARIMAX model, which may have the ability to scale up to multiple domains for prediction and anomaly detection.

[0019] Many companies and governments are currently engaged in deep digital transformation processes to evolve traditional business for the future. In this transition, digital transformation is being done in order to emulate the real world, mimic behavioral models, and apply a real-world vision of the future.

[0020] Aspects disclosed herein may facilitate prediction of traffic emissions and detection of anomalous behavior in DT solutions. Conventional prediction approaches may not be able to address certain constraints or provide certain capabilities to the DT environment, such as special capabilities that enable realistic digital rehearsals or advanced simulations in DT.

[0021] Aspects disclosed herein can accomplish the following: 1. Scalable periodic forecasting for high demand systems with small granularity of information and near real-time response. 2. Context-aware traffic forecasting at various levels, enabling digital rehearsal and pre-simulation aimed at detecting complex patterns that impact future emissions. 3. Cycle- and trend-independent anomaly detection, e.g. highlighting scenarios that penalize NZT achievement.

[0022] As disclosed herein, "traffic" may refer to vehicular traffic. The terms "mobility" and "traffic" can both be used to refer to traffic data from sensors and / or digital twins.

[0023] A first implementation will now be described, which includes a number of implementation examples and aspects.

[0024] Figure 1 shows a representation of a practical use case. An analysis of mobility in an area is performed to obtain insights on how mobility behaves. For such an analysis, specific data from the DT environment with historical and real-time data can be used. This DT data and mobility insights are used to obtain / generate predictions of mobility (traffic) at different levels, taking into account a broader view of mobility, with information such as traffic, correlation regions, citizen mobility flows, events, weather, cross-sectional transport modalities, etc. Then, detection of anomalous behavior of mobility (having a direct impact on pollutant emissions) is performed. These anomalies reveal scenarios to be further investigated: (1) a decision in the application of measures for potential emission reduction, or (2) an analysis of the scenarios to obtain best practices for the reduction of emission levels. Furthermore, an estimation of pollutant emissions is obtained based on the predicted traffic, including the impact of potential anomalies.

[0025] DT systems can provide early recognition of future emission levels and anomalous behavior by incorporating aspects disclosed herein. The following advantages can be achieved by the aspects disclosed herein: providing near real-time response for demanding systems such as DT environments, and reducing the complexity of the prediction problem (i.e., the models used and the calculations required).

[0026] 2 shows a general overview of an architecture incorporating the system 100. The system 100 can communicate with interfaces to receive (live) data from real-world sensors. The interfaces can also connect to a data storage layer where data from such sensors is stored over time. The interfaces may also be connected to simulation systems, for example to provide new insights into simulated scenarios.

[0027] System 100, in an example implementation, includes the following components having the following functionality:

[0028] The DT Mobility Analysis module 20 is responsible for providing parameters for fine-tuning the predictive model (as will be further explained below). The parameters may be called "optimal" parameters, e.g. to indicate that they are generated based on an analysis and are considered useful. This module receives live information from sensors deployed in the real world and historical information stored in the persistence layer (data storage) and executes a logic to obtain optimal settings (parameters) of the predictive model. As output, the optimal parameters are provided to the DT Mobility Prediction module 40.

[0029] The DT Mobility Prediction module 40 generates predictions, i.e. this module is in charge of predicting the behavior of different mobility trends. It uses data such as historical data, mobility flows, events, weather, etc. and performs the prediction after analysing all these aspects. As an output, this module provides a prediction of the traffic behavior (traffic prediction), including specific features used by the DT Mobility Anomaly Detection module 60 and the Emission Estimation Engine 80.

[0030] · The DT Mobility Anomaly Detection module 60 detects anomalous behavior in the predicted traffic (that may affect pollutant emissions). This module uses the predicted behavior from the DT Mobility Prediction module (and e.g. historical traffic data) as input and detects anomalous behavior patterns, regardless of cyclical elements or current trends in the mobility (traffic). As output, this module provides the detected anomalous behavior together with information about the pattern (expected behavior, type of anomaly predicted, deviations). This information is provided to the Emission Estimation Engine 80.

[0031] The Emissions Estimation Engine 80 estimates emissions such as CO2 or other pollutants based on the output of the DT Mobility Anomaly Detection Module 60 and the DT Mobility Prediction Module 40. This may include specific information about the traffic such as number of vehicles, vehicle type, passengers, driving behavior, etc. Furthermore, this module may be able to estimate the impact of the detected anomalous behavior and measure the impact of each anomaly in terms of sustainability. This module may provide a quantification of emissions for a given traffic situation.

[0032] FIG. 3 illustrates the system 100 and uses arrows to indicate process flow.

[0033] Figure 4 illustrates the system 100 and some of the input / communication capabilities. The DT-Mob footprint represents data that can be received from a digital twin of a geographical area and can include any of the following (each over time, i.e., time series data): number of vehicles, type of vehicle, speed, average speed, maximum speed, minimum speed, congestion level, traffic jams (number), utilization level of the road network, maximum capacity, etc.

[0034] FIG. 5 illustrates a method that may be performed by the system 100. Step S20 includes performing a traffic prediction process for a first time period. Step S40 includes performing a traffic prediction process for a second time period. Here, performing a traffic prediction process for a time period means performing a traffic prediction process using traffic data covering that time period. For example, the time period may be the past 90 days or the past year or more or less. The first time period includes the second time period and an additional time period that is more recent than the second time period. For example, the additional time period may be the previous hour or the previous day or the previous week, etc. In some respects, the (historical) traffic data covering the additional time period may be considered as "live" traffic data.

[0035] Traffic data as used herein (throughout the application, for the first and second implementations) may include data collected by real-world sensors and stored, for example, in a storage device (such as a storage layer). Alternatively or additionally, traffic data as used herein may include data from a DT. That is, traffic data for a geographical area may come from a DT that models / simulates that geographical area. The DT itself may utilize traffic data obtained from real-world sensors. Traffic data from a DT may have undergone several pre-processing steps (e.g., compared to traffic data obtained from real-world sensors), in which, for example, intermediate variables and values ​​may have been generated. That is, congestion levels for a specific area and a specific transport modality given the number of vehicles and the average speed of the vehicles may have been generated.

[0036] Furthermore, traffic data as used herein may, in some implementations, include traffic data from a DT that models / simulates specific scenarios (and, e.g., not just real world). That is, a DT can model / simulate scenarios such as, for example, intervention scenarios (i.e., what would happen (to the traffic data) if a specific intervention were made) and other "what if" scenarios. Such scenarios can be thought of as representing a mirror of the simulated near or future given a set of parameters to be simulated. Multiple DT scenarios (and thus multiple sets of corresponding traffic data) may be simulated in parallel.

[0037] As mentioned above, traffic data as used herein may represent / include traffic data obtained from real-world sensors (eg, via storage devices) as well as data from the DT.

[0038] Steps S20 and S40 result in a first traffic forecast and a second traffic forecast, respectively.

[0039] Step S60 includes decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component, and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component. Step S60 may include using a periodic trend decomposition (STL) technique with locally estimated scatterplot smoothing (LOESS).

[0040] Step S80 includes comparing the noise components to detect at least one anomaly, i.e., step S80 includes comparing a first noise component of the first traffic forecast to a second noise component of the second traffic forecast to detect at least one anomaly, step S80 includes comparing at least one deviation between the first noise component and the second noise component to an anomaly threshold, and classifying the deviation as an anomaly if the magnitude of the deviation is greater than the anomaly threshold.

[0041] Step S100 includes estimating emissions, the estimation being based on the first traffic forecast and may include an effect of at least one detected anomaly.

[0042] Figure 6 shows the method. Steps S22-S28 can be regarded as a traffic prediction process. Steps S20 and S40 of the method in Figure 5 can each include the traffic prediction process of steps S22-S28.

[0043] Step S22 includes generating a periodic traffic forecast for the target area. Step S22 includes generating at least one other periodic traffic forecast for at least one other area, the area being a geographic area of ​​a plurality of geographic areas within the geographic area. The generation of the periodic traffic forecast is described below.

[0044] Step S26 includes analyzing mobility flows of the target region and at least one other region to determine a correlation between the target region and the at least one other region.

[0045] Step S28 includes adjusting the periodic traffic forecast for the target region based on at least one other forecast for at least one other region and based on the determined correlation. The adjusted forecast may be called a combined traffic forecast, a final forecast, an accumulated forecast, or an overall forecast, and may be considered as a final traffic forecast for the target region. These terms may be used interchangeably herein (in the first and second embodiments).

[0046] Step S101 includes estimating emissions based on the adjusted / final forecast. The method may or may not include step S101.

[0047] DT Mobility Analysis Module 20 FIG. 7 illustrates a specific implementation of the DT mobility analysis module 20. As shown in FIG.

[0048] This module receives information / data from real-world sensors and (historical) traffic data that may be stored in a data storage layer. This module analyzes the behavior of the DT and finds optimal parameters for the predictive model. The DT mobility analysis module 20 can be considered to receive as input the DT-Mob footprint and "historical" data that may provide all the information required from the DT (as mentioned above, the traffic data may be received from sensors and / or storage and / or the DT). The mobility analysis module 20 analyzes the received information, identifies the current trends of mobility and sets optimal parameters for the predictive model. This module has the ability to analyze mobility from different angles (e.g., either number of vehicles, speed, average speed, maximum speed, minimum speed, level of congestion, traffic jams (number), level of use of the road network, maximum capacity, etc.). The number of vehicles (i.e., number of vehicles or number of vehicles of each of multiple types) and (average) speed may be preferred in some examples (e.g., because these aspects may be used to calculate emissions). The type of vehicle may correspond to either engine size, average emissions, and manufacturer.

[0049] The optimal parameters may be determined by testing runs using a set of parameters, comparing the metrics and error for each configuration, and selecting the best one (i.e., selecting the parameters that give the smallest error).

[0050] The DT mobility analysis module 20 may be considered to perform part of step S22 and part of step S24 in the sense that generating a periodic forecast may include first analyzing relevant traffic data to determine optimal parameters for the periodic forecast.

[0051] As can be seen in FIG. 7, in this implementation, the DT mobility analysis module 20 has four components.

[0052] Stationarity Checker 22 This component analyzes traffic data to obtain different levels of stationarity of annual trends. That is, the component provides the variation of historical trends (trends) depending on the period of the year. To obtain optimal parameters, the module can use methods such as Canova-Hansen to determine the optimal degree of periodic differencing.

[0053] Regression Checker 24 This component analyzes historical data added to the current live information of mobility. It can be thought of as simply analyzing traffic data. This module establishes the relationship between the historical data, the current mobility trends, and future predictions. It can be thought of as determining the relationship between past mobility trends, more recent mobility trends, and tentatively predicted mobility trends. Finally, in this implementation, this component provides the number of "waves" that the analyzed mobility trends will follow to achieve an accurate prediction.

[0054] Examples of methods that can be used with this module are the Kwiatkowski-Phillips-Schmidt-Shin, Augmented Dicky-Fuller, or Phillips-Perron methods.

[0055] Prediction Checker 26 This component establishes the range of values ​​within which the current mobility trend is oscillating. In other words, this module determines one or more parameters that establish the degree of magnitude of future values ​​or the rate of increase / decrease in the mobility trend. Some techniques for determining these parameters include the autocorrelation function (ACF), partial autocorrelation function (PACF), and / or extended autocorrelation function (EACF) methods.

[0056] To find the optimal parameters, these modules 22, 24, 26 use techniques such as the Akaike Information Criterion, the Modified Akaike Information Criterion, the Bayesian Information Criterion, the Hannan-Quinn Information Criterion, or "out-of-bag" for validation.

[0057] Optimal Parameter Engine 28 This component aggregates stationary, regression, and forecast parameters (eg, it can set and order parameters in the correct fields for use in the next stage).

[0058] 8 shows example data for the DT mobility analysis module 20. A shows an example of data contained in the DT-mob footprint. B shows an example of traffic data stored in the data storage. Both A and B together can be considered traffic data. C shows an example of parameters aggregated together by the optimal parameter engine 28.

[0059] The optimal parameters may be provided to the DT mobility prediction module 40 .

[0060] DT Mobility Prediction Module 40 FIG. 9 illustrates a specific implementation of the DT mobility prediction module 40.

[0061] The DT mobility prediction module 40 predicts the behavior of various mobility trends. This module uses data from datasets such as historical mobility, mobility flows, events, and weather (collectively sometimes referred to as traffic data) as input and performs predictions after analyzing such aspects. As output, this module provides a prediction of traffic behavior (traffic prediction), including specific features used by the DT mobility anomaly detection module 60 and the emission estimation engine 80. Mobility data can be considered as traffic data.

[0062] The DT mobility prediction module 40 of this implementation shown in FIG. 9 has five components.

[0063] Regressor Builder 41 The Regressor Builder 41 contributes to scalable periodic forecasting. This component analyzes historical mobility (i.e. traffic) data and generates regressors that define the periodic aspects of the forecast. These regressors are provided to the Periodic Forecasting Model module 44. Thus, the computational cost of estimating the periodic component of the forecast is decoupled from the generation of the periodic forecast and is performed by the Regressor Builder 41 instead of the Periodic Forecasting Model module 44. This allows reducing the complexity and computational cost. The Regressor Builder 41 can provide the regressors to the Periodic Forecasting Model module 44 as Fourier coefficients. The Regressor Builder 41 can use techniques such as Fourier transforms, sinusoidal decompositions, and functions of time and frequency.

[0064] Periodic Prediction Model Module 44 The Periodic Prediction Model module 44 is responsible for the periodic prediction of mobility. This module uses historical traffic information (i.e. traffic data) and regressors. Using Fourier coefficients as regressors may reduce the training complexity. During the training phase, the parameters of the model (periodic prediction model) can be configured based on the training dataset. The model has several components: stationary / periodic and non-stationary / non-periodic. The regressors can be considered to provide information of the periodic component. This simplifies the model setup since only the non-stationary / non-periodic part needs to be trained, reducing the time and complexity of the setup / training of the periodic prediction model.

[0065] The use of regressors also reduces the complexity during the execution of the model. The model is defined by the following equation: The Periodic Prediction Model module 44 uses the optimal parameters provided by the DT Mobility Analysis module 20 with the aim of providing an optimal setup of the model used in the prediction / forecasting.

[0066] The model used by the periodic forecast model module 44 to generate the periodic forecasts may be defined as:

number

[0067] The generation of the regressors can be performed by the periodic forecast model module 44 instead of the regressor builder 41. Equations other than those above can be used to generate the periodic forecasts.

[0068] Mobility Flow Analysis Department 42 The mobility flow analysis unit 42 provides insight into how regions are related to each other, how the mobility of one region correlates with the mobility of other regions, and finally how the traffic forecasts of different regions correlate with each other. This component enables mobility flow analysis at three levels, for example as shown in FIG. 12. For example, correlations can be searched considering: (1) at the region level (i.e., within a geographical region), (2) considering adjacent regions, or (3) considering regions that influence the target region based on the mobility flow analysis. Finally, this module provides as a final result the regions that need to be considered for a realistic prediction. That is, the mobility flow analysis unit 42 can output information indicating at least one region whose traffic data influences the traffic data of the target region based on the analyzed traffic data. Also, the mobility flow analysis unit 42 determines at least one correlation and outputs this information, or the at least one correlation can be determined by the prediction engine 45.

[0069] Mobility flows can be defined as the impact of traffic data of one area on traffic data of any other area. Mobility flows can be defined as how people move from one area to another (e.g., their destinations) regardless of the transport modality.

[0070] To analyze mobility or origin-destination (OD) matrices from call detail records (CDRs), two approaches can be considered: (a) in time-based matrices, the analysis can focus on estimating mobility directly from the sequence of CDRs; (b) in routine-based matrices (OD by purpose), the analysis can focus on routine types of trips, such as home-work commutes, derived from trip generation models. Furthermore, the module can map OD coordinates for specific time periods to a territorial distribution.

[0071] The mobility flow analysis unit 42 can be considered to execute step S26.

[0072] External Event Analysis Section 43 The external event analyzer 43 provides insight into how external factors are taken into account for the final prediction. This component analyzes the correlation of such external factors (e.g. events, weather conditions, complex correlation patterns, or other factors affecting mobility, i.e. traffic information) with traffic information. Such correlations are sent to the prediction engine 45 and may be considered together with possible future events to adjust the traffic prediction.

[0073] The external event analysis unit 43 can analyze the correlation between events and several variables using techniques such as time series correlation (Pearson correlation method or linear regression method), network correlation (graphlet network distance, alignment-based methods, spectral methods, or other methods), and data mining methods.

[0074] Prediction Engine 45 The prediction engine 45 provides the final prediction output by the DT mobility prediction module 40. The prediction engine 45 receives information from the periodic prediction model module 44, the mobility flow analyzer 42, and the external event analyzer 43. The prediction engine 45 selects a periodic prediction for a target area and analyzes the influence of: (1) mobility flows, (2) behavior of surrounding / correlated areas, (3) future occurring events, (4) weather effects, and (5) correlations between other complex patterns that affect the mobility prediction, among other possible factors. The prediction engine 45 adjusts the periodic prediction based on any of these factors. The prediction engine 45 may use the following formula to calculate the adjusted prediction, which may be referred to as the final prediction: The prediction engine 45 may receive correlation weights, which may be determined in advance (e.g., through training and / or statistical interference) and may be used as coefficients in the following formula:

number

[0075] The prediction engine 45 can be considered to perform step S28.

[0076] Equations other than those shown above may also be used, for example equations that do not include all of the terms of the above equations, among other possible equations.

[0077] FIG. 10 shows some example data that may be input to the DT mobility prediction module 40. A shows an example of "historical traffic data". B shows an example of area information. C shows an example of "mobility flow" data. FIG. 11 shows further example data. A shows an example of external event data. B shows an example of data related to regressors. C shows an example of data related to traffic prediction (in this example, specifically the number of vehicles is displayed). The data input to the DT mobility prediction module 40 may simply be referred to as traffic data. That is, traffic data as used herein may include data related to events, mobility flows, area information, etc.

[0078] DT mobility anomaly detection unit 60 FIG. 13 illustrates a particular embodiment of the DT mobility anomaly detector 60. As shown in FIG.

[0079] The DT mobility anomaly detection module detects anomalous behavior in the predicted traffic that may affect pollutant emissions. The DT mobility anomaly detection unit 60 uses the historical traffic data and the specific predicted behavior from the DT mobility prediction module as inputs and detects anomalous behavior patterns agnostic to the periodic component or current trends of mobility. As shown in FIG. 13, the DT mobility anomaly detection unit 60 includes three modules in this implementation: two periodic decomposition modules 62, 64: one that decomposes the "expected" traffic behavior and the other that decomposes the "forecasted" traffic behavior; and an anomaly detection module 65 that detects anomalies based on the output of the two periodic decomposition modules 62, 64.

[0080] As an output, the DT mobility anomaly detector 60 can provide the detected anomalous behavior along with information about the pattern (expected behavior, predicted, type of anomaly, deviation). This information can be provided to the emission estimation engine 80.

[0081] Periodic decomposition modules 62, 64 Each cyclical decomposition module 62, 64 decomposes traffic behavior into three aspects / components: (1) cyclical components, (2) trends, and (3) noise.

[0082] The seasonal aspect / component contains information about recurring variations in the traffic data based on a selected time scale. That is, the seasonal component indicates recurring temporal patterns present in the traffic data based on a selected periodicity. The selected periodicity can be, for example, monthly, yearly, or weekly, among other periodicities.

[0083] The trend aspect / component contains information about the fluctuations of the traffic data over a longer period than the period defined by the periodicity. For example, if a monthly periodicity is selected, the trend component shows the long-term fluctuations of the traffic data over several months. The trend component can be considered (and defined) as the traffic data with the periodic component removed / subtracted or the traffic data with the periodic fluctuations removed.

[0084] The noise aspect / component can be referred to as the remainder component. The remainder component can be calculated by subtracting the values ​​of the cyclical and trend components from the traffic data. The remainder value indicates the amount of noise present in the traffic data. Values ​​close to zero indicate that the cyclical and trend components are accurate in describing the traffic data, while larger remainder values ​​indicate the presence of noise.

[0085] The periodic decomposition unit 62 uses the historical traffic behavior as input and decomposes the expected behavior into periodic, trend and noise components. The other periodic decomposition unit 64 uses the historical traffic data and the predictions from the DT mobility prediction module as input and decomposes the predicted behavior into periodic, trend and noise components.

[0086] This decomposition can be performed using, for example, Periodic Decomposition with LOESS (STL) (https: / / doc.arcgis.com / en / insights / latest / analyze / stl.htm). The output from both Periodic Decomposition modules 62, 64 is used as input to an anomaly detection module 65.

[0087] Anomaly Detection Module 65 The anomaly detection module 65 is responsible for highlighting potential anomalies detected based on the difference between the "expected" and "predicted" behavior. This estimation is done from a periodic and trend-independent perspective, where only the "noise" or "deviation" aspects / components of the behavior are analyzed. Thus, traffic fluctuations due to periodic or trend factors, which are normal and should not be considered as anomalies, are excluded from consideration when detecting anomalies.

[0088] As described above, the DT mobility anomaly detector 60 can receive as input the historical data (traffic data) and the predicted behavior. Alternatively, the DT mobility anomaly detector 60 can receive as input the first and second traffic predictions generated by the DT mobility prediction module 40, where the first traffic prediction is based on traffic data covering a first period and the second traffic prediction is based on traffic data covering a second period, the first period including the second period and an additional period newer than the second period, as described with respect to FIG. 5. The first traffic prediction can be considered as the "predicted" behavior and the second traffic prediction can be considered as the "expected" behavior. The second traffic prediction can be called the traffic-based prediction.

[0089] Figure 14 shows the first and second traffic forecasts in terms of how the number of vehicles in the area changes over time. The light grey dashed "Traffic Forecast" is the "predicted behavior" or first traffic forecast, and the dark grey / black dashed "Traffic Reference Forecast" is the "expected behavior" or second traffic forecast. The solid black line shows the actual traffic data up to a certain point in time (marked "now" in Figure 14). The graph in Figure 14 shows the deviation / noise component of the forecast / data after decomposition.

[0090] The anomaly detection module 65 compares the deviation / difference between the noise components of the first and second traffic forecasts with a threshold (anomaly threshold) and classifies the deviation as an anomaly if the magnitude of the deviation is greater than the threshold. A percentage difference can be used, i.e. the deviation as a percentage of the value of the noise component of the first or second traffic forecast at the time of interest.

[0091] The anomaly detection module 65 may provide as output the baseline traffic forecast, the traffic forecast, the detected anomalies, and the percentage of deviation for each anomaly. This output is provided to the emission estimation engine 80 to highlight the anomalies in terms of pollutants. The anomaly detection module 65 may detect anomalies based on different aspects of the traffic (e.g. any of the number of vehicles, speed (average), type of vehicles, number of vehicles of each type, road network density, maximum speed, minimum speed, traffic congestion level, number of traffic jams, etc.). In an embodiment, anomalies are detected based on the number of vehicles and the average speed of the vehicles.

[0092] The DT anomaly detection unit 60 can be considered to execute steps S60 and S80.

[0093] Figure 15 shows example data for the DT anomaly detector 60. A shows example "historical" traffic data. B shows example predicted behavior. Figure 16 shows example output data of the anomaly detection module 65.

[0094] Emissions Estimate Engine 65 The emission estimation engine 65 estimates emissions or other pollutant emissions based on predicted traffic information such as number of vehicles, vehicle type, passengers, driving behavior, etc. This module thus provides a prediction of future emissions based on the output of the DT mobility prediction module (i.e., the first traffic prediction). Any emission can be predicted. For example, any of the following emissions can be predicted: CO2, NO, NO2, CO, hydrocarbons, CH4, benzene, formaldehyde, chlorofluorocarbons, CHCIF2, radon. For example, any of the following emissions can be predicted: carbon dioxide (CO2), carbon monoxide (CO), nitrogen oxides (NO and / or NO2), methane (CH4), chlorodifluoromethane (CHClF2), radon (Rn). Any of the following emissions can be predicted instead of or in addition to CH4: Any of the following hydrocarbons can be predicted. Any of the following emissions can be predicted instead of or in addition to CHClF2:

[0095] Furthermore, this module estimates the impact of the abnormal behavior / anomalies detected by the DT mobility anomaly detector 60. That is, the emission estimation engine 65 measures the impact of each anomaly in terms of sustainability. This module provides a quantification of emissions for a given traffic situation.

[0096] In an example implementation, the emissions are calculated based on the following: information indicative of the average emissions of the vehicles based on speed, the number of vehicles from the traffic forecast, and the average speed of the vehicles from the traffic forecast. The emission estimation can also take into account different types of vehicles and can utilize, for each type of vehicle, information indicative of the average emissions based on the speed of that type of vehicle. The information indicative of the average emissions can be based on information provided by the vehicle manufacturer or determined by testing. Instead of or in addition to the average speed, an average behavior can also be used. The average behavior can include speed changes, acceleration, frequency of trips, etc.

[0097] A second implementation will now be described, which includes a number of implementation examples and aspects.

[0098] Aspects disclosed herein can facilitate management of large-scale solutions to forecast traffic emissions and, if necessary, detect anomalous behavior of the DT solution. Conventional techniques for traffic forecasting may not be able to address certain constraints and provide certain features for DT environments. Conventional techniques may struggle with large-scale solutions, i.e., handling many domains simultaneously, possibly with short response times.

[0099] Aspects disclosed herein can achieve the following benefits, among others: 1. Large scale, scalable execution for forecasting and anomaly detection for high demand systems. 2.Reduced complexity in large scale scenarios.

[0100] FIG. 17 shows a representation of traffic forecast for a large number of regions. In some cases, using traditional approaches with hundreds or thousands of regions to manage all the models required for DT and traffic forecasting is not optimal. In the presently disclosed aspects, the mobility of multiple regions is analyzed to gain insight into how the mobility is behaving. Data from the DT environment, along with "historical" and "real-time" data, can be used. Such data, along with mobility insights, can be used to obtain a classification of multiple regions covered and managed by the DT. This classification is inferred based on the mobility behavior patterns for each region. Each region is assigned a category, the number of categories being less than the number of regions managed.

[0101] For each category, a representative behavior is selected, which is a behavior pattern used to represent all areas in the category. This representative behavior can be determined by selecting the behavior pattern of the area that has the smallest difference from the behavior patterns of other areas of the same category or from the average of the behavior patterns (traffic data) of the area. That is, the areas can be selected to minimize the difference.

[0102] A relationship between the number of categories and the error introduced by the classification process can be determined (e.g., based on testing of the second implementation and / or an example implementation). If the number of categories is low, the error introduced is high and vice versa. A configuration parameter defining the number of categories and / or the "acceptable" error can be set (from which the number of categories can be determined based on the relationship). A closeness threshold associated with the number of categories and / or the "acceptable" error can be implemented. The closeness threshold can be used when clustering traffic datasets. This means that the closeness threshold can define the level of similarity that two traffic datasets must share to be grouped into the same category / cluster. In other words, in an example implementation, the similarity between two traffic datasets (e.g., between a centroid and another traffic dataset) can be compared to a closeness threshold, and if the similarity exceeds the threshold, the traffic datasets can be clustered / classified together, otherwise they can not be clustered / classified. A closeness threshold can be implemented instead of or in addition to the number of categories and / or the "acceptable" error.

[0103] Considering the errors introduced by the classification process, it can be said that the classification process reduces the complexity of the problem at the cost of introducing errors, which can be reduced / kept small to some extent (and thus the accuracy of the final output (emissions estimation) can be maintained to some extent) by selecting an appropriate number of categories.As mentioned before, the classification process realizes the management of large-scale execution in the DT environment.

[0104] The representative behavior is used to determine optimal parameters for periodic forecasting. Forecasting and anomaly detection are performed based on the determined optimal parameters (e.g., as described in relation to the first embodiment). This saves computational power and time compared to determining optimal parameters for forecasting each region individually.

[0105] The clustering / classification and selection of representative behaviors may be repeated at intervals or whenever new traffic data is acquired from the DT / real-world sensors.

[0106] Figure 18 shows a general schematic diagram of an architecture incorporating a system 300. The architecture and system 300 are similar to the architecture and system 100 of Figure 2, and a repeated description will be omitted.

[0107] The difference between system 100 and system 300 is that system 300 includes an additional component, a DT mobility classification module 310 .

[0108] The DT mobility classification module 310 supports the processing of a large number of regions. This module analyzes historical and live information (which can be collectively called traffic data), classifies multiple regions based on their mobility behavior patterns, and selects a number of representative regions that represent the behavior of all the considered regions. As an output, the representative regions are provided to the DT mobility analysis module 320 and the DT mobility prediction module 340.

[0109] Figure 19 illustrates a system 300 according to an example implementation. Figure 20 illustrates the system 300 and some input / communication capabilities. The DT-Mob footprint represents data that can be received from a digital twin of a geographical area and can include any of the following: number of vehicles, type of vehicle, speed, average speed, maximum speed, minimum speed, congestion level, traffic jams (number), utilization level of the road network, maximum capacity, etc.

[0110] FIG 21 illustrates the method. Step S212 includes clustering traffic data sets. Each traffic data set is traffic data corresponding to a geographical region. The traffic data may be, for example, the traffic data described herein in the first embodiment.

[0111] Step S214 involves selecting a representative traffic data set (a representative set of traffic data) for each cluster.

[0112] Step S216 includes performing a mobility analysis on the representative traffic data set to generate optimal parameters for periodic forecasting (for the area corresponding to the representative traffic data set).

[0113] 22 is a diagram showing an implementation example of the DT mobility classifier 10. The DT mobility classifier 10 can be considered to execute steps S212 and S214.

[0114] As mentioned above, the DT mobility classifier 10 analyses historical and live information (which can be collectively referred to as traffic data), classifies regions based on their mobility behavior patterns, and selects a number of representative regions / behavior patterns that represent the behavior of all regions. The mobility behavior patterns can be analyzed from different angles, i.e. based on different aspects including number of vehicles, average speed, max-min speed, congestion level, traffic jams (number), usage level of the road network, maximum capacity, etc. In some instances, number of vehicles and (average) speed are preferred (e.g. because these aspects can be used to calculate emissions).

[0115] The DT mobility classifier 10 receives as input historical traffic data and information (collectively referred to as traffic data) for the area managed by the DT. As shown in Figure 22, the DT mobility classifier 10 in a particular implementation includes three components: a TS similarity engine 12, a mobility classifier 14, and a representative behavior pattern module 16.

[0116] As output, the DT Mobility Classifier 10 may provide: (1) a set of categories / clusters into which domains (e.g. those managed by the associated DT solution) are classified, and (2) a set of representative profiles (representative behaviors / behavior patterns, or representative traffic datasets) representing each category / cluster.

[0117] TS Similarity Engine 12 The TS similarity engine 12 analyzes the historical mobility data (i.e., traffic data) of all regions and provides similarity features between all regions. The TS similarity engine 12 analyzes the traffic data and determines the average behavior of each region. The TS similarity engine 12 then performs a similarity search and provides a set of similarity features that are used by the mobility classifier 14. Similarity features are a set of metrics that determine how similar two time series are, i.e., the value of the difference (positive and negative) of the time series over time. Similarity features can include differences based on vector configurations or differences based on one or more patterns in the time series.

[0118] Figure 23 shows two techniques that can be used to generate similarity features: DTW looks for the closest value in a time range, while Euclidean looks for the similarity of each specific value in that time.

[0119] Mobility Classification Unit 14 The mobility classifier 14 uses as input information from all domains (e.g., managed by the DT) and historical mobility data (which may be referenced together with the traffic data), along with similarity features provided by the TS similarity engine 12. Additionally, the mobility classifier 14 may receive parameters defining the number of categories and / or an "acceptable" error. This parameter may take the form of a "closeness" threshold that defines how similar traffic datasets must be to be clustered into the same category.

[0120] For the generation (clustering) of the different categories, the mobility classifier 14 executes the following algorithm:

number

[0121] Here, dynamic time warping is used to compute the similarity, but other techniques can also be used, e.g., Euclidean distance, vector space similarity, shapelets similarity, bag of features / patterns, etc.

[0122] After execution of the algorithm, the mobility classifier 14 provides information about the different categories that have been generated and to which category each region belongs.

[0123] Representative behavior pattern module 16 The representative behavior pattern module 16 provides for each category the traffic data set that is closest to the centroid of that cluster / category (the centroid that corresponds to the "average behavior" of the cluster / category). The representative behavior pattern module 16 uses as input the historical traffic data (i.e. traffic data / traffic data set), the category information, and the information about which category each region belongs to.

[0124] FIG. 24 shows traffic data for multiple clusters. Where a cluster includes multiple regions, FIG. 24 shows a representative historical data set or "representative behavior pattern" (using a darker line) and other traffic data sets corresponding to other regions in the cluster (using a lighter line). As can be seen in FIG. 24, there are some categories that aggregate multiple behavior patterns or regions, and other categories that have few or only one of them. Such specific details will vary from example to example and may depend, at least in part, on the proximity threshold or other parameters that define the categories or the number of "acceptable" errors.

[0125] Finally, the representative behavior patterns of the clusters or the representative traffic dataset are provided to the DT mobility analyzer 320 and may also be provided to the prediction and anomaly detection modules 340, 360. Classification / clustering of traffic datasets of multiple regions and selection of a representative traffic dataset facilitates improved efficiency, especially when managing a large number of regions, and facilitates scalability, both while maintaining a high level of accuracy.

[0126] For each region that the DT mobility analyzer 320 analyzes, the DT mobility analyzer 320 uses a representative traffic data set for the clusters / categories of that region (rather than the actual traffic data of the region) to determine the optimal parameters. Otherwise, the DT mobility analyzer 320 may operate in the same manner as described with respect to the first implementation, and repeated description will be omitted.

[0127] The DT mobility prediction module 340 (specifically, the periodic prediction model module 44) uses the optimal parameters provided by the DT mobility analysis unit 320. Otherwise, the operation may be the same as described with respect to the first implementation. Figure 25 shows the DT mobility prediction module 340, in which the periodic prediction model module 44 receives a "representative profile" to indicate that the module 44 receives optimal parameters from the DT mobility analysis unit 320 based on a representative traffic data set. Apart from the optimal parameters, the periodic prediction model module 44 and the DT mobility prediction model 340 may generally use actual data for the area concerned (i.e., not a representative profile).

[0128] The second implementation further includes a method comprising the method described with respect to Fig. 21 and the method described with respect to Fig. 6 and / or Fig. 5. For example, the method described with respect to Fig. 21 may be implemented as method step S21 in the method described with respect to Fig. 6. In this case, the periodic traffic forecast may use optimal parameters determined based on a representative data set corresponding to the area under consideration, rather than the actual traffic data of the area itself. As mentioned above, apart from the optimal parameter determination, the forecast can generally use actual traffic data of the area concerned, i.e. not a representative traffic data set.

[0129] Similarly, aspects of a first implementation may include aspects of a second implementation or a portion thereof.

[0130] In some aspects, the second implementation may not include the DT anomaly detection unit 360 or method steps that may be performed by the DT anomaly detection unit 360 (as indicated by this module outlined in dashed lines in Figures 18-20).

[0131] In some aspects of the first or second implementations, generating a traffic forecast may be simplified compared to the traffic forecast generation described above, and adjustment of a forecast for a target area based on at least one other forecast for at least one other area may not be performed.

[0132] The traffic data in the first and second implementations is time series data.

[0133] The aspects disclosed herein in the first and second implementations improve the predicted and simulated scenarios for traffic mobility. Emission forecasts / estimations are provided. The aspects disclosed herein can also improve the predicted and simulated scenarios for human mobility. The geographical area including the geographical region may be an urban area, for example, a town or city.

[0134] Advantages of some aspects disclosed herein include the following: Scalable periodic forecasting: Maintain accuracy for forecasts and anomalies over large periodic ranges and small granularity (1 year / 5-15 minute granularity). · Context-aware prediction and anomaly detection at multiple levels, the key to simulation and digital rehearsal. · Detects abnormal behavior more accurately than other approaches by not relying on cycles or trends. High Demand Response. Near real-time response on demand for high frequency systems (2-5 seconds) such as DT solutions. Reduces incremental training time. Reduced complexity, reduced space. Minimize the size and complexity of models suitable for the digital twin environment. Suitability for large scale implementations: When prediction and / or anomaly detection models need to be scaled up, the time, size and computational resource needs can be minimized. Reduced Complexity: Minimize the number and complexity of models to fit into the digital twin environment while maintaining high accuracy.

[0135] Aspects disclosed herein may include any of the following: ·DT Mobility Forecasting: A mechanism for forecasting the near-future behavior of mobility, taking into account certain characteristics important for a DT solution, such as periodicity over large time horizons with small granularity, rapidity of forecasting, minimized models, mobility flow analysis, and context-aware forecasting at various levels, important for digital rehearsal, and complex pattern recognition (events, correlated regions, weather, etc.). · DT Mobility Anomaly Detection: A process to detect traffic anomalies from expected behavior and disruptions independent of periodic and trend aspects against predictions. DT Mobility Classification: A mechanism to increase the digital twin solution's ability to manage large-scale scenarios to predict pollutant emissions and potential behavioral anomalies in traffic. This component minimizes the computational requirements and models needed for scale-up while maintaining the accuracy and rapidity of response, which is important for real-time environments.

[0136] Fig. 26 is a block diagram of an information processing appliance 900 or computing device 900, such as a data storage server, that may be used to embody the present invention, to perform some or all of the operations of the methods embodying the present invention, and to perform some or all of the tasks of the apparatus of the embodiments. Fig. 26 and its description relate to / apply to the first and second embodiments.

[0137] The computing device 900 may be used to implement any of the above method steps, such as any of steps S20-S100, S22-S28, S101, S212-S216.

[0138] Computing device 900 includes a processor 993 and a memory 994. Optionally, the computing device also includes a network interface 997 for communicating with other such computing devices, such as other computing devices of embodiments of the present invention. Optionally, the computing device also includes one or more input mechanisms, such as a keyboard and mouse 996, and a display unit, such as one or more monitors 995. These elements may facilitate user interaction. The components may be connected to each other via a bus 992.

[0139] The memory 994 may include a computer-readable medium, which term may refer to a single medium or multiple media (e.g., centralized or distributed databases and / or associated caches and servers) configured to carry computer-executable instructions. Computer-executable instructions may include, for example, instructions and data that are accessible by a computer (e.g., one or more processors) and cause it to perform one or more functions or operations. For example, computer-executable instructions may include instructions for implementing the methods disclosed herein, or any method disclosed herein, such as steps S20-S100, S22-S28, S101, S212-S216. The term "computer-readable storage medium" may include any medium capable of storing, encoding, or retaining a set of instructions for execution by a machine and causing a machine to perform any one or more of the steps of the methods disclosed herein, and thus the term "computer-readable storage medium" includes, but is not limited to, solid-state memory, optical media, and magnetic media. For example, but not limited to, the term "computer-readable storage medium" may include solid-state memory, optical media, and magnetic media. By way of example, and not limitation, such computer-readable media may include non-transitory computer-readable media including random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), compact disk read-only memory (CD-ROM) or other optical disk storage devices, magnetic disk storage devices or other magnetic storage devices, flash memory devices (e.g., solid-state memory devices).

[0140] The processor 993 is configured to control the computing device and perform processing operations such as executing computer program code stored in the memory 994 to perform any of the method steps described herein. The memory 994 stores data that is read and written by the processor 993 and can store the traffic data and / or mobility information and / or mobility flow information and / or coefficients / parameters and / or predictions and / or forecast information and / or external event information / data and / or region information and / or correlation information and / or category information and / or thresholds and / or similarity information and / or formulas / algorithms and / or input data and / or other data and / or programs for performing any of the method steps described above. As referred to herein, a processor may include one or more general-purpose processing devices such as a microprocessor, a central processing unit, etc. The processor may include a processor implementing a Complex Instruction Set Computing (CISC) microprocessor, a Reduced Instruction Set Computing (RISC) microprocessor, a Very Long Instruction Word (VLIW) microprocessor, or other instruction set or processor implementing a combination of instruction sets. The processor may also include one or more special purpose processing devices, such as an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), a network processor, etc. In one or more embodiments, the processor is configured to execute the operations discussed herein and instructions to perform the operations. The processor 993 may be considered to include any of the modules described above. Any operation described as being performed by a module may be implemented as a method by a computer, such as the processor 993.

[0141] The display unit 995 may display traffic data and / or mobility information and / or mobility flow information and / or coefficients / parameters and / or predictions and / or forecast information and / or external event information / data and / or area information and / or correlation information and / or category information and / or thresholds and / or similarity information and / or formulas / algorithms and / or input data and / or other data and / or displays of data stored by the computing device, such as GUI windows and / or interactive displays that allow a user to interact with the device 900, for example by drag and drop or selection operations, and / or displays of any other output as described above, and may also display cursors and dialog boxes and screens that allow interaction between a user and programs and data stored on the computing device. The input mechanism 996 may enable a user to input data and instructions into the computing device, such as allowing a user to enter any user input as described above.

[0142] The network interface (network I / F) 997 can be connected to a network such as the Internet, and can be connected to other such computing devices via the network. The network I / F 997 can control input and output data to and from other devices via the network.

[0143] Other peripheral devices such as microphones, speakers, printers, power supply units, fans, cases, scanners, tracker balls, etc. may also be included in the computing device.

[0144] A method embodying the present invention may be performed on a computing device / appliance 900 as shown in FIG. 26. Such a computing device need not have all the components shown in FIG. 26, but may be comprised of a subset of these components. For example, the device 900 includes a processor 993 and a memory 994 coupled to the processor 993. Alternatively, the device 900 includes a processor 993, a memory 994 coupled to the processor 993, and a display 995. A method embodying the present invention may be performed by a single computing device in communication with one or more data storage servers over a network. The computing device may be a data storage device that stores at least a portion of the data.

[0145] Methods embodying the invention may be performed by a plurality of computing devices operating in cooperation with one another, one or more of which may be data storage servers that store at least a portion of the data.

[0146] The invention can be implemented in digital electronic circuitry, or in computer hardware, firmware, software, or in combinations of them. The invention can be implemented as a computer program or computer program product, i.e. a computer program embodied in a non-transitory information carrier, for example a machine-readable storage device or a propagated signal, for execution by or controlling the operation of one or more hardware modules.

[0147] A computer program may be in the form of a stand-alone program, computer program portions, or multiple computer programs, may be written in any type of programming language, including compiled or interpreted languages, and may be deployed in any form, such as as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a data processing environment. A computer program may be deployed to be executed on one module or on multiple modules that are located at one site or distributed across multiple sites and interconnected by a communication network.

[0148] The steps of the methods of the present invention may be performed by one or more programmable processors executing computer programs to perform the functions of the present invention by operating on input data and generating output. The apparatus of the present invention may be implemented as programmed hardware or as special purpose logic circuitry including Field Programmable Gate Arrays (FPGAs) or Application-Specific Integrated Circuits (ASICs), etc.

[0149] Processors suitable for the execution of a computer program include, by way of example, both general and special purpose microprocessors, and any one or more processors of any kind of digital computer. Typically, a processor receives instructions and data from a read-only memory or a random access memory or both. The essential elements of a computer are a processor for executing instructions coupled to one or more memories that store instructions and data.

[0150] The above-described embodiments of the invention may be advantageously used independently of any other embodiment of the embodiments or in any feasible combination with one or more other embodiments of the embodiments, for example features and / or aspects of the first / second implementation may be incorporated into other features and / or aspects of the first / second implementation.

[0151] The disclosure extends to the following statement (S): (S1) 1. A computer-implemented method comprising: Executing a traffic forecasting process using historical traffic data for a first time period to generate a first traffic forecast for the target geographic area (predicting traffic data for a future time period), the traffic forecasting process comprising: generating a periodic traffic forecast for the target geographic area (by estimating a periodic component of traffic data for the target geographic area and for the first time period) based on traffic data for the target geographic area and for the first time period; generating at least one other periodic traffic forecast for the at least one other geographical area based on traffic data for the at least one other geographical area and for the first time period (by estimating a periodic component of the traffic data for the at least one other geographical area and the traffic data for the first time period); analyzing mobility flows of the target geographical area and of the at least one other geographical area based on traffic data of the target geographical area and traffic data of the at least one other geographical area to determine at least one correlation between (traffic data trends of) the target geographical area and the at least one other geographical area; generating a (first) traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; and running the traffic forecasting process using (historical) traffic data for a second time period instead of the first time period to generate a second traffic forecast (predicting traffic data for a future time period) for the target geographic area, the first time period including the second time period and an additional time period newer than the second time period; decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component, and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component; comparing the first noise component of the first traffic forecast with the second noise component of the second traffic forecast to detect at least one anomaly, the comparison including comparing at least one deviation between the first noise component and the second noise component to an anomaly threshold (and classifying the deviation as an anomaly if the magnitude of the deviation is greater than the anomaly threshold); predicting emissions generated by traffic in the target geographic area based on the first traffic prediction, comprising, when at least one anomaly is detected, predicting an impact of the at least one detected anomaly on emissions; 4. A computer-implemented method comprising: (S2) the traffic data is in the form of a traffic data set each corresponding to a plurality of geographic regions, the plurality of geographic regions including the target geographic region and the at least one other geographic region; The computer-implemented method includes executing a parameter determination process, the parameter determination process comprising: clustering the traffic dataset based on similarity to one another to generate a plurality of clusters of the traffic dataset; For each cluster including multiple traffic datasets, selecting a traffic dataset of the cluster that is most similar to the average of the traffic datasets of the cluster as a representative traffic dataset; performing a mobility analysis based on a representative traffic data set corresponding to the target geographic area to determine optimal parameters for a periodic traffic forecast for the target geographic area, and performing a mobility analysis based on at least one representative traffic data set corresponding to the at least one other geographic area to determine optimal parameters for the at least one other periodic traffic forecast for the at least one other geographic area; Including, The computer-implemented method of claim 1, wherein the traffic forecasting process includes generating a periodic traffic forecast for the target geographic area using the determined optimal parameters for a periodic traffic forecast for the target geographic area, and generating at least one other periodic traffic forecast for the at least one other geographic area using optimal parameters for the at least one other periodic traffic forecast for the at least one other geographic area. (S3) The computer-implemented method of claim 1, wherein the traffic data for the first time period may be in the form of traffic data sets each corresponding to a plurality of geographic regions, the plurality of geographic regions including the target geographic region and the at least one other geographic region. performing a parameter determination process, the parameter determination process comprising: clustering the traffic data set for the first time period based on similarity to one another to generate a plurality of clusters of the traffic data set; For each cluster including multiple traffic datasets, selecting a traffic dataset of the cluster that is most similar to the average of the traffic datasets of the cluster as a representative traffic dataset; performing a mobility analysis based on a representative traffic data set corresponding to the target geographic area to determine optimal parameters for a periodic traffic forecast for the target geographic area, and performing a mobility analysis based on at least one representative traffic data set corresponding to the at least one other geographic area to determine optimal parameters for the at least one other periodic traffic forecast for the at least one other geographic area; Including, The computer-implemented method of statement S1, wherein the traffic forecasting process includes generating a periodic traffic forecast for the target geographic area using the determined parameters for a periodic traffic forecast for the target geographic area, and generating at least one other periodic traffic forecast for the at least one other geographic area using optimal parameters for the at least one other periodic traffic forecast for the at least one other geographic area. (S5) The computer-implemented method of claim 2 or S4, wherein generating the second traffic forecast by performing the traffic forecast process using traffic data for the second time period (instead of traffic data for the first time period) includes using determined parameters for a periodic traffic forecast for the target geographic area and determined parameters for a periodic traffic forecast for the at least one other geographic area. (S6) the parameter is a first parameter, and the computer-implemented method further comprises: determining second parameters for a periodic traffic forecast for the target geographic area and second parameters for a periodic traffic forecast for the at least one other geographic area by performing the parameter determination process using traffic data for the second time period (instead of traffic data for the first time period); The computer-implemented method of statement S2 or S4, wherein generating the second traffic forecast by performing the traffic forecasting process using traffic data for the second time period (instead of traffic data for the first time period) includes using the determined second parameters. (S7) The computer-implemented method of any of statements S1-6, wherein the traffic data includes vehicular traffic data. (S8) The computer-implemented method of any one of statements S1-7, wherein the traffic data includes data obtained from sensors in (a geographical area including) the geographical area. (S9) The computer-implemented method of any of statements S1-8, wherein the traffic data includes data obtained from a digital twin of the associated geographic area. (S10) The computer-implemented method of any one of statements S1-9, wherein the traffic data includes data obtained from a digital twin of a geographic area that includes the associated geographic area. (S11) The computer-implemented method of statements S1-10, wherein the traffic data includes data obtained from sensors in the relevant geographic area (including the geographic area) and data obtained from a digital twin of the relevant geographic area (including the geographic area). (S12) The computer-implemented method of statement S9 or S10 or S11, wherein data obtained from the digital twin is based on data obtained from sensors in the associated geographic region / area. (S13) The computer-implemented method of any of statements S9-12, wherein the digital twin is a model / simulation of the geographic region / area. (S14) The computer-implemented method of any of statements S9-13, wherein the digital twin is a model / simulation of (a geographic area including) the geographic area, with some modifications compared to an actual geographic area / region (to model a scenario). (S15) The computer-implemented method of any of statements S9-14, wherein the data obtained from the digital twin based on data obtained from the sensors includes data that has undergone at least one pre-processing stage, optionally including generation of intermediate variables and / or values. (S16) The computer-implemented method of any of statements S1-15, wherein the at least one other geographical area includes a plurality of geographical areas (and the traffic forecasting process may include generating a plurality of other periodic traffic forecasts for the plurality of other geographical areas) (and the parameter determination process may include determining (optimal) parameters for each of the plurality of other periodic traffic forecasts for the plurality of other geographical areas). (S17) 17. The computer-implemented method of any of statements 1-16, wherein the target geographic area and the at least one other geographic area are each areas of an urban area or a town or a city. (S18) The computer-implemented method of any of statements S1-17, wherein the geographic region includes a metropolitan area or a town or a city. (S19) traffic data for said geographical area (first and / or second traffic data and / or traffic data for said geographical area and / or traffic data set for said geographical area), the number of vehicles in said geographic area (over time); the number of each of a plurality of types of vehicles in said geographic area (over time); the location of each of a plurality of vehicles within said geographic region (over time); the heading of each of a plurality of vehicles within the geographic region (over time); the speed of each of a plurality of vehicles within said geographic region (over time); an average speed of each of a plurality of vehicles in said geographic region (over time); the minimum and / or maximum speed of each of a plurality of vehicles in said geographic region (over time); the level of congestion within said geographic area (over time); the number of traffic jams in said geographic area (over time); the level of use of the road network within said geographical area (over time); the maximum transport capacity (of the road network) within said geographical area (over time); an identifier for each of a plurality of vehicles within said geographic region (over time); 19. The computer-implemented method of any one of statements 1 to 18, comprising: (S20) The computer-implemented method of any of statements S1-19, wherein the traffic data for the geographical region (the first and / or second traffic data and / or the traffic data set for the geographical region) includes: a number of vehicles in the geographical region (over time) or a number of each of a plurality of types of vehicles in the geographical region (over time); and an average speed of each of a plurality of vehicles in the geographical region (over time). (S21) The traffic forecast includes: the number of vehicles in said geographic area (over time); the number of each of a plurality of types of vehicles in said geographic area (over time); the location of each of a plurality of vehicles within said geographic region (over time); the heading of each of a plurality of vehicles within the geographic region (over time); the speed of each of a plurality of vehicles within said geographic region (over time); an average speed of each of a plurality of vehicles in said geographic region (over time); the minimum and / or maximum speed of each of a plurality of vehicles in said geographic region (over time); the level of congestion within said geographic area (over time); the number of traffic jams in said geographic area (over time); the level of use of the road network within said geographical area (over time); the maximum transport capacity (of the road network) within said geographical area (over time); an identifier for each of a plurality of vehicles within said geographic region (over time); 21. The computer-implemented method of any of statements 1-20, comprising predictive traffic data including: (S22) The computer-implemented method of any of statements S1-21, wherein the traffic forecast includes predicted traffic data including: a number of vehicles (over time) in the geographic region or a number of each of a plurality of types of vehicles in the geographic region (over time); and an average speed of each of a plurality of vehicles in the geographic region (over time). (S23) The computer-implemented method of any of statements S8-S22, wherein the sensors include any of the following: at least one vehicle-mounted sensor, at least one user device, at least one camera, and at least one speed sensor. (S24) The computer-implemented method / traffic forecasting process comprises: performing a mobility analysis based on a traffic dataset for the target geographical area to determine (optimal) parameters for a periodic traffic forecast for the target geographical area and performing a mobility analysis based on a traffic dataset for the at least one other geographical area to determine (optimal) parameters for the at least one other periodic traffic forecast for the at least one other geographical area; and / or The computer-implemented method of any of statements S1-23, wherein the traffic forecasting process includes steps of generating a periodic traffic forecast for the target geographical area using the determined parameters for a periodic traffic forecast for the target geographical area, and generating at least one other periodic traffic forecast for the at least one other geographical area using optimal parameters for the at least one other periodic traffic forecast for the at least one other geographical area. (S25) The computer-implemented method of statement S24, wherein performing the mobility analysis includes identifying trends in the associated traffic data. (S26) The step of performing a mobility analysis includes obtaining the level of stationarity of the annual trends (determining the degree of periodic differences (using the Canova-Hansen method)); and / or Obtaining relationships between data from different time periods (using either the Kwiatkowski-Phillips-Schmidt-Shin method, the extended Dicky-Fuller method, or the Phillips-Perron method); and / or Obtaining a range of values ​​around which the mobility trend oscillates (using the autocorrelation function, partial autocorrelation function and / or extended autocorrelation function methods); The computer-implemented method of claim 24 or S25, comprising: (S27) The step of performing a mobility analysis includes obtaining the level of stationarity of the annual trends and determining the stationarity parameters (determining the degree of periodic differences (using the Canova-Hansen method)); and / or Obtaining the relationship between data from different time periods and determining regression parameters (using either the Kwiatkowski-Phillips-Schmidt-Shin method, the extended Dicky-Fuller method, or the Phillips-Perron method); and / or Obtaining the range of values ​​around which the mobility trend oscillates and determining prediction parameters (using the autocorrelation function, partial autocorrelation function and / or extended autocorrelation function methods); The computer-implemented method of claim 24, S25, or S26, comprising: (S28) The computer-implemented method of any of statements S1-27, wherein generating a periodic traffic forecast for the target geographic region (or for the geographic region) comprises generating regressors based on associated traffic data that define periodic aspects of the data (optionally using a Fourier transform or a sinusoidal decomposition, where the regressors are optionally generated as Fourier factors). (S29) The computer-implemented method of statement S28, wherein generating a periodic traffic forecast for the target geographic region (or the geographic region) includes generating a periodic traffic forecast based on associated regressors, associated traffic data, and associated parameters. (S30) 30. The computer-implemented method of any of statements 1-29, further comprising: analyzing mobility flows of the target geographic area and a plurality of other geographic areas to determine the at least one other geographic area relevant to a traffic forecast for the target geographic area. (S31) the traffic data includes information regarding events occurring in the relevant geographic region, and the computer-implemented method / traffic prediction process further comprises: - analyzing traffic data of said target area to obtain at least one correlation between the traffic data and at least one (external) event; Including, generating a (first) traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined correlation, The computer-implemented method of any of statements 1-30, further comprising adjusting a periodic forecast for the target geographic area based on at least one correlation between traffic data and at least one external event and based on the predicted at least one event. (S32) The computer-implemented method of statement S31, wherein the at least one event includes any of a weather event, a traffic event, an accident event, and an entertainment event. (S33) 33. The computer-implemented method of any of statements 1-32, wherein decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component includes using a Periodic Trend Decomposition with LOESS (STL) technique. (S34) 34. The computer-implemented method of any of statements 1-33, wherein comparing at least one deviation between the first noise component and the second noise component to an abnormality threshold comprises comparing a percentage deviation to the abnormality threshold. (S35) The computer-implemented method of statement S34, wherein the percentage deviation is a deviation as a percentage of the value of the first or second noise component (at the relevant time point). (S36) Decomposing the first and second traffic forecasts into components may include decomposing at least one aspect of the first and second traffic forecasts, the at least one aspect comprising: the number of vehicles in said geographic area (over time); the number of each of a plurality of types of vehicles in said geographic area (over time); the location of each of a plurality of vehicles within said geographic region (over time); the heading of each of a plurality of vehicles within the geographic region (over time); the speed of each of a plurality of vehicles within said geographic region (over time); an average speed of each of a plurality of vehicles in said geographic region (over time); the minimum and / or maximum speed of each of a plurality of vehicles in said geographic region (over time); the level of congestion within said geographic area (over time); the number of traffic jams in said geographic area (over time); the level of use of the road network within said geographical area (over time); the maximum transport capacity (of the road network) within said geographical area (over time); an identifier for each of a plurality of vehicles within said geographic region (over time); The computer-implemented method of any one of statements 1 to 35, comprising: (S37) Decomposing the first and second traffic forecasts into components may include decomposing at least one aspect of the first and second traffic forecasts, the at least one aspect comprising: The computer-implemented method of any of statements S1-36, including any of: a number of vehicles in the geographic region (over time) or a number of each of a plurality of types of vehicles in the geographic region (over time); and an average speed of each of a plurality of vehicles in the geographic region (over time). (S38) The step of estimating emissions comprises: information indicative of average emissions of the vehicle based on said speed; the number of vehicles from said traffic forecast; an average speed of vehicles from said traffic forecast; The computer-implemented method of any of statements 1-37, comprising calculating emissions based on: (S39) A computer-implemented method comprising executing the computer-implemented method set forth in any of statements 1-38 multiple times, with a different geographic area of ​​the geographic region as a target area each time. (S40) The computer-implemented method of statement S39 includes aggregating the forecasted emissions (and forecasted impacts of detected anomalies) for the multiple regions to obtain a forecast of emissions occurring in the geographic region (including the impacts of detected anomalies). (S41) Clustering the traffic data sets based on their mutual similarity includes clustering the traffic data sets based on at least one aspect of the traffic data, the at least one aspect being one of: the number of vehicles in said geographic area (over time); the number of each of a plurality of types of vehicles in said geographic area (over time); the location of each of a plurality of vehicles within said geographic region (over time); the heading of each of a plurality of vehicles within the geographic region (over time); the speed of each of a plurality of vehicles within said geographic region (over time); an average speed of each of a plurality of vehicles in said geographic region (over time); the minimum and / or maximum speed of each of a plurality of vehicles in said geographic region (over time); the level of congestion within said geographic area (over time); the number of traffic jams in said geographic area (over time); the level of use of the road network within said geographical area (over time); the maximum transport capacity (of the road network) within said geographical area (over time); The computer-implemented method of any one of statements S2 or S4 to S40, comprising: (S42) Clustering the traffic data sets based on their mutual similarity includes clustering the traffic data sets based on at least one aspect of the traffic data, the at least one aspect being one of: The computer-implemented method of any of statements S2 or S4-40, including any of: a number of vehicles in the geographic region (over time) or a number of each of a plurality of types of vehicles in the geographic region (over time); and an average speed of each of a plurality of vehicles in the geographic region (over time). (S43) The parameter determination process comprises, prior to the clustering, obtaining a behavioral representation for each traffic data set, the behavioral representation being based on at least one aspect of the data, the at least one aspect being one of: the number of vehicles in said geographic area (over time); the number of each of a plurality of types of vehicles in said geographic area (over time); the location of each of a plurality of vehicles within said geographic region (over time); the heading of each of a plurality of vehicles within the geographic region (over time); the speed of each of a plurality of vehicles within said geographic region (over time); an average speed of each of a plurality of vehicles in said geographic region (over time); the minimum and / or maximum speed of each of a plurality of vehicles in said geographic region (over time); the level of congestion within said geographic area (over time); the number of traffic jams in said geographic area (over time); the level of use of the road network within said geographical area (over time); the maximum transport capacity (of the road network) within said geographical area (over time); The computer-implemented method of any one of statements S2 or S4 to S42, comprising: (S44) The parameter determination process comprises, prior to the clustering, obtaining a behavioral representation for each traffic data set, the behavioral representation being based on at least one aspect of the data, the at least one aspect being one of: The computer-implemented method of any of statements S2 or S4-43, including either the number of vehicles in the geographic region (over time), or the number of each of a plurality of types of vehicles in the geographic region (over time), and the average speed of each of a plurality of vehicles in the geographic region (over time). (S45) The computer-implemented method of S43 or S44, wherein clustering the traffic data sets based on similarity to each other comprises clustering the traffic data sets based on similarity of corresponding behavior representations to each other. (S46) The computer-implemented method of any of S2 or S4-S45, wherein the step of clustering the traffic datasets based on their similarity to each other comprises using dynamic time warp matching or Euclidean matching between the traffic datasets or between the behavior representations. (S47) 1. A computer-implemented method comprising: generating a first traffic forecast for the target geographical area based on the (historical) traffic data for a first time period and the (historical) traffic data for the target geographical area (predicting traffic data for a future time period); generating a second traffic forecast (predicting traffic data for a future time period) for the target geographic area based on the (past) traffic data for a second time period and the (past) traffic data for the target geographic area, the first time period including the second time period and an additional time period more recent than the second time period; decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component, and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component; comparing the first noise component of the first traffic forecast with a second noise component of the second traffic forecast to detect at least one anomaly, the comparison including comparing at least one deviation between the first noise component and the second noise component with an anomaly threshold (and classifying the deviation as an anomaly if the magnitude of the deviation is greater than the anomaly threshold); forecasting emissions generated by traffic in the target geographical area based on the first traffic forecast, including, if at least one anomaly is detected, forecasting an impact of the at least one detected anomaly on emissions; 23. A computer-implemented method comprising: (S48) the traffic data for the first time period includes traffic data for at least one other region of the geographic region (over the first time period), and the traffic data for the second time period includes traffic data for the at least one other region of the geographic region (over the second time period), and generating the first traffic forecast includes performing a traffic forecasting process using the traffic data for the first time period, the traffic forecasting process including: generating a periodic traffic forecast for the target geographic region based on the traffic data for the target geographic region (by estimating a periodic component of the traffic data for the target geographic region) (predicting traffic data for a future time period); generating a periodic traffic forecast for the target geographical area (by estimating a periodic component of traffic data for the target geographical area) based on traffic data for the geographical area (predicting traffic data for a future period of time); generating at least one other periodic traffic forecast for the at least one other area based on traffic data of the at least one other area of ​​the geographical region (by estimating a periodic component of the traffic data of the at least one other area) (predicting traffic data for a future period); Based on traffic data of the target geographical area and traffic data of the at least one other geographical area, analyzing mobility flows of the target geographical area and the at least one other geographical area to determine at least one correlation between (traffic data trends of) the target geographical area and the at least one other geographical area; generating a (first) traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; Including, The computer-implemented method of statement S47, wherein generating the second traffic forecast includes performing the traffic forecasting process using traffic data for the second time period (instead of traffic data for the first time period). (S49) 1. A computer-implemented method comprising: performing a traffic prediction process using the (historical) traffic data, said traffic prediction process comprising: generating a periodic traffic forecast for a target geographical area (by estimating a periodic component of traffic data for the target geographical area) based on traffic data for the target geographical area (for a geographical region); generating at least one other periodic traffic forecast for the at least one other geographical area based on traffic data of the at least one other geographical area (by estimating a periodic component of the traffic data of the at least one other geographical area) (predicting traffic data for a future period); analyzing mobility flows of the target geographical area and the at least one other geographical area based on traffic data of the target geographical area and traffic data of the at least one other geographical area to determine at least one correlation between (traffic data trends of) the target geographical area and the at least one other geographical area; generating a multi-modal traffic forecast (or final, cumulative, or overall forecast) for the target geographic region by adjusting the periodic forecast for the target geographic region based on the determined at least one correlation with the at least one other periodic forecast; forecasting emissions generated by traffic in the target geographic area based on the final traffic forecast; 4. A computer-implemented method comprising: (S50) the traffic data is traffic data for a first time period, the final traffic forecast is a first traffic forecast, and the computer-implemented method further comprises: generating a second traffic forecast by running the traffic forecasting process using traffic data for a second time period (instead of traffic data for the first time period), the first time period including the second time period and an additional time period more recent than the second time period; The computer-implemented method comprises: decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component, and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component; comparing the first noise component of the first traffic prediction with a second noise component of the second traffic prediction to detect at least one anomaly, the step including comparing at least one deviation between the first traffic prediction and the second traffic prediction with an anomaly threshold (and classifying the deviation as an anomaly if a magnitude of the deviation is greater than the anomaly threshold); The computer-implemented method of statement S49, wherein predicting emissions generated by traffic in the target geographic area may include, if at least one anomaly is detected, predicting an impact of the at least one detected anomaly on emissions. (S51) A computer program which, when executed on a computer, causes the computer to carry out a method, the method comprising: Executing a traffic forecasting process using historical traffic data for a first time period to generate a first traffic forecast for the target geographic area (predicting traffic data for a future time period), the traffic forecasting process comprising: generating a periodic traffic forecast for the target geographic area (by estimating a periodic component of traffic data for the target geographic area and for the first time period) based on traffic data for the target geographic area and for the first time period; generating at least one other periodic traffic forecast for the at least one other geographical area based on traffic data for the at least one other geographical area and for the first time period (by estimating a periodic component of the traffic data for the at least one other geographical area and the traffic data for the first time period); analyzing mobility flows of the target geographical area and of the at least one other geographical area based on traffic data of the target geographical area and traffic data of the at least one other geographical area to determine at least one correlation between (traffic data trends of) the target geographical area and the at least one other geographical area; generating a (first) traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; and running the traffic forecasting process using (historical) traffic data for a second time period instead of the first time period to generate a second traffic forecast (predicting traffic data for a future time period) for the target geographic area, the first time period including the second time period and an additional time period newer than the second time period; decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component, and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component; comparing the first noise component of the first traffic forecast with the second noise component of the second traffic forecast to detect at least one anomaly, the comparison including comparing at least one deviation between the first noise component and the second noise component to an anomaly threshold (and classifying the deviation as an anomaly if the magnitude of the deviation is greater than the anomaly threshold); predicting emissions generated by traffic in the target geographic area based on the first traffic prediction, comprising, when at least one anomaly is detected, predicting an impact of the at least one detected anomaly on emissions; A computer program comprising: (S52) An information processing device comprising: a memory; and a processor coupled to the memory, the processor comprising: Executing a traffic forecasting process using historical traffic data for a first time period to generate a first traffic forecast for the target geographic area (predicting traffic data for a future time period), the traffic forecasting process comprising: generating a periodic traffic forecast for the target geographic area (by estimating a periodic component of traffic data for the target geographic area and for the first time period) based on traffic data for the target geographic area and for the first time period; generating at least one other periodic traffic forecast for the at least one other geographical area based on traffic data for the at least one other geographical area and for the first time period (by estimating a periodic component of the traffic data for the at least one other geographical area and the traffic data for the first time period); analyzing mobility flows of the target geographical area and of the at least one other geographical area based on traffic data of the target geographical area and traffic data of the at least one other geographical area to determine at least one correlation between (traffic data trends of) the target geographical area and the at least one other geographical area; generating a (first) traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; and running the traffic forecasting process using (historical) traffic data for a second time period instead of the first time period to generate a second traffic forecast (predicting traffic data for a future time period) for the target geographic area, the first time period including the second time period and an additional time period newer than the second time period; decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component, and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component; comparing the first noise component of the first traffic forecast with the second noise component of the second traffic forecast to detect at least one anomaly, the comparison including comparing at least one deviation between the first noise component and the second noise component to an anomaly threshold (and classifying the deviation as an anomaly if the magnitude of the deviation is greater than the anomaly threshold); predicting emissions generated by traffic in the target geographic area based on the first traffic prediction, comprising, when at least one anomaly is detected, predicting an impact of the at least one detected anomaly on emissions; An information processing device configured to execute the above steps.

[0152] The disclosure extends to statement (B) below. (B1) 1. A computer-implemented method comprising: executing a parameter determination process based on a plurality of (historical) traffic data sets for a plurality of geographical areas, said parameter determination process comprising: clustering the traffic datasets based on their similarity to each other to generate a plurality of clusters of traffic datasets, and for each cluster including a plurality of traffic datasets, selecting a traffic dataset of the cluster that is most similar to an average of the traffic datasets of the cluster as a representative traffic dataset; Based on a representative traffic data set corresponding to a target geographical area, performing a mobility analysis to determine (optimal) parameters for a periodic traffic forecast for said target geographical area, and based on at least one representative traffic data set corresponding to at least one other geographical area, performing a mobility analysis to determine (optimal) parameters for at least one other periodic traffic forecast for at least one other geographical area; and performing a traffic prediction process for the target geographic region using a traffic dataset for the target geographic region and at least one traffic dataset for at least one other geographic region, the traffic prediction process comprising: generating the periodic traffic forecast for the target geographical area (by estimating the periodic component of traffic data for the target geographical area) based on the traffic dataset for the target geographical area and using the determined parameters for periodic traffic forecast for the target geographical area (predicting traffic data for a future time period); generating a periodic traffic forecast for the at least one other geographical area (by estimating a periodic component of traffic data for the at least one other geographical area) based on the at least one traffic data set for the at least one other geographical area and using the determined parameters for the at least one other periodic traffic forecast for the at least one other geographical area (predicting traffic data for a future period); analyzing mobility flows in the geographical area based on a traffic dataset of the geographical area and at least one traffic dataset of the at least one other geographical area to determine at least one correlation between the target geographical area and (traffic data trends of) the at least one other geographical area; generating a combined traffic forecast (or final forecast or cumulative forecast or overall forecast) by adjusting the periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; and forecasting emissions generated by traffic in the target geographic area based on the final traffic forecast; The computer-implemented method of claim 2, comprising: (B2) The traffic dataset for the target geographic region and at least one traffic dataset for the at least one other geographic region are traffic data for a first time period, and the final traffic forecast is a first traffic forecast, and the computer-implemented method further comprises: generating a second traffic forecast by running the traffic forecasting process using traffic data for a second time period (instead of traffic data for the first time period), the first time period including the second time period and an additional time period more recent than the second time period; The computer-implemented method comprises: decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component, and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component; comparing the first noise component of the first traffic prediction with a second noise component of the second traffic prediction to detect at least one anomaly, the step including comparing at least one deviation between the first traffic prediction and the second traffic prediction with an anomaly threshold (and classifying the deviation as an anomaly if a magnitude of the deviation is greater than the anomaly threshold); The computer-implemented method of statement B1, wherein the step of predicting emissions generated by traffic in the target geographic area includes, if at least one anomaly is detected, predicting an impact of the at least one detected anomaly on emissions. (B3) The computer-implemented method of statement B2, wherein generating the second traffic forecast by performing the traffic forecast process using traffic data for the second time period (instead of traffic data for the first time period) includes using determined parameters for the periodic traffic forecast for the target geographic area and determined parameters for at least one other periodic traffic forecast for the at least one other geographic area. (B4) the parameter is a first parameter, and the computer-implemented method further comprises: determining second parameters for a periodic traffic forecast for the target geographic area and second parameters for a periodic traffic forecast for the at least one other geographic area by performing the parameter determination process using traffic data for the second time period (instead of traffic data for the first time period); The computer-implemented method of any of statements B2-B4, wherein generating the second traffic forecast by performing the traffic forecasting process using traffic data for the second time period (instead of traffic data for the first time period) includes using the determined second parameters. (B5) Clustering the traffic data sets based on their mutual similarity includes clustering the traffic data sets based on at least one aspect of the traffic data, the at least one aspect being one of: the number of vehicles in said geographic area (over time); the number of each of a plurality of types of vehicles in said geographic area (over time); the location of each of a plurality of vehicles within said geographic region (over time); the heading of each of a plurality of vehicles within the geographic region (over time); the speed of each of a plurality of vehicles within said geographic region (over time); an average speed of each of a plurality of vehicles in said geographic region (over time); the minimum and / or maximum speed of each of a plurality of vehicles in said geographic region (over time); the level of congestion within said geographic area (over time); the number of traffic jams in said geographic area (over time); the level of use of the road network within said geographical area (over time); the maximum transport capacity (of the road network) within said geographical area (over time); The computer-implemented method of any one of statements B1 to B4, comprising: (B6) Clustering the traffic data sets based on their mutual similarity includes clustering the traffic data sets based on at least one aspect of the traffic data, the at least one aspect being one of: The computer-implemented method of any of statements B1-B5, including any of: a number of vehicles in the geographic area (over time) or a number of each of a plurality of types of vehicles in the geographic area (over time); and an average speed of each of a plurality of vehicles in the geographic area (over time). (B7) The parameter determination process comprises, prior to the clustering, obtaining a behavioral representation for each traffic data set, the behavioral representation being based on at least one aspect of the data, the at least one aspect being one of: the number of vehicles in said geographic area (over time); the number of each of a plurality of types of vehicles in said geographic area (over time); the location of each of a plurality of vehicles within said geographic region (over time); the heading of each of a plurality of vehicles within the geographic region (over time); the speed of each of a plurality of vehicles within said geographic region (over time); an average speed of each of a plurality of vehicles in said geographic region (over time); the minimum and / or maximum speed of each of a plurality of vehicles in said geographic region (over time); the level of congestion within said geographic area (over time); the number of traffic jams in said geographic area (over time); the level of use of the road network within said geographical area (over time); the maximum transport capacity (of the road network) within said geographical area (over time); The computer-implemented method of any one of statements B1 to B6, comprising: (B8) The parameter determination process comprises, prior to the clustering, obtaining a behavioral representation for each traffic data set, the behavioral representation being based on at least one aspect of the data, the at least one aspect being one of: The computer-implemented method of any of statements B1-B7, including any of the number of vehicles in the geographical area (over time), or the number of each of a plurality of types of vehicles in the geographical area (over time), and the average speed of each of a plurality of vehicles in the geographical area (over time). (B9) The computer-implemented method of any one of B1 to B8, wherein clustering the traffic data sets based on similarity to one another includes clustering the traffic data sets based on similarity of corresponding behavior representations to one another. (B10) The computer-implemented method of any of B1 to B9, wherein the step of clustering the traffic datasets based on their similarity to each other includes using dynamic time warp matching or Euclidean matching between the traffic datasets or between the behavior representations. (B11) The computer-implemented method of any of statements B1-B10, wherein the traffic data includes vehicular traffic data. (B12) The computer-implemented method of any one of statements B1 to B11, wherein the traffic data includes data obtained from sensors in (a geographical area including) the geographical area. (B13) The computer-implemented method of any of statements B1-B12, wherein the traffic data includes data obtained from a digital twin of the associated geographic area. (B14) The computer-implemented method of any one of statements B1 to B13, wherein the traffic data includes data obtained from a digital twin of a geographic area that includes the associated geographic area. (B15) The computer-implemented method of any one of statements B1 to B14, wherein the traffic data includes data obtained from sensors in the relevant geographic area (including the geographic area) and data obtained from a digital twin of the relevant geographic area (including the geographic area). (B16) The computer-implemented method of any one of statements B13 to B15, wherein the data obtained from the digital twin is based on data obtained from sensors in the associated geographic region / area. (B17) The computer-implemented method of any of statements B13-B16, wherein the digital twin is a model / simulation of the geographic region / area. (B18) The computer-implemented method of any of statements B13-B17, wherein the digital twin is a model / simulation of (a geographic area including) the geographic area, with some modifications compared to an actual geographic area / region (to model a scenario). (B19) The computer-implemented method of any of statements B13-B18, wherein the data obtained from the digital twin based on data obtained from the sensors includes data that has undergone at least one pre-processing stage, optionally including generation of intermediate variables and / or values. (B20) The computer-implemented method of any of statements B1-B19, wherein the at least one other geographic area includes a plurality of geographic areas (and the traffic forecasting process includes generating a plurality of other periodic traffic forecasts for the plurality of other geographic areas) (and the parameter determination process includes determining (optimal) parameters for each of the plurality of other periodic traffic forecasts for the plurality of other geographic areas). (B21) The computer-implemented method of any of statements B1-B20, wherein the target geographic area and the at least one other geographic area are each areas of an urban area or a town or a city. (B22) The computer-implemented method of any of statements B1-B21, wherein the geographic region includes a metropolitan area or a town or a city. (B23) traffic data for said geographical area (first and / or second traffic data and / or traffic data for said geographical area and / or traffic data set for said geographical area), the number of vehicles in said geographic area (over time); the number of each of a plurality of types of vehicles in said geographic area (over time); the location of each of a plurality of vehicles within said geographic region (over time); the heading of each of a plurality of vehicles within the geographic region (over time); the speed of each of a plurality of vehicles within said geographic region (over time); an average speed of each of a plurality of vehicles in said geographic region (over time); the minimum and / or maximum speed of each of a plurality of vehicles in said geographic region (over time); the level of congestion within said geographic area (over time); the number of traffic jams in said geographic area (over time); the level of use of the road network within said geographical area (over time); the maximum transport capacity (of the road network) within said geographical area (over time); an identifier for each of a plurality of vehicles within said geographic region (over time); The computer-implemented method of any of statements B1 to B22, comprising: (B24) The computer-implemented method of any of statements B1-B23, wherein the traffic data for the geographical region (the first and / or second traffic data and / or the traffic data set for the geographical region) includes a number of vehicles in the geographical region (over time) or a number of each of a plurality of types of vehicles in the geographical region (over time), and an average speed of each of a plurality of vehicles in the geographical region (over time). (B25) The traffic forecast includes: the number of vehicles in said geographic area (over time); the number of each of a plurality of types of vehicles in said geographic area (over time); the location of each of a plurality of vehicles within said geographic region (over time); the heading of each of a plurality of vehicles within the geographic region (over time); the speed of each of a plurality of vehicles within said geographic region (over time); an average speed of each of a plurality of vehicles in said geographic region (over time); the minimum and / or maximum speed of each of a plurality of vehicles in said geographic region (over time); the level of congestion within said geographic area (over time); the number of traffic jams in said geographic area (over time); the level of use of the road network within said geographical area (over time); the maximum transport capacity (of the road network) within said geographical area (over time); an identifier for each of a plurality of vehicles within said geographic region (over time); The computer-implemented method of any of statements B1-B24, comprising predictive traffic data including: (B26) The computer-implemented method of any of statements B1-B25, wherein the traffic forecast includes predicted traffic data including: a number of vehicles in the geographic region (over time) or a number of each of a plurality of types of vehicles in the geographic region (over time); and an average speed of each of a plurality of vehicles in the geographic region (over time). (B27) The computer-implemented method of any of statements B12-B26, wherein the sensors include any of the following: at least one vehicle-mounted sensor, at least one user device, at least one camera, and at least one speed sensor. (B28) performing a mobility analysis based on a traffic dataset for the target geographical area to determine (optimal) parameters for a periodic traffic forecast for the target geographical area and performing a mobility analysis based on a traffic dataset for the at least one other geographical area to determine (optimal) parameters for the at least one other periodic traffic forecast for the at least one other geographical area; Including, The computer-implemented method of any of statements B1-B27, wherein the traffic forecasting process includes steps of generating a periodic traffic forecast for the target geographical area using the determined parameters for a periodic traffic forecast for the target geographical area, and generating at least one other periodic traffic forecast for the at least one other geographical area using optimal parameters for the at least one other periodic traffic forecast for the at least one other geographical area. (B29) The computer-implemented method of statement B28, wherein performing the mobility analysis includes identifying trends in the associated traffic data. (B30) The step of performing a mobility analysis includes obtaining the level of stationarity of the annual trends (determining the degree of periodicity difference (using the Canova-Hansen method)); and / or Obtaining relationships between data from different time periods (using either the Kwiatkowski-Phillips-Schmidt-Shin method, the extended Dicky-Fuller method, or the Phillips-Perron method); and / or Obtaining a range of values ​​around which the mobility trend oscillates (using the autocorrelation function, partial autocorrelation function and / or extended autocorrelation function methods); The computer-implemented method of any one of statements B28 to B29, comprising: (B31) The step of performing a mobility analysis includes obtaining the level of stationarity of the annual trends and determining the stationarity parameters (determining the degree of periodic differences (using the Canova-Hansen method)); and / or Obtaining the relationship between data from different time periods and determining regression parameters (using either the Kwiatkowski-Phillips-Schmidt-Shin method, the extended Dicky-Fuller method, or the Phillips-Perron method); and / or Obtaining the range of values ​​around which the mobility trend oscillates and determining prediction parameters (using the autocorrelation function, partial autocorrelation function and / or extended autocorrelation function methods); The computer-implemented method of claim 28 or B29 or B30, comprising: (B32) The computer-implemented method of any of statements B1-B31, wherein generating a periodic traffic forecast for the target geographic region (or for the geographic region) comprises generating regressors based on associated traffic data that define periodic aspects of the data (optionally using a Fourier transform or a sinusoidal decomposition, where the regressors are optionally generated as Fourier factors). (B33) The computer-implemented method of any of statements B1-B32, wherein generating a periodic traffic forecast for the target geographic region (or the geographic region) includes generating a periodic traffic forecast based on associated regressors, associated traffic data, and associated parameters. (B34) The computer-implemented method of any of statements B1-B33, wherein the computer-implemented method / traffic prediction process includes analyzing mobility flows of the target geographical area and a plurality of other geographical areas to determine the at least one other geographical area that is relevant for traffic prediction of the target geographical area. (B35) the traffic data includes information regarding events occurring in the relevant geographic region, and the computer-implemented method / traffic prediction process further comprises: - analyzing traffic data of said target area to obtain at least one correlation between the traffic data and at least one (external) event; Including, generating a (first) traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined correlation, The computer-implemented method of any of statements B1-B34, further comprising adjusting a periodic forecast for the target geographic area based on at least one correlation between traffic data and at least one external event and based on the predicted at least one event. (B36) The computer-implemented method of statement B35, wherein the at least one event includes any of a weather event, a traffic event, an accident event, and an entertainment event. (B37) The computer-implemented method of any of statements B2-B36, wherein decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component includes using a Periodic Trend Decomposition with LOESS (STL) technique. (B38) The computer-implemented method of any of statements B2-B37, wherein comparing at least one deviation between the first noise component and the second noise component to an abnormality threshold includes comparing a percentage deviation to the abnormality threshold. (B39) The computer-implemented method of statement B1, wherein the percentage deviation is a deviation as a percentage of the value of the first or second noise component (at the relevant time point). (B40) Decomposing the first and second traffic forecasts into components may include decomposing at least one aspect of the first and second traffic forecasts, the at least one aspect comprising: the number of vehicles in said geographic area (over time); the number of each of a plurality of types of vehicles in said geographic area (over time); the location of each of a plurality of vehicles within said geographic region (over time); the heading of each of a plurality of vehicles within the geographic region (over time); the speed of each of a plurality of vehicles within said geographic region (over time); an average speed of each of a plurality of vehicles in said geographic region (over time); the minimum and / or maximum speed of each of a plurality of vehicles in said geographic region (over time); the level of congestion within said geographic area (over time); the number of traffic jams in said geographic area (over time); the level of use of the road network within said geographical area (over time); the maximum transport capacity (of the road network) within said geographical area (over time); an identifier for each of a plurality of vehicles within said geographic region (over time); The computer-implemented method of any of statements B2 to B39, comprising any of the following: (B41) Decomposing the first and second traffic forecasts into components may include decomposing at least one aspect of the first and second traffic forecasts, the at least one aspect comprising: The computer-implemented method of any of statements B2-B40, including any of: a number of vehicles in the geographic region (over time) or a number of each of a plurality of types of vehicles in the geographic region (over time); and an average speed of each of a plurality of vehicles in the geographic region (over time). (B42) The step of estimating emissions comprises: information indicative of average emissions of the vehicle based on said speed; the number of vehicles from said traffic forecast; an average speed of vehicles from said traffic forecast; The computer-implemented method of any of Statements B1-B41, comprising calculating emissions based on: (B43) A computer-implemented method comprising executing the computer-implemented method set forth in any of statements B1-B42 multiple times, with a different geographical area of ​​the geographic region as a target area each time. (B44) The computer-implemented method of statement B43, comprising aggregating the forecasted emissions (and forecasted impacts of detected anomalies) for the multiple regions to obtain a forecast of emissions occurring in the geographic region (including the impacts of detected anomalies). (B45) A computer program which, when executed on a computer, causes the computer to carry out a method, the method comprising: executing a parameter determination process based on a plurality of (historical) traffic data sets for a plurality of geographical areas, said parameter determination process comprising: clustering the traffic datasets based on their similarity to each other to generate a plurality of clusters of traffic datasets, and for each cluster including a plurality of traffic datasets, selecting a traffic dataset of the cluster that is most similar to an average of the traffic datasets of the cluster as a representative traffic dataset; Based on a representative traffic data set corresponding to a target geographical area, performing a mobility analysis to determine (optimal) parameters for a periodic traffic forecast for said target geographical area, and based on at least one representative traffic data set corresponding to at least one other geographical area, performing a mobility analysis to determine (optimal) parameters for at least one other periodic traffic forecast for at least one other geographical area; and performing a traffic prediction process for the target geographic region using a traffic dataset for the target geographic region and at least one traffic dataset for at least one other geographic region, the traffic prediction process comprising: generating the periodic traffic forecast for the target geographical area (by estimating the periodic component of traffic data for the target geographical area) based on the traffic dataset for the target geographical area and using the determined parameters for periodic traffic forecast for the target geographical area (predicting traffic data for a future time period); generating a periodic traffic forecast for the at least one other geographical area (by estimating a periodic component of traffic data for the at least one other geographical area) based on the at least one traffic data set for the at least one other geographical area and using the determined parameters for the at least one other periodic traffic forecast for the at least one other geographical area (predicting traffic data for a future period); analyzing mobility flows in the geographical area based on a traffic dataset of the geographical area and at least one traffic dataset of the at least one other geographical area to determine at least one correlation between the (traffic data trend of) the target geographical area and the at least one other geographical area; generating a combined traffic forecast (or final forecast or cumulative forecast or overall forecast) by adjusting the periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; and forecasting emissions generated by traffic in the target geographic area based on the final traffic forecast; A computer program comprising: (B46) An information processing device comprising: a memory; and a processor coupled to the memory, the processor comprising: executing a parameter determination process based on a plurality of (historical) traffic data sets for a plurality of geographical areas, said parameter determination process comprising: clustering the traffic datasets based on their similarity to each other to generate a plurality of clusters of traffic datasets, and for each cluster including a plurality of traffic datasets, selecting a traffic dataset of the cluster that is most similar to an average of the traffic datasets of the cluster as a representative traffic dataset; Based on a representative traffic data set corresponding to a target geographical area, performing a mobility analysis to determine (optimal) parameters for a periodic traffic forecast for said target geographical area, and based on at least one representative traffic data set corresponding to at least one other geographical area, performing a mobility analysis to determine (optimal) parameters for at least one other periodic traffic forecast for at least one other geographical area; Including, performing a traffic prediction process for the target geographic region using a traffic dataset for the target geographic region and at least one traffic dataset for at least one other geographic region, the traffic prediction process comprising: generating the periodic traffic forecast for the target geographical area (by estimating the periodic component of traffic data for the target geographical area) based on the traffic dataset for the target geographical area and using the determined parameters for periodic traffic forecast for the target geographical area (predicting traffic data for a future time period); generating a periodic traffic forecast for the at least one other geographical area (by estimating a periodic component of traffic data for the at least one other geographical area) based on the at least one traffic data set for the at least one other geographical area and using the determined parameters for the at least one other periodic traffic forecast for the at least one other geographical area (predicting traffic data for a future period); analyzing mobility flows in the geographical area based on a traffic dataset of the geographical area and at least one traffic dataset of the at least one other geographical area to determine at least one correlation between the (traffic data trend of) the target geographical area and the at least one other geographical area; generating a combined traffic forecast (or final forecast or cumulative forecast or overall forecast) by adjusting the periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; Including, forecasting emissions generated by traffic in the target geographic area based on the final traffic forecast; An information processing device configured as follows. In addition to the above-described embodiment, the following supplementary notes are further disclosed. (Appendix 1) 1. A computer-implemented method, comprising: executing a parameter determination process based on a plurality of traffic data sets for a plurality of geographic regions, said parameter determination process comprising: clustering the traffic data set based on similarity to one another to generate a plurality of clusters of the traffic data set; For each cluster including multiple traffic datasets, selecting a traffic dataset of the cluster that is most similar to an average of the traffic datasets of the cluster as a representative traffic dataset; performing a mobility analysis based on a representative traffic data set corresponding to a target geographic area to determine parameters for a periodic traffic forecast for the target geographic area and performing a mobility analysis based on at least one representative traffic data set corresponding to at least one other geographic area to determine parameters for at least one other periodic traffic forecast for the at least one other geographic area; and performing a traffic prediction process for the target geographic region using a traffic dataset for the target geographic region and at least one traffic dataset for the at least one other geographic region, the traffic prediction process comprising: generating a periodic traffic forecast for the target geographic region based on a traffic dataset for the target geographic region and using the determined parameters for a periodic traffic forecast for the target geographic region; generating at least one other periodic traffic forecast for the at least one other geographical area based on at least one traffic dataset for the at least one other geographical area and using the determined parameters for the at least one other periodic traffic forecast for the at least one other geographical area; analyzing mobility flows in the geographical region based on a traffic dataset of the geographical region and at least one traffic dataset of the at least one other geographical region to determine at least one correlation between the target geographical region and the at least one other geographical region; generating a combined traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; and forecasting emissions generated by traffic in the target geographic region based on the multi-modal traffic forecast; 23. A computer-implemented method comprising: (Appendix 2) Clustering the traffic data sets based on their mutual similarity includes clustering the traffic data sets based on at least one aspect of the traffic data, the at least one aspect being one of: The number of vehicles within the geographic area; a number of each of a plurality of types of vehicles within said geographic area; a location of each of a plurality of vehicles within the geographic region; a heading for each of a plurality of vehicles within the geographic area; a speed of each of a plurality of vehicles within the geographic region; an average speed of each of a plurality of vehicles within the geographic region; a minimum and / or maximum speed of each of a plurality of vehicles within said geographic region; the level of congestion within said geographic area; the number of traffic jams within said geographic area; the level of use of the road network within said geographic region; maximum shipping capacity within said geographic area; 2. The computer-implemented method of claim 1, (Appendix 3) 3. The computer-implemented method of claim 1 or 2, wherein clustering the traffic datasets based on their similarity to each other comprises using dynamic time warp matching or Euclidean matching between the traffic datasets or behavior representations. (Appendix 4) 3. The computer-implemented method of claim 1 or 2, wherein the traffic data includes data obtained from sensors within the relevant geographic region. (Appendix 5) 3. The computer-implemented method of claim 1 or 2, wherein the traffic data includes data obtained from a digital twin of the relevant geographic area. (Appendix 6) 3. The computer-implemented method of claim 1 or 2, wherein the target geographic area and the at least one other geographic area are each areas of a metropolitan area or a town or a city. (Appendix 7) traffic data for said geographical area, The number of vehicles within the geographic area; a number of each of a plurality of types of vehicles within said geographic area; a location of each of a plurality of vehicles within the geographic region; a heading for each of a plurality of vehicles within the geographic area; a speed of each of a plurality of vehicles within the geographic region; an average speed of each of a plurality of vehicles within the geographic region; a minimum and / or maximum speed of each of a plurality of vehicles within said geographic region; the level of congestion within said geographic area; the number of traffic jams within said geographic area; the level of use of the road network within said geographic region; maximum shipping capacity within said geographic area; an identifier for each of a plurality of vehicles within the geographic region; 3. The computer-implemented method of claim 1 or 2, (Appendix 8) 3. The computer-implemented method of claim 1 or 2, wherein generating a periodic traffic forecast for the target geographic area comprises generating a regressor based on associated traffic data that defines periodic aspects of the traffic data. (Appendix 9) 3. The computer-implemented method of claim 1 or 2, wherein the traffic prediction process includes analyzing mobility flows of the target geographic area and a plurality of other geographic areas to determine the at least one other geographic area relevant to traffic prediction for the target geographic area. (Appendix 10) The traffic data includes information regarding events occurring in the geographic region of interest, and the traffic prediction process comprises: analyzing traffic data for the target geographic area to obtain at least one correlation between the traffic data and at least one event; Further comprising: generating a traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined correlation, 3. The computer-implemented method of claim 1 or 2, further comprising adjusting a periodic forecast for the target geographic area based on at least one correlation between traffic data and at least one external event and based on the predicted at least one event. (Appendix 11) The step of estimating emissions comprises: information indicating the average emissions of a vehicle based on speed; the number of vehicles from said traffic forecast; an average speed of vehicles from said traffic forecast; 3. The computer-implemented method of claim 1 or 2, comprising calculating emissions based on: (Appendix 12) The traffic dataset for the target geographic region and at least one traffic dataset for the at least one other geographic region are traffic data for a first time period, the combined traffic forecast is a first traffic forecast, and the computer-implemented method includes: generating a second traffic forecast by running the traffic forecasting process using traffic data for a second time period, the first time period including the second time period and an additional time period more recent than the second time period; The computer-implemented method comprises: decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component, and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component; comparing the first noise component of the first traffic forecast with a second noise component of the second traffic forecast to detect at least one anomaly, the comparing at least one deviation between the first traffic forecast and the second traffic forecast with an anomaly threshold; 3. The computer-implemented method of claim 1 or 2, wherein the step of predicting emissions generated by traffic in the target geographic area includes, if at least one anomaly is detected, predicting an impact of the at least one detected anomaly on emissions. (Appendix 13) 3. A computer-implemented method comprising performing the computer-implemented method of claim 1 or 2 multiple times, with a different geographic area of ​​the geographic region as a target area each time. (Appendix 14) A computer program which, when executed on a computer, causes the computer to carry out a method, the method comprising: executing a parameter determination process based on a plurality of traffic data sets for a plurality of geographic regions, said parameter determination process comprising: clustering the traffic data set based on similarity to one another to generate a plurality of clusters of the traffic data set; For each cluster including multiple traffic datasets, selecting a traffic dataset of the cluster that is most similar to an average of the traffic datasets of the cluster as a representative traffic dataset; performing a mobility analysis based on a representative traffic data set corresponding to a target geographic area to determine parameters for a periodic traffic forecast for the target geographic area; and performing a mobility analysis based on at least one representative traffic data set corresponding to at least one other geographic area to determine parameters for at least one other periodic traffic forecast for the at least one other geographic area. and performing a traffic prediction process for the target geographic region using a traffic dataset for the target geographic region and at least one traffic dataset for the at least one other geographic region, the traffic prediction process comprising: generating a periodic traffic forecast for the target geographic region based on a traffic dataset for the target geographic region and using the determined parameters for a periodic traffic forecast for the target geographic region; generating at least one other periodic traffic forecast for the at least one other geographical area based on at least one traffic dataset for the at least one other geographical area and using the determined parameters for the at least one other periodic traffic forecast for the at least one other geographical area; analyzing mobility flows in the geographical area based on a traffic dataset of the geographical area and at least one traffic dataset of the at least one other geographical area to determine at least one correlation between the target geographical area and the at least one other geographical area; generating a combined traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; and forecasting emissions generated by traffic in the target geographic region based on the multi-modal traffic forecast; A computer program comprising: (Appendix 15) An information processing device comprising: a memory; and a processor coupled to the memory, the processor comprising: executing a parameter determination process based on a plurality of traffic data sets for a plurality of geographic regions, said parameter determination process comprising: clustering the traffic dataset based on similarity to one another to generate a plurality of clusters of the traffic dataset; For each cluster including multiple traffic datasets, selecting a traffic dataset of the cluster that is most similar to an average of the traffic datasets of the cluster as a representative traffic dataset; performing a mobility analysis based on a representative traffic data set corresponding to a target geographic area to determine parameters for a periodic traffic forecast for the target geographic area, and performing a mobility analysis based on at least one representative traffic data set corresponding to at least one other geographic area to determine parameters for at least one other periodic traffic forecast for the at least one other geographic area; Including, performing a traffic prediction process for the target geographic region using a traffic dataset for the target geographic region and at least one traffic dataset for the at least one other geographic region, the traffic prediction process comprising: generating a periodic traffic forecast for the target geographic region based on a traffic dataset for the target geographic region and using the determined parameters for a periodic traffic forecast for the target geographic region; generating at least one other periodic traffic forecast for the at least one other geographical area based on at least one traffic dataset for the at least one other geographical area and using the determined parameters for the at least one other periodic traffic forecast for the at least one other geographical area; analyzing mobility flows in the geographical region based on a traffic dataset of the geographical region and at least one traffic dataset of the at least one other geographical region to determine at least one correlation between the target geographical region and the at least one other geographical region; generating a combined traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined at least one correlation; Including, forecasting emissions generated by traffic in the target geographic region based on the multimodal traffic forecast; An information processing device configured as follows. [Explanation of symbols]

[0153] 20 DT Mobility Analysis Department 40 DT Mobility Prediction 60 DT Mobility Anomaly Detection Unit 80 Emissions Estimate Engine

Claims

1. 1. A computer-implemented method comprising: executing a parameter determination process based on a plurality of traffic data sets for a plurality of geographic regions, said parameter determination process comprising: clustering the traffic data set based on similarity to one another to generate a plurality of clusters of the traffic data set; For each cluster including multiple traffic datasets, selecting a traffic dataset of the cluster that is most similar to an average of the traffic datasets of the cluster as a representative traffic dataset; performing a mobility analysis based on a representative traffic data set corresponding to a target geographic area to determine parameters for a periodic traffic forecast for the target geographic area and performing a mobility analysis based on at least one representative traffic data set corresponding to at least one other geographic area to determine parameters for at least one other periodic traffic forecast for the at least one other geographic area; and performing a traffic prediction process for the target geographic region using a traffic dataset for the target geographic region and at least one traffic dataset for the at least one other geographic region, the traffic prediction process comprising: generating a periodic traffic forecast for the target geographic region based on a traffic dataset for the target geographic region and using the determined parameters for a periodic traffic forecast for the target geographic region; generating at least one other periodic traffic forecast for the at least one other geographical area based on at least one traffic dataset for the at least one other geographical area and using the determined parameters for the at least one other periodic traffic forecast for the at least one other geographical area; analyzing mobility flows in a geographical region based on a traffic dataset of the geographical region and at least one traffic dataset of the at least one other geographical region to determine at least one correlation between the target geographical region and the at least one other geographical region; generating a combined traffic forecast by adjusting a periodic forecast for the target geographic region based on the at least one other periodic forecast and the determined at least one correlation; and forecasting emissions generated by traffic in the target geographic region based on the multi-modal traffic forecast; 23. A computer-implemented method comprising:

2. Clustering the traffic data sets based on their mutual similarity includes clustering the traffic data sets based on at least one aspect of the traffic data, the at least one aspect being one of: The number of vehicles within the geographic area; a number of each of a plurality of types of vehicles within said geographic area; a location of each of a plurality of vehicles within the geographic region; a heading for each of a plurality of vehicles within the geographic area; a speed of each of a plurality of vehicles within said geographic region; an average speed of each of a plurality of vehicles within the geographic region; a minimum and / or maximum speed of each of a plurality of vehicles within said geographic region; the level of congestion within said geographic area; the number of traffic jams within said geographic area; the level of use of the road network within said geographic region; maximum shipping capacity within said geographic area; The computer-implemented method of claim 1 , comprising:

3. The computer-implemented method of claim 1 or 2, wherein clustering the traffic datasets based on their similarity to each other comprises using dynamic time warp matching or Euclidean matching between the traffic datasets or behavior representations.

4. The computer-implemented method of claim 1 or 2, wherein the traffic data comprises data obtained from sensors within the geographical region of interest.

5. The computer-implemented method of claim 1 or 2, wherein the traffic data comprises data obtained from a digital twin of the associated geographic area.

6. The computer-implemented method of claim 1 or 2, wherein the target geographic area and the at least one other geographic area are each areas of a metropolitan area or a town or a city.

7. traffic data for said geographical area, The number of vehicles within the geographic area; a number of each of a plurality of types of vehicles within said geographic area; a location of each of a plurality of vehicles within the geographic region; a heading for each of a plurality of vehicles within the geographic area; a speed of each of a plurality of vehicles within said geographic region; an average speed of each of a plurality of vehicles within the geographic region; a minimum and / or maximum speed of each of a plurality of vehicles within said geographic region; the level of congestion within said geographic area; the number of traffic jams within said geographic area; the level of use of the road network within said geographic region; maximum shipping capacity within said geographic area; an identifier for each of a plurality of vehicles within the geographic region; 3. The computer-implemented method of claim 1, further comprising:

8. The computer-implemented method of claim 1 or 2, wherein generating a periodic traffic forecast for the target geographic area comprises generating regressors based on associated traffic data that define periodic aspects of the traffic data.

9. 3. The computer-implemented method of claim 1, wherein the traffic prediction process includes analyzing mobility flows of the target geographical area and a plurality of other geographical areas to determine the at least one other geographical area relevant to traffic prediction for the target geographical area.

10. The traffic data includes information regarding events occurring in the geographic region of interest, and the traffic prediction process comprises: analyzing traffic data for the target geographic area to obtain at least one correlation between the traffic data and at least one event; Further comprising: generating a traffic forecast by adjusting a periodic forecast for the target geographic area based on the at least one other periodic forecast and the determined correlation, 3. The computer-implemented method of claim 1, further comprising adjusting a periodic forecast for the target geographic area based on at least one correlation between traffic data and at least one external event and based on the predicted at least one event.

11. The step of estimating emissions comprises: information indicating the average emissions of a vehicle based on speed; the number of vehicles from said traffic forecast; an average speed of vehicles from said traffic forecast; The computer-implemented method of claim 1 or 2, comprising calculating an emission amount based on:

12. The traffic dataset for the target geographic region and at least one traffic dataset for the at least one other geographic region are traffic data for a first time period, the combined traffic forecast is a first traffic forecast, and the computer-implemented method includes: generating a second traffic forecast by running the traffic forecasting process using traffic data for a second time period, the first time period including the second time period and an additional time period more recent than the second time period; The computer-implemented method comprises: decomposing the first traffic forecast into a first periodic component, a first trend component, and a first noise component, and decomposing the second traffic forecast into a second periodic component, a second trend component, and a second noise component; comparing the first noise component of the first traffic prediction with a second noise component of the second traffic prediction to detect at least one anomaly, the comparison comprising comparing at least one deviation between the first traffic prediction and the second traffic prediction with an anomaly threshold; 3. The computer-implemented method of claim 1 or 2, wherein predicting emissions generated by traffic in the target geographical area comprises, if at least one anomaly is detected, predicting an impact of the at least one detected anomaly on emissions.

13. 3. A computer-implemented method comprising performing the computer-implemented method of claim 1 or 2 a plurality of times, each time with a different geographical area of ​​the geographic region as a target area.

14. A computer program which, when executed on a computer, causes the computer to carry out a method, the method comprising: executing a parameter determination process based on a plurality of traffic data sets for a plurality of geographic regions, said parameter determination process comprising: clustering the traffic data set based on similarity to one another to generate a plurality of clusters of the traffic data set; For each cluster including multiple traffic datasets, selecting a traffic dataset of the cluster that is most similar to an average of the traffic datasets of the cluster as a representative traffic dataset; performing a mobility analysis based on a representative traffic data set corresponding to a target geographic region to determine parameters for a periodic traffic forecast for the target geographic region; and performing a mobility analysis based on at least one representative traffic data set corresponding to at least one other geographic region to determine parameters for at least one other periodic traffic forecast for the at least one other geographic region. and performing a traffic prediction process for the target geographic region using a traffic dataset for the target geographic region and at least one traffic dataset for the at least one other geographic region, the traffic prediction process comprising: generating a periodic traffic forecast for the target geographic region based on a traffic dataset for the target geographic region and using the determined parameters for a periodic traffic forecast for the target geographic region; generating at least one other periodic traffic forecast for the at least one other geographical area based on at least one traffic dataset for the at least one other geographical area and using the determined parameters for the at least one other periodic traffic forecast for the at least one other geographical area; analyzing mobility flows in the geographical area based on a traffic dataset of the geographical area and at least one traffic dataset of the at least one other geographical area to determine at least one correlation between the target geographical area and the at least one other geographical area; generating a combined traffic forecast by adjusting a periodic forecast for the target geographic region based on the at least one other periodic forecast and the determined at least one correlation; and forecasting emissions generated by traffic in the target geographic region based on the multi-modal traffic forecast; A computer program comprising:

15. An information processing device comprising: a memory; and a processor coupled to the memory, the processor comprising: executing a parameter determination process based on a plurality of traffic data sets for a plurality of geographic regions, said parameter determination process comprising: clustering the traffic dataset based on similarity to one another to generate a plurality of clusters of the traffic dataset; for each cluster including multiple traffic datasets, selecting a traffic dataset of the cluster that is most similar to an average of the traffic datasets of the cluster as a representative traffic dataset; performing a mobility analysis based on a representative traffic data set corresponding to a target geographical area to determine parameters for a periodic traffic forecast for the target geographical area and performing a mobility analysis based on at least one representative traffic data set corresponding to at least one other geographical area to determine parameters for at least one other periodic traffic forecast for the at least one other geographical area; Including, performing a traffic prediction process for the target geographic region using a traffic dataset for the target geographic region and at least one traffic dataset for the at least one other geographic region, the traffic prediction process comprising: generating a periodic traffic forecast for the target geographic region based on a traffic dataset for the target geographic region and using the determined parameters for a periodic traffic forecast for the target geographic region; generating at least one other periodic traffic forecast for the at least one other geographical area based on at least one traffic dataset for the at least one other geographical area and using the determined parameters for the at least one other periodic traffic forecast for the at least one other geographical area; analyzing mobility flows in a geographical region based on a traffic dataset of the geographical region and at least one traffic dataset of the at least one other geographical region to determine at least one correlation between the target geographical region and the at least one other geographical region; generating a combined traffic forecast by adjusting a periodic forecast for the target geographic region based on the at least one other periodic forecast and the determined at least one correlation; Including, forecasting emissions generated by traffic in the target geographic region based on the multimodal traffic forecast; An information processing device configured as follows.

Images