Rail transit linkage scheme prediction method, device and equipment and storage medium

Abstract

The application provides a rail transit linkage scheme prediction method, device and equipment and a storage medium, and is applied to the technical field of rail transit. The method comprises the following steps: acquiring first monitoring data collected by a first sensor at a target object in rail transit; if it is detected that the first monitoring data is abnormal, acquiring abnormal monitoring data existing in the first monitoring data; inputting the abnormal monitoring data into a preset deep learning model for linkage scheme prediction processing, and determining at least one candidate linkage scheme corresponding to the target object; the deep learning model is obtained through supervised training according to a plurality of sample linkage schemes and a known working condition corresponding to each sample linkage scheme and / or unsupervised training according to a plurality of known working conditions; and determining a target linkage scheme corresponding to the target object according to the at least one candidate linkage scheme of the target object. The technical scheme of the application can improve the intelligence of the predicted linkage scheme.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, and in particular to a method, apparatus, equipment, and storage medium for predicting rail transit linkage schemes. Background Technology

[0002] With the rapid development of urban rail transit, the operational complexity of rail transit is constantly increasing, and passengers' requirements for the rail transit travel experience are also rising. For example, when the temperature of the rail transit platform is too high, the rail transit system needs to determine a plan to adjust the platform temperature based on the platform temperature in order to cool down the platform and ensure that the platform temperature is suitable.

[0003] In related technologies, when a rail transit system needs to set an adjustment plan for a certain area or equipment, it usually determines the appropriate adjustment plan for that area or equipment based on a fixed correspondence between the parameters of the area or equipment and the adjustment plan.

[0004] However, the adjustment scheme determined by the aforementioned technology is not intelligent enough. Summary of the Invention

[0005] This invention provides a method, apparatus, device, and storage medium for predicting linkage schemes in rail transit. It addresses the shortcomings of existing technologies where determining the appropriate adjustment scheme for a region or device through a fixed correspondence between pre-set parameters and adjustment schemes leads to insufficient intelligence. The invention utilizes a pre-trained deep learning model to predict linkage schemes based on anomaly monitoring data at target locations within rail transit. This intelligently generates more reasonable linkage schemes, enhancing the intelligence of the predicted linkage schemes and improving the overall linkage capability of the rail transit system. Ultimately, this achieves the goals of improving operational efficiency, ensuring system safety, and optimizing passenger experience.

[0006] This invention provides a method for predicting the linkage scheme of rail transit, comprising:

[0007] Acquire first monitoring data collected by a first sensor at a target object in rail transit; the target object includes a target area and / or a target vehicle.

[0008] Anomaly detection is performed on the first monitoring data. If anomalies are detected in the first monitoring data, the abnormal monitoring data containing the anomalies in the first monitoring data is obtained.

[0009] Anomaly monitoring data is input into a pre-set deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object; each candidate linkage scheme includes multiple first candidate actions for linkage, and each first candidate action is an adjustment action to adjust the abnormal situation of the target object. The above deep learning model is obtained in advance by supervised training and / or unsupervised training based on multiple sample linkage schemes and the known working conditions corresponding to each sample linkage scheme.

[0010] Based on at least one candidate linkage scheme for the target object, determine the target linkage scheme corresponding to the target object; the target linkage scheme includes multiple linkage target actions and control parameters for each target action.

[0011] According to the present invention, a method for predicting linkage schemes in rail transit includes determining a target linkage scheme corresponding to the target object based on at least one candidate linkage scheme for the target object, comprising:

[0012] Based on the priority of each first candidate action in each candidate linkage scheme, determine the initial quantization value of each first candidate action in each candidate linkage scheme; the priorities of each first candidate action in the above candidate linkage schemes are not completely the same;

[0013] Based on the initial quantization value corresponding to each first candidate action in each candidate linkage scheme, each candidate linkage scheme is quantized to determine the target quantization value corresponding to each candidate linkage scheme.

[0014] Based on the target quantification value corresponding to each candidate linkage scheme, the target linkage scheme is determined from the candidate linkage schemes.

[0015] According to the present invention, a method for predicting linkage schemes in rail transit includes determining a target linkage scheme corresponding to the target object based on at least one candidate linkage scheme for the target object, comprising:

[0016] Display at least one candidate linkage scheme;

[0017] Get the user's selection action input from at least one candidate linkage scheme;

[0018] In response to the selection operation, the candidate linkage scheme corresponding to the selection operation is determined as the target linkage scheme.

[0019] According to the method for predicting rail transit linkage schemes provided by the present invention, the method further includes:

[0020] Acquire performance data of the target object and / or performance data of the target device within the target object;

[0021] Based on the performance data of the target object and / or the performance data of the target equipment, adjust the control parameters of at least one target linkage action in the target linkage scheme.

[0022] According to the method for predicting rail transit linkage schemes provided by the present invention, the training method of the above-mentioned deep learning model includes:

[0023] Obtain the sample linkage schemes for multiple sample objects and the known operating conditions corresponding to each sample linkage scheme; the known operating conditions include sample anomaly monitoring data at the sample objects;

[0024] Each known working condition is input into the initial deep learning model for linkage scheme prediction processing, and the predicted linkage scheme corresponding to the sample object for each known working condition is determined.

[0025] Calculate the loss between each predicted linkage scheme and the corresponding sample linkage scheme;

[0026] The initial deep learning model is trained in a supervised manner based on each loss and / or in an unsupervised manner based on each known working condition to obtain the deep learning model. The supervised training is used to learn the mapping relationship between the known working conditions and the corresponding sample linkage scheme, while the unsupervised training is used to mine the inherent correlation between various types of data in the known working conditions.

[0027] According to the method for predicting rail transit linkage schemes provided by the present invention, the method further includes:

[0028] Acquire second monitoring data collected by the second sensor at the target object;

[0029] The above-mentioned process involves inputting anomaly monitoring data into a pre-defined deep learning model for predicting linkage schemes, thereby determining at least one candidate linkage scheme corresponding to the target object, including:

[0030] The abnormal monitoring data and the second monitoring data are input into a preset deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object; the above candidate linkage scheme includes each first candidate action and second candidate action, and the above second candidate action is used to perform energy-saving actions on energy-related equipment in the target object.

[0031] According to the method for predicting rail transit linkage schemes provided by the present invention, the method further includes:

[0032] Acquire the execution result obtained after the device at the target object at a historical time performs the first candidate action, and acquire the third monitoring data collected by the third sensor at the target object;

[0033] The above-mentioned process involves inputting anomaly monitoring data into a pre-defined deep learning model for predicting linkage schemes, thereby determining at least one candidate linkage scheme corresponding to the target object, including:

[0034] Anomaly monitoring data, execution results, and third-party monitoring data are input into a pre-set deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object.

[0035] The present invention also provides a predictive device for rail transit linkage schemes, comprising the following modules:

[0036] The first monitoring data acquisition module is used to acquire first monitoring data collected by a first sensor at a target object in the rail transit; the target object includes a target area and / or a target vehicle.

[0037] An anomaly detection module is used to perform anomaly detection on the first monitoring data. If an anomaly is detected in the first monitoring data, the abnormal monitoring data containing the anomaly in the first monitoring data is obtained.

[0038] The linkage scheme prediction module is used to input anomaly monitoring data into a preset deep learning model for linkage scheme prediction processing, and to determine at least one candidate linkage scheme corresponding to the target object. Each candidate linkage scheme includes multiple first candidate actions for linkage, and each first candidate action is an adjustment action to adjust the abnormal situation of the target object. The aforementioned deep learning model is obtained in advance by supervised training and / or unsupervised training based on multiple sample linkage schemes and the known working conditions corresponding to each sample linkage scheme.

[0039] The target linkage scheme determination module is used to determine the target linkage scheme corresponding to the target object based on at least one candidate linkage scheme for the target object; the target linkage scheme includes multiple linkage target actions and control parameters for each target action.

[0040] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the rail transit linkage scheme prediction method as described above.

[0041] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the rail transit linkage scheme prediction method as described above.

[0042] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the rail transit linkage scheme prediction method as described above.

[0043] The present invention provides a method, apparatus, device, and storage medium for predicting linkage schemes in rail transit. It acquires first monitoring data collected by a first sensor at a target object in the rail transit system, including a target area and / or a target vehicle. Anomaly detection is performed on the first monitoring data. If anomalies are detected, abnormal monitoring data is acquired from the first monitoring data, and this abnormal monitoring data is input into a pre-trained deep learning model for linkage scheme prediction. At least one candidate linkage scheme is determined for the target object. Then, a target linkage scheme for the target object is determined based on the at least one candidate linkage scheme. The deep learning model is pre-trained using supervised and / or unsupervised training based on multiple sample linkage schemes and the known operating conditions corresponding to each sample linkage scheme. Each candidate linkage scheme includes multiple first candidate actions for linkage, and each first candidate action corresponds to an adjustment action to address anomalies in the target object. The target linkage scheme includes multiple target actions for linkage and control parameters for each target action. In this method, since a deep learning model pre-trained using a combination of supervised and unsupervised methods can predict linkage schemes based on monitoring data at rail transit target objects, the linkage schemes predicted by the technical solution of this invention are more intelligent and reasonable compared to adjustment schemes determined by fixed correspondences. This improves the intelligence of the predicted linkage schemes and enhances the linkage of the rail transit system. Furthermore, since anomaly monitoring can be performed on the data at rail transit target objects, and linkage scheme prediction can be performed using a pre-trained deep learning model when anomalies are detected, the efficiency of predicting linkage schemes is higher. Therefore, this allows the rail transit system to respond to anomalies promptly and quickly, ensuring the operational safety of rail transit and improving the passenger experience. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0045] Figure 1 This is one of the flowcharts illustrating the predictive method for rail transit linkage schemes provided by the present invention.

[0046] Figure 2 This is the second flowchart of the rail transit linkage scheme prediction method provided by the present invention.

[0047] Figure 3 This is the third flowchart of the rail transit linkage scheme prediction method provided by the present invention.

[0048] Figure 4 This is a schematic diagram of the structure of the rail transit linkage scheme prediction device provided by the present invention.

[0049] Figure 5 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0051] Currently, when rail transit systems need to set adjustment plans for a certain area or equipment, the appropriate adjustment plan is usually determined based on a pre-defined fixed correspondence between the parameters of the area or equipment and the adjustment plan. However, the adjustment plans determined by the above technology are not intelligent enough. For example, when the platform temperature rises, only the air conditioning can be activated to cool it down, without comprehensively considering other related factors such as ventilation; when equipment malfunctions, it cannot promptly coordinate with related systems to make a comprehensive response; during peak passenger flow, the diversion measures are simplistic and difficult to efficiently divert passengers. In addition, the above technology cannot automatically associate and recommend more optimized linkage plans based on real-time changes, and cannot meet the needs of efficient, safe, and comfortable operation of modern rail transit. Based on this, embodiments of the present invention provide a method, device, equipment, and storage medium for predicting linkage plans for rail transit, which can solve the above-mentioned technical problems.

[0052] It should be noted that the executing entity in the embodiments of the present invention can be a rail transit linkage scheme prediction device, an electronic device including a rail transit linkage scheme prediction device, a rail transit system including a rail transit linkage scheme prediction device, or other equipment, systems, or devices, without specific limitations. The following embodiments will use a rail transit system as an example for illustration. In addition, the aforementioned electronic device can be a terminal or a server.

[0053] Figure 1 This is one of the flowcharts illustrating the rail transit linkage scheme prediction method provided by the present invention, such as... Figure 1 As shown, the method includes the following steps:

[0054] Step 102: Obtain the first monitoring data collected by the first sensor at the target object in the rail transit; the target object includes the target area and / or the target vehicle.

[0055] In rail transit, target objects can include target areas or target vehicles. Target areas include specific areas within the rail transit system, such as platforms, station entrances, station exits, and vehicle transfer channels. A target area can also be a sub-area within a platform area. Target vehicles can be vehicles operating on the rail transit line, such as trains or buses. One or more types of sensors can be pre-distributed or installed at key locations of any of these target objects to collect relevant data. The number of sensors of each type can be one or more. These sensors include, but are not limited to, temperature sensors, humidity sensors, light sensors, passenger flow counters, equipment status monitors, smoke sensors, and position sensors. Each type of sensor can collect corresponding monitoring data. These sensors can all use high-precision, high-reliability industrial-grade sensor elements, possessing the ability to collect data in real time and continuously. Their sampling frequency is set according to the changing characteristics of different data types. For example, temperature and humidity sensors collect temperature data every 2 minutes, and passenger flow counters update passenger flow data every 30 seconds.

[0056] Specifically, one or more first sensors at the aforementioned target object can collect monitoring data of the corresponding type. The collected monitoring data is recorded as the first monitoring data. Here, the first monitoring data can be monitoring data at a certain moment or monitoring data within a certain time period.

[0057] Of course, other types of sensors at the target object can also collect their corresponding monitoring data. Here, both the primary sensor and other types of sensors can collect data in real time to ensure data real-time performance and the real-time nature of subsequent predictive linkage schemes.

[0058] Step 104: Perform anomaly detection on the first monitoring data. If anomalies are detected in the first monitoring data, obtain the abnormal monitoring data in the first monitoring data that contains anomalies.

[0059] In this step, after obtaining the first monitoring data collected by the first sensor at the target object, a conditional triggering mechanism can be set to facilitate subsequent determination of when the linkage scheme prediction process needs to be initiated. Specifically, after obtaining the first monitoring data, the mechanism can detect whether there is an anomaly in the first monitoring data through set abnormal conditions. If an anomaly is detected in the first monitoring data, the subsequent linkage scheme prediction function is triggered, and the subsequent linkage scheme prediction process can continue. If no anomaly is detected in the first monitoring data, the subsequent linkage scheme prediction function is not triggered, and monitoring data can continue to be acquired for anomaly detection.

[0060] When detecting anomalies in the first monitoring data, the data can be matched against predefined anomaly conditions. If a match is successful, the anomaly in the first monitoring data is confirmed, triggering subsequent linkage scheme prediction functions. These anomaly conditions can include multiple trigger points, such as thresholds or threshold ranges related to the data type, data change thresholds within a set time period, or fault count thresholds within a set time period. The set time can be configured according to actual conditions, such as 2 minutes, 5 minutes, or 10 minutes. For example, in a station environment, if a temperature sensor in the middle of a station detects a temperature increase from 28 degrees to 32 degrees within 10 minutes (a 4-degree increase), exceeding the preset data change threshold of 3 degrees / 5 minutes (a 3-degree increase within 5 minutes), then the temperature data in that station is determined to be anomaly. For example, if a train in operation suddenly issues a fault warning through its door system, such as a door closing time exceeding 3 seconds and four consecutive error reports, exceeding the door closing time threshold by 2 seconds, and the number of faults exceeding the threshold of zero, then the train's door closing data is determined to be abnormal. Similarly, in terms of equipment operation, if a train's door system detects a fault warning, such as a door closing time exceeding 2 seconds or more frequent error reports exceeding 3 within 10 minutes, then the train's operational data is determined to be abnormal, and the system immediately triggers and initiates the associative recommendation function / process.

[0061] When an anomaly is detected in the first monitoring data, the abnormal monitoring data within the first monitoring data can be acquired and recorded as abnormal monitoring data. For example, all temperature data collected within 10 minutes can be used as abnormal monitoring data, or temperature data at any time within 10 minutes can be used as abnormal monitoring data, etc.

[0062] Alternatively, the monitoring data collected by other types of sensors at the target object can be monitored for anomalies in the manner described above to obtain the results of whether there are any anomalies in other types of monitoring data, and to obtain the monitoring data with anomalies in other types of monitoring data when anomalies are found.

[0063] Step 106: Input the anomaly monitoring data into a preset deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object; each candidate linkage scheme includes multiple first candidate actions for linkage, and each first candidate action is an adjustment action to adjust the abnormal situation of the target object. The above deep learning model is obtained in advance by supervised training and / or unsupervised training based on multiple sample linkage schemes and the known working conditions corresponding to each sample linkage scheme.

[0064] In this step, to facilitate more intelligent prediction of linkage schemes, a deep learning model can be pre-trained. This deep learning model can be trained using either unsupervised or supervised learning methods, or a combination of both. During deep learning model training, supervised training can be performed using multiple known working conditions and the optimal linkage scheme under each known working condition. This allows the model to quickly and accurately learn the mapping relationship between working conditions and linkage schemes. Simultaneously, unsupervised training can be performed using multiple known working conditions, allowing the deep learning model to utilize its imagination / association capabilities, associating possible linkage schemes based on these known working conditions, thus improving the intelligence and diversity of the associated schemes. Furthermore, the architecture and type of the aforementioned deep learning model can be set according to the actual situation. For example, it can be a deep learning model built based on deep learning frameworks such as TensorFlow or PyTorch, which can adopt a multi-layer neural network structure, including an input layer, hidden layers, and an output layer. The number of hidden layers can be set to 3-5 layers depending on the complexity of data association, and the number of neurons in each layer can be determined through experiments and optimization.

[0065] The optimal linkage scheme under each of the above-mentioned known operating conditions can be denoted as a sample linkage scheme. Each sample linkage scheme can include multiple linked sample actions. Each sample action includes an adjustment action to address the abnormal situation under the known operating conditions. Through a series of linked sample actions in each sample linkage scheme, the abnormal situation under the known operating conditions can be improved, allowing the rail transit to return to normal as soon as possible. The above-mentioned known operating conditions can be abnormal monitoring data of sample areas or sample vehicles in the rail transit at a historical time. Of course, it can also include normal monitoring data. Here, the monitoring data of the known operating conditions can be obtained by multiple sensors in the corresponding sample areas or sample vehicles.

[0066] The deep learning model can be trained using the above method. The input to this trained model is monitoring data of the target object, and the output is the associated / predicted linkage scheme for the target object. After obtaining the anomaly monitoring data at the target object, this data can be directly input into the trained deep learning model, or it can be preprocessed before being input into the model. Alternatively, the anomaly monitoring data can be input into the model along with other types of anomaly monitoring data and normal monitoring data at the target object. The deep learning model can then predict the linkage scheme based on the input monitoring data. The predicted linkage scheme can include one or more schemes, each designated as a candidate scheme. Each candidate scheme can include multiple linkage actions, designated as first candidate actions. Each first candidate action includes an adjustment action to address the anomaly at the target object. Through a series of first candidate actions within each candidate scheme, the anomaly at the target object can be improved, allowing the target object to quickly return to normal. For example, taking a platform environment as an example, if a temperature sensor in the middle of a platform detects that the temperature rises from 28°C to 32°C within 10 minutes, which is 4 degrees Celsius higher than the preset data change threshold, indicating an anomaly, a well-trained deep learning model can predict the linkage scheme for that platform. This includes multiple linkage actions, such as turning on the air conditioning at that platform for cooling, turning on / adjusting the wind speed and direction of the ventilation system, etc.

[0067] Step 108: Determine the target linkage scheme corresponding to the target object based on at least one candidate linkage scheme for the target object; the target linkage scheme includes multiple linkage target actions and control parameters for each target action.

[0068] In this step, after obtaining at least one candidate linkage scheme for the target object, if only one candidate linkage scheme is obtained through the deep learning model, this candidate linkage scheme can be directly used as the target linkage scheme for the target object. If multiple candidate linkage schemes are obtained through the deep learning model, a target linkage scheme can be selected from the multiple candidate linkage schemes based on the order of the identifiers (e.g., numbers) and the probabilities of each candidate linkage scheme given by the deep learning model. Alternatively, a target linkage scheme can be selected from the multiple candidate linkage schemes by comprehensively analyzing the first candidate actions in each candidate linkage scheme and using the comprehensive analysis results.

[0069] Understandably, the final selected candidate linkage scheme can be denoted as the target linkage scheme, and each first candidate action in the candidate linkage scheme is also denoted as a target action. In other words, the target linkage scheme can include multiple linked target actions, and can also include control parameters for each target action. These control parameters can be, for example, the specific start time and cooling temperature when the air conditioner is turned on, or the specific wind direction, magnitude, and wind speed when the ventilation is turned on.

[0070] Furthermore, to facilitate the implementation of the technical solution of this invention, a rail transit linkage prediction system can be pre-built. This system comprises a three-layer architecture: a data acquisition layer, a data analysis and association layer, and a decision-making and linkage execution layer. The data acquisition layer can collect monitoring data, i.e., execute step 102 above, and transmit the collected monitoring data to the data analysis and association layer via a wired and wireless hybrid data transmission network (such as using a combination of low-power Bluetooth, Wi-Fi, and wired Ethernet to ensure the stability and timeliness of data transmission). The data analysis and association layer can utilize big data storage and processing technologies to build a big data platform based on distributed file systems (such as Ceph, GlusterFS, etc.) and distributed computing frameworks (such as Apache Flink, Spark, etc.), aggregating massive monitoring data from the data acquisition layer, and performing data detection, analysis, and linkage prediction processing on the monitoring data, i.e., executing steps 104 and 106 above. When receiving real-time data change signals from the data acquisition layer, based on the constructed and trained deep learning model, a series of possible linkage actions and linkage schemes are quickly and automatically associated. The decision-making and linkage execution layer is responsible for receiving the association recommendation results from the data analysis and association layer. It can make decisions on the candidate linkage schemes predicted by the deep learning model of the data analysis and association layer to select the target linkage scheme, which is to execute step 108 above. Furthermore, after selecting the target linkage scheme, the decision-making and linkage execution layer can also execute the target linkage scheme. Specifically, it can immediately trigger the corresponding systems and equipment to perform linkage operations, such as controlling the air conditioning system, ventilation system, lighting system, train operation control system, broadcasting system, and gate system, to achieve intelligent rail transit operation management. The above linkage operation execution process can achieve precise control of equipment or systems through industrial automation control protocols (such as Modbus, OPC UA / OPC Unified Architecture, etc.) to ensure the accuracy and timeliness of the target action execution in the target linkage scheme.

[0071] In this embodiment, first monitoring data collected by a first sensor at a target object in the rail transit system, including the target area and / or the target vehicle, is acquired. Anomaly detection is performed on the first monitoring data. If an anomaly is detected in the first monitoring data, the abnormal monitoring data containing the anomaly is acquired and input into a pre-trained deep learning model for linkage scheme prediction processing. At least one candidate linkage scheme corresponding to the target object is determined. Then, the target linkage scheme of the target object is determined based on the at least one candidate linkage scheme of the target object. The deep learning model is obtained in advance through supervised and / or unsupervised training based on multiple sample linkage schemes and the known working conditions corresponding to each sample linkage scheme. Each candidate linkage scheme includes multiple first candidate actions for linkage. Each first candidate action is an adjustment action that adjusts the target object in response to the anomaly. The target linkage scheme includes multiple target actions for linkage and control parameters for each target action. In this method, since a deep learning model pre-trained using a combination of supervised and unsupervised methods can predict linkage schemes based on monitoring data at rail transit target objects, the linkage schemes predicted by the technical solution of this invention are more intelligent and reasonable compared to adjustment schemes determined by fixed correspondences. This improves the intelligence of the predicted linkage schemes and enhances the linkage of the rail transit system. Furthermore, since anomaly monitoring can be performed on the data at rail transit target objects, and linkage scheme prediction can be performed using a pre-trained deep learning model when anomalies are detected, the efficiency of predicting linkage schemes is higher. Therefore, this allows the rail transit system to respond to anomalies promptly and quickly, ensuring the operational safety of rail transit and improving the passenger experience.

[0072] The following embodiment illustrates an implementation method that involves comprehensively analyzing each candidate action in each candidate linkage scheme and selecting a target linkage scheme from multiple candidate linkage schemes based on the comprehensive analysis results.

[0073] Figure 2 This is the second flowchart illustrating the predictive method for rail transit linkage schemes provided by this invention, as shown below. Figure 2 As shown, step 108 above, which determines the target linkage scheme corresponding to the target object based on at least one candidate linkage scheme, may include the following steps:

[0074] Step 202: Determine the initial quantization value corresponding to each first candidate action in each candidate linkage scheme according to the priority of each first candidate action in each candidate linkage scheme; the priorities of each first candidate action in the above candidate linkage schemes are not completely the same.

[0075] In this step, at least one candidate linkage scheme includes multiple candidate linkage schemes. Each candidate linkage scheme may include multiple first candidate actions for linkage. The priority of each first candidate action can be determined based on factors such as rail transit operation safety, rail transit operation efficiency, and rail transit passenger experience. Among these factors, rail transit operation safety has a higher priority than rail transit operation efficiency, which in turn has a higher priority than rail transit passenger experience. Generally, the priority of each first candidate action is determined based on the highest priority factor it involves. The priorities of the first candidate actions in each candidate linkage scheme are not entirely the same; they may be partially the same or completely different. For example, if a first candidate action involves rail transit operation safety, its priority is determined to be the rail transit operation safety priority, which is the highest priority. Conversely, if a first candidate action only involves rail transit passenger experience, its priority is determined to be the rail transit passenger experience priority, which is the lowest priority.

[0076] Regarding the factors mentioned above, such as rail transit operation safety, rail transit operation efficiency, and rail transit passenger experience, these factors can be pre-quantified with weights, with the sum of the weights of these factors being 1. Then, the weights are assigned sequentially according to the number and priority of each factor. For example, the weight corresponding to rail transit operation safety is 0.5, the weight corresponding to rail transit operation efficiency is 0.3, and the weight corresponding to rail transit passenger experience is 0.2.

[0077] After determining the priority of each first candidate action in each candidate linkage scheme, the first candidate action can be quantified according to the weight of the factor corresponding to its priority to determine its corresponding initial quantization value, that is, the weight of the factor corresponding to the priority of each first candidate action is used as its initial quantization value.

[0078] Step 204: Based on the initial quantization value corresponding to each first candidate action in each candidate linkage scheme, perform quantization processing on each candidate linkage scheme to determine the target quantization value corresponding to each candidate linkage scheme.

[0079] In this process, after obtaining the initial quantization value corresponding to each first candidate action in each candidate linkage scheme, the initial quantization values ​​of the multiple first candidate actions included in each candidate linkage scheme can be summed. The sum obtained can be used as the target quantization value corresponding to the candidate linkage scheme. In this way, the target quantization value corresponding to each candidate linkage scheme can be obtained.

[0080] Step 206: Determine the target linkage scheme from the candidate linkage schemes based on the target quantification value corresponding to each candidate linkage scheme.

[0081] After obtaining the target quantization value corresponding to each candidate linkage scheme, the target quantization values ​​can be sorted directly, for example, from largest to smallest, and the candidate linkage scheme corresponding to the first target quantization value (i.e. the largest target quantization value) can be taken as the target linkage scheme.

[0082] Optionally, after obtaining the target quantification value corresponding to each candidate linkage scheme, the target quantification value of each candidate linkage scheme can be combined with the current state of the rail transit system to determine the target linkage scheme from among the candidate schemes. The current state of the rail transit system may include, for example, whether it is currently operating or whether its energy consumption is too high. By combining the current state of the rail transit system to determine the target linkage scheme, the determined linkage scheme is more in line with the current actual situation of the rail transit system, thereby making the subsequent system control more precise.

[0083] In this embodiment, the quantization value of each first candidate action is determined by the priority of each first candidate action in the candidate linkage scheme, and then the quantization value of each candidate linkage scheme is determined and the target linkage scheme is decided. In this way, by comprehensively analyzing each candidate linkage scheme by the priority of each candidate action in the candidate linkage scheme, the analysis of each candidate linkage scheme can be more comprehensive and reasonable. Thus, the final decision on the target linkage scheme is more reasonable and accurate, which can improve the effect of subsequent linkage scheme execution.

[0084] In some embodiments, after obtaining multiple candidate linkage schemes, in order to facilitate decision-making / screening and make the target linkage scheme more in line with the user's actual needs, optionally, step 108 above, based on at least one candidate linkage scheme for the target object, may include the following steps:

[0085] Display at least one candidate linkage scheme;

[0086] Get the user's selection action input from at least one candidate linkage scheme;

[0087] In response to the selection operation, the candidate linkage scheme corresponding to the selection operation is determined as the target linkage scheme.

[0088] After obtaining multiple candidate linkage schemes, each scheme can be displayed on the equipment interface of the rail transit system for users to view in a timely manner. Users can then select a candidate linkage scheme by inputting a selection operation. This selection can be achieved through selection controls provided on the equipment interface (such as physical buttons or touch buttons) or through voice input. Upon receiving the user's selection operation, the rail transit system can determine the specific candidate linkage scheme chosen and respond to the selection group by using the selected scheme as the target linkage scheme.

[0089] In this embodiment, by displaying multiple candidate linkage schemes on the device interface for the user to select the target linkage scheme, the final target linkage scheme can be more in line with the user's actual needs.

[0090] As the rail transit system continues to operate, new monitoring data will be generated continuously. Similarly, the performance of the equipment in the rail transit system will also change. In order to control the equipment of the rail transit system more accurately based on the target linkage scheme, this embodiment proposes a technical solution that can further adjust the target linkage scheme based on the performance data of the equipment. The following embodiment will explain this process.

[0091] In some embodiments, the above method may further include the following steps:

[0092] Acquire performance data of the target object and / or performance data of the target device within the target object;

[0093] Based on the performance data of the target object and / or the performance data of the target equipment, adjust the control parameters of at least one target linkage action in the target linkage scheme.

[0094] If the target object is a target area, such as a platform, then performance data of the target equipment within the platform after a period of operation can be obtained. Target equipment could be, for example, air conditioning or ventilation systems. Air conditioning performance data could include cooling and heating performance data after a period of operation, while ventilation system performance data could include wind speed and direction data after a period of operation. If the target object is a target vehicle, then performance data of the target vehicle after a period of operation can be obtained, such as door closing time, number of malfunctions, braking distance, and start-up time.

[0095] After obtaining the performance data of the target object after running for a period of time and / or the performance data of the target device in the target object after running for a period of time, if the target linkage scheme includes a target action that adjusts the target device or target vehicle at the target object, the control parameters of the target action can be adjusted through the performance data. For example, the value of the control parameter can be adjusted to better match the actual situation, so that the final adjusted target action can improve the abnormal situation at the target object more quickly and accurately.

[0096] Optionally, after obtaining the performance data of the target object after running for a period of time and / or the performance data of the target device in the target object after running for a period of time, the abnormal conditions in the condition triggering mechanism in the above embodiment can also be adjusted according to these performance data. For example, the original abnormal condition is to trigger the linkage scheme prediction function when the temperature exceeds 30 degrees. Here, the abnormal condition can be adjusted to trigger the linkage scheme prediction function when the temperature exceeds 28 degrees using performance data.

[0097] For example, during the high-temperature period in summer, after accumulating data over a period of time, the rail transit system found that the air conditioning cooling effect of a certain platform was slightly insufficient during peak hours. By automatically adjusting the associated linkage scheme, such as increasing the frequency of ventilation-assisted cooling, adjusting the triggering condition / abnormal condition for ventilation-assisted cooling from the original temperature above 30 degrees to above 28 degrees, and appropriately advancing the time of turning on the air conditioning from the original 10 minutes in advance to 15 minutes in advance, in order to ensure passenger comfort.

[0098] In addition, as the rail transit system continues to operate, new operational data will be constantly collected and fed back to data analysis and association. Here, online learning algorithms (such as online learning algorithms based on stochastic gradient descent, combined with a mini-batch gradient update strategy, which updates the deep learning model every time a certain number (e.g., 1000) of new data are collected) can be used to update and optimize the deep learning model in real time, so that it can adapt to the long-term operational changes of the rail transit system, such as seasonal passenger flow fluctuations and performance changes caused by equipment aging, thereby improving the self-optimization and adaptive adjustment capabilities of the rail transit system.

[0099] In this embodiment, the target linkage scheme is further adjusted by using the performance data of the target object and the performance data of the target equipment at the target object. This allows the final adjusted target linkage scheme to improve the abnormal situation at the target object more quickly and accurately, and further enhances the intelligence of the linkage scheme and the intelligence of controlling the rail transit system.

[0100] The above embodiments mentioned that multiple types of sensors can be set at the target objects of rail transit to collect data. The following embodiments will explain the participation of data collected by multiple types of sensors in the prediction process of linkage schemes.

[0101] In some embodiments, the above method may further include the following steps:

[0102] Acquire second monitoring data collected by the second sensor at the target object;

[0103] Accordingly, step 106 above, which involves inputting the anomaly monitoring data into a preset deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object, may include the following steps:

[0104] The abnormal monitoring data and the second monitoring data are input into a preset deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object; the above candidate linkage scheme includes each first candidate action and second candidate action, and the above second candidate action is used to perform energy-saving actions on energy-related equipment in the target object.

[0105] The second sensor is a sensor of a different type from the first sensor, or it can be a sensor located at the target object. The second sensor can include multiple sensors, which can be sensors of the same or different types, such as, but not limited to, temperature sensors, humidity sensors, light sensors, passenger flow counters, equipment status monitors, smoke sensors, and position sensors.

[0106] The second sensor at the target object can also monitor or collect the corresponding type of data at the target object in real time to obtain the second monitoring data. The second monitoring data is the monitoring data corresponding to the type of the second sensor, and its type can be different from the first monitoring data. For example, the first monitoring data is temperature data, while the second monitoring data can be humidity data, passenger flow count data, etc.

[0107] Specifically, after obtaining the second monitoring data, it can be input together with the abnormal monitoring data obtained from the first monitoring data into a pre-trained deep learning model. The deep learning model then fuses these two types of data and uses the fused data to comprehensively predict the linkage scheme at the target object, ultimately obtaining candidate linkage schemes for the target object. These candidate linkage schemes can include not only the aforementioned first candidate action but also a second candidate action linked to the first candidate action. The first candidate action is primarily used to improve the abnormal situation at the target object, while the second candidate action mainly involves energy-saving actions on energy-related equipment within the target object, such as reducing the brightness of lights in areas with low passenger traffic to save energy.

[0108] For example, taking platform environment control as an example, on a hot summer afternoon, the temperature sensor in the middle of the platform detected that the temperature rose from 28 degrees to 32 degrees within 10 minutes, triggering the associative recommendation mechanism of the rail transit system. Based on real-time data and a trained deep learning model, the data analysis and association layer quickly associates this with turning on the central air conditioning on both sides of the platform and recommends adjusting the fan speed to medium and the airflow direction to blow horizontally towards the middle of the platform. Simultaneously, based on passenger flow data monitored by the current passenger flow counter, it determines that there are fewer passengers at both ends of the platform and automatically reduces the lighting brightness in these two areas to save energy. Upon receiving the recommended solution, the decision-making and execution layer immediately sends instructions to the air conditioning system, ventilation system, and lighting system to achieve intelligent control of the platform environment, providing passengers with a comfortable waiting environment while also achieving energy savings.

[0109] In this embodiment, monitoring data collected by multiple second sensors at the target object and abnormal monitoring data obtained from data collected by the first sensor are input into a deep learning model for linkage scheme prediction processing. This yields actions that include energy saving at the target object and actions that improve abnormal conditions at the target object. This can save energy consumption while improving abnormal conditions at the target object, further enhancing the intelligence of the predicted linkage scheme.

[0110] The above embodiments mention that multiple types of sensors can be set at the target objects of rail transit to collect data. At the same time, the rail transit system is constantly running and will execute the predicted linkage scheme to obtain the execution result. The following embodiments will explain the content of the data collected by multiple types of sensors and the execution result obtained after executing the predicted linkage scheme participating in the subsequent linkage scheme prediction process.

[0111] In some embodiments, the above method may further include the following steps:

[0112] Acquire the execution result obtained after the device at the target object at a historical time performs the first candidate action, and acquire the third monitoring data collected by the third sensor at the target object;

[0113] Accordingly, step 106 above, which involves inputting the anomaly monitoring data into a preset deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object, may include the following steps:

[0114] Anomaly monitoring data, execution results, and third-party monitoring data are input into a pre-set deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object.

[0115] The third sensor is a sensor of a different type from the first and second sensors, and can also be a sensor located at the target object. This third sensor can include multiple sensors, which can be of the same or different types, such as, but not limited to, temperature sensors, humidity sensors, light sensors, passenger flow counters, equipment status monitors, smoke sensors, and position sensors.

[0116] The third sensor at the target object can also monitor or collect the corresponding type of data at the target object in real time to obtain third monitoring data. This third monitoring data is the monitoring data corresponding to the type of the third sensor. It can be different from the first and second monitoring data. For example, the first monitoring data is temperature data, the second monitoring data can be humidity data, passenger flow count data, etc., and the third monitoring data can be the location data of the target area in the target object.

[0117] Specifically, in historical time periods prior to the current time, the data analysis and association layer deployed in the rail transit system can continuously predict candidate linkage schemes and output them to the decision-making and linkage execution layer for linkage scheme decision-making and target linkage scheme execution. After the decision-making and linkage execution layer executes the target linkage scheme for each decision, it will obtain the corresponding execution results. In this way, the execution results of each target linkage scheme executed in historical time periods can be obtained. Then, the above-mentioned third monitoring data, the abnormal monitoring data obtained from the first monitoring data, and the execution results of the target linkage schemes executed in historical time periods are input together into a pre-trained deep learning model. In the deep learning model, these three types of data are fused and processed, and the linkage scheme at the target object is predicted by comprehensively using the fused data, and finally, candidate linkage schemes at the target object are obtained. The candidate linkage schemes may include various first candidate actions. Each first candidate action is mainly used to improve the abnormal situation at the target object. These first candidate actions are determined by comprehensively considering the execution results of historical target linkage schemes. For example, they may be actions after further adjusting their control parameters based on the first candidate actions in step 106 above, thus being more accurate and in line with the needs of the actual scenario.

[0118] For example, taking a platform environment as an example, when a temperature sensor in a certain area detects a temperature change exceeding a set threshold (such as a 3-degree increase within 5 minutes), the rail transit system uses this as a trigger condition. At this time, the data analysis and association layer, based on historical execution results and other data, as well as deep learning models, can not only quickly associate this with turning on the air conditioning in that area, but also recommend simultaneously adjusting the airflow speed and direction of the ventilation system. Specifically, if the area is near the edge of the platform, and past data analysis indicates that ventilation in this area has a significant impact on overall comfort, it is recommended to increase the ventilation volume by 20% and adjust the airflow direction towards the center of the platform to accelerate air circulation, quickly balance the platform temperature, and prevent passengers from experiencing discomfort due to stuffiness.

[0119] In this embodiment, data such as data collected by multiple sensors at the target object, anomaly monitoring data, and execution results obtained after executing the target linkage scheme in historical time are input into a deep learning model for linkage scheme prediction processing. This yields candidate linkage schemes that further adjust the first candidate action, thereby dynamically adjusting the candidate linkage schemes and further improving the intelligence and accuracy of the predicted linkage schemes.

[0120] The above embodiments briefly illustrate the training process of deep learning models. The following embodiments will describe the specific training process of deep learning models.

[0121] Figure 3 This is the third flowchart of the rail transit linkage scheme prediction method provided by the present invention, as shown below. Figure 3 As shown, the training method for the above deep learning model can include the following steps:

[0122] Step 302: Obtain the sample linkage schemes for multiple sample objects and the known operating conditions corresponding to each sample linkage scheme; the known operating conditions include sample anomaly monitoring data at the sample object.

[0123] The sample objects, similar to the target objects mentioned above, can include sample areas and / or sample vehicles. The sample area can be a large area like a platform, or a smaller area within a platform. Known operating conditions refer to known environmental parameters or known vehicle operating parameters at the sample object location. These could include historical platform temperature, humidity, passenger flow data, equipment operation logs, environmental monitoring records, vehicle malfunction counts, and vehicle door closing time. Furthermore, these known environmental parameters or known vehicle operating parameters can be data from abnormal conditions, such as temperature data when temperatures are high, passenger flow data exceeding passenger flow thresholds, malfunction data when the number of vehicle malfunctions exceeds a malfunction threshold, and vehicle door closing time exceeding a set duration.

[0124] Specifically, massive amounts of operational data accumulated over a long period of time in the rail transit system can be collected to obtain multiple known operating conditions over historical periods. The optimal linkage schemes adopted under these known operating conditions can be collected and recorded as sample linkage schemes. Then, each known operating condition is bound to its corresponding sample linkage scheme.

[0125] Step 304: Input each known working condition into the initial deep learning model for linkage scheme prediction processing, and determine the predicted linkage scheme corresponding to the sample object for each known working condition.

[0126] In this step, after obtaining multiple known working conditions and their corresponding sample linkage schemes, the relevant data of each known working condition can be input into the initial deep learning model (i.e., the untrained deep learning model). The initial deep learning model will predict the corresponding linkage scheme based on the data of the known working conditions. Then, the optimal linkage scheme can be selected as the predicted linkage scheme, or all the linkage schemes output by the initial deep learning model can be used as the predicted linkage scheme.

[0127] Step 306: Calculate the loss between each prediction linkage scheme and the corresponding sample linkage scheme.

[0128] In this step, after obtaining the predicted linkage scheme for each known working condition, the predicted linkage scheme and its corresponding sample linkage scheme can be input into the loss function for loss calculation to obtain the loss corresponding to each known working condition. The loss function here can be set according to the actual situation; for example, it can be a loss function that calculates the difference between the predicted linkage scheme for the known working condition and its corresponding sample linkage scheme, such as error, variance, or mean squared error loss functions.

[0129] Step 308: Supervised training of the initial deep learning model is performed based on each loss, and / or unsupervised training of the initial deep learning model is performed based on each known working condition to obtain the deep learning model; the aforementioned supervised training is used to learn the mapping relationship between the known working conditions and the corresponding sample linkage scheme, and the aforementioned unsupervised training is used to mine the inherent correlation relationship between various types of data in the known working conditions.

[0130] In this step, after obtaining the loss between the predicted linkage scheme and the sample linkage scheme for each known working condition, the parameters of the initial deep learning model can be adjusted using these losses to conduct supervised training on the initial deep learning model. For example, given the known linkage scheme for turning on the air conditioning and ventilation when the platform is at high temperature, the initial deep learning model can learn the mapping relationship between the known working conditions and the corresponding sample linkage scheme.

[0131] Meanwhile, after inputting the known operating conditions into the initial deep learning model, the initial deep learning model can also be allowed to perform unsupervised learning. Unsupervised learning focuses on mining hidden patterns and correlations in the data. Specifically, the initial deep learning model can perform in-depth analysis on various types of data included in the known operating conditions to mine the inherent correlations between various types of data in the known operating conditions. For example, it can learn the relationship between passenger flow at different times and platform temperature and humidity, equipment energy consumption, and train departure intervals; learn the logical relationship between equipment failure modes and corresponding countermeasures; and learn / discover the potential relationship between passenger behavior preferences and equipment energy consumption under different passenger flow densities through cluster analysis.

[0132] By repeatedly training the initial deep learning model using a combination of supervised and unsupervised training, a well-trained deep learning model can eventually be obtained.

[0133] Alternatively, the relevant data of the aforementioned known working conditions can be preprocessed before being input into the initial deep learning model. This includes data cleaning (removing outliers and duplicates), normalization (unifying data of different magnitudes to the same range, such as using the Min-Max normalization method), and feature engineering (extracting key features, such as extracting peak and off-peak features from passenger flow data), to improve the training efficiency and accuracy of the model.

[0134] In this embodiment, the initial deep learning model is trained in both supervised and unsupervised manner using multiple known working conditions and their corresponding sample linkage schemes. This combination of supervised and unsupervised training, along with repeated training and optimization of the model, enables the finally trained deep learning model to possess strong associative capabilities. When faced with new real-time data changes, it can quickly and accurately recommend a series of reasonable linkage schemes, thereby improving the intelligence and accuracy of the predicted linkage schemes of the trained deep learning model.

[0135] To facilitate a detailed explanation of the technical solutions of the embodiments of the present invention, several application examples of the embodiments of the present invention are given below.

[0136] Scenario 1: Platform Environment Control

[0137] In a typical urban rail transit station, a three-layer architecture system according to an embodiment of this invention was deployed, comprising a data acquisition layer, a data analysis and association layer, and a decision-making and execution layer. On a hot summer afternoon, a temperature sensor in the middle of the platform detected a temperature increase from 28 degrees Celsius to 32 degrees Celsius within 10 minutes, triggering the rail transit system's associative recommendation function / mechanism. Based on real-time data and a trained deep learning model, the data analysis and association layer quickly associated the activation of the central air conditioning on both sides of the platform, recommending that the fan speed be adjusted to medium and the airflow direction be adjusted to blow horizontally towards the middle of the platform. Simultaneously, considering the current passenger flow, it determined that there were fewer passengers at both ends of the platform and automatically reduced the lighting brightness in these two areas to save energy. Upon receiving the recommended solution, the decision-making and execution layer immediately sent instructions to the air conditioning system, ventilation system, and lighting system to achieve intelligent control of the platform environment, providing passengers with a comfortable waiting environment.

[0138] Scenario 2: Handling Train Equipment Failures

[0139] A train in operation experienced a sudden malfunction warning in its door system, with the door closing time extended to 3 seconds and four consecutive error messages. Upon detecting this, the rail transit system immediately initiated its adaptive recommendation process. Firstly, it automatically reduced the train's speed from 80 km / h to 60 km / h and alerted the driver via an audible and visual alarm through the train's internal communication system. Secondly, it sent a detailed maintenance request to the control center, listing potentially faulty components such as door lock sensors, drive motors, and controllers. This request was sent in standardized fault codes and descriptive text. Thirdly, it coordinated with the backup train dispatch system to pre-arrange a backup train at the next station, ready to replace the faulty train, based on current passenger flow and timetable. Once the faulty train arrived at the next station, passengers quickly and orderly transferred to the backup train, ensuring operational continuity and passenger safety. The process for dispatching backup vehicles is as follows: First, query the location and status information of backup vehicles, select backup vehicles that meet the conditions from the vehicle management database, then determine the optimal route for the backup vehicle to the station where the faulty train is located based on the route planning algorithm (such as Dijkstra's algorithm, which combines real-time passenger flow and running time cost of the line to calculate the optimal path), and finally issue dispatch instructions to the backup vehicle driver to ensure that the overall operation is not significantly affected.

[0140] Scenario 3: Peak Passenger Flow Management

[0141] During weekday morning and evening rush hours, passenger flow monitoring equipment / sensors at the entrance of a certain rail transit station detected a sharp increase in the number of people entering the station within 15 minutes, reaching the set peak threshold. The rail transit system quickly activated its associative recommendation mechanism, in addition to opening all turnstiles as usual, it also considered the real-time passenger flow distribution of surrounding platforms. Analysis revealed that a platform on an adjacent line had lower passenger flow, and it was recommended to adjust the departure interval of trains on that line from 5 minutes to 3 minutes to increase capacity. At the same time, the station's public address system automatically switched to a high-frequency broadcast mode, providing passengers with real-time guidance information such as the most convenient transfer routes and waiting locations. Combined with the dynamic changes in signage, passengers were efficiently diverted to various waiting areas, thereby effectively alleviating platform congestion and improving passenger throughput.

[0142] Based on the descriptions of the above embodiments, the present invention has the following beneficial effects:

[0143] (1) Improve operational efficiency: Through intelligent association recommendation and linkage schemes, it is possible to respond quickly in case of emergencies such as equipment failure and peak passenger flow, rationally allocate resources, reduce train delays, and improve the overall operational efficiency of rail transit. For example, during peak passenger flow, the train departure interval and the number of gate channels opened can be precisely adjusted to effectively shorten the waiting time and entry time for passengers.

[0144] (2) Ensuring system safety: When equipment malfunctions or the environment becomes abnormal (such as a smoke alarm), the system can quickly associate and execute a series of safety measures to reduce the risk of accidents. For example, when the door malfunctions, the system can automatically reduce speed and dispatch a replacement vehicle; when smoke appears, the system can promptly activate the fire sprinkler system and evacuate passengers to ensure the safety of personnel and facilities.

[0145] (3) Optimize passenger experience: Provide a comfortable and convenient travel environment based on real-time passenger needs and environmental changes. For example, automatically adjust ventilation, lighting and other facilities according to platform temperature and passenger flow to create a pleasant waiting environment for passengers; and provide accurate route guidance during transfers to improve passenger experience and satisfaction.

[0146] In summary, the technical solution of this invention, through its innovative three-layer architecture design, powerful associative recommendation function, and adaptive optimization mechanism, brings a brand-new intelligent solution to the rail transit field, effectively improving the operational efficiency of rail transit, ensuring system operation safety, and optimizing passenger experience.

[0147] The following describes the rail transit linkage scheme prediction device provided by the present invention. The rail transit linkage scheme prediction device described below and the rail transit linkage scheme prediction method described above can be referred to and correspond to each other.

[0148] Figure 4This is a schematic diagram of the predictive device for rail transit linkage schemes provided by the present invention. See also: Figure 4 As shown, the device may include:

[0149] The first monitoring data acquisition module 410 is used to acquire first monitoring data collected by a first sensor at a target object in rail transit; the target object includes a target area and / or a target vehicle.

[0150] The anomaly detection module 420 is used to perform anomaly detection on the first monitoring data. If an anomaly is detected in the first monitoring data, the abnormal monitoring data containing the anomaly in the first monitoring data is obtained.

[0151] The linkage scheme prediction module 430 is used to input the anomaly monitoring data into a preset deep learning model for linkage scheme prediction processing, and to determine at least one candidate linkage scheme corresponding to the target object; each candidate linkage scheme includes multiple first candidate actions for linkage, and each first candidate action is an adjustment action to adjust the anomaly situation of the target object. The aforementioned deep learning model is obtained in advance by supervised training and / or by unsupervised training based on multiple sample linkage schemes and the known working conditions corresponding to each sample linkage scheme.

[0152] The target linkage scheme determination module 440 is used to determine the target linkage scheme corresponding to the target object based on at least one candidate linkage scheme of the target object; the target linkage scheme includes multiple linkage target actions and control parameters for each target action.

[0153] In some embodiments, the target linkage scheme determination module 440 is specifically used to determine the initial quantization value corresponding to each first candidate action in each candidate linkage scheme according to the priority corresponding to each first candidate action in each candidate linkage scheme; the priorities corresponding to each first candidate action in the above candidate linkage scheme are not completely the same;

[0154] Based on the initial quantization value corresponding to each first candidate action in each candidate linkage scheme, each candidate linkage scheme is quantized to determine the target quantization value corresponding to each candidate linkage scheme.

[0155] Based on the target quantification value corresponding to each candidate linkage scheme, the target linkage scheme is determined from the candidate linkage schemes.

[0156] In some embodiments, the target linkage scheme determination module 440 is specifically used to display at least one candidate linkage scheme; obtain the selection operation input by the user in at least one candidate linkage scheme; and, in response to the selection operation, determine the candidate linkage scheme corresponding to the selection operation as the target linkage scheme.

[0157] In some embodiments, the above-described apparatus further includes:

[0158] The performance data acquisition module is used to acquire the performance data of the target object and / or the performance data of the target device in the target object.

[0159] The scheme adjustment module is used to adjust the control parameters of at least one target linkage action in the target linkage scheme based on the performance data of the target object and / or the performance data of the target equipment.

[0160] In some embodiments, the apparatus further includes a training module for training a deep learning model. Specifically, the training module acquires sample linkage schemes for multiple sample objects and known operating conditions corresponding to each sample linkage scheme. The known operating conditions include sample anomaly monitoring data at the sample objects. Each known operating condition is input into an initial deep learning model for linkage scheme prediction processing to determine the predicted linkage scheme corresponding to each known operating condition for the corresponding sample object. The loss between each predicted linkage scheme and the corresponding sample linkage scheme is calculated. The initial deep learning model is then subjected to supervised training based on each loss, and / or unsupervised training is performed based on each known operating condition to obtain the deep learning model. The supervised training is used to learn the mapping relationship between known operating conditions and corresponding sample linkage schemes, while the unsupervised training is used to mine the inherent correlation relationships between various types of data in the known operating conditions.

[0161] In some embodiments, the above-described apparatus further includes:

[0162] The second monitoring data acquisition module is used to acquire the second monitoring data collected by the second sensor at the target object.

[0163] The aforementioned linkage scheme prediction module 430 is specifically used to input abnormal monitoring data and second monitoring data into a preset deep learning model for linkage scheme prediction processing, and to determine at least one candidate linkage scheme corresponding to the target object; the aforementioned candidate linkage scheme includes each first candidate action and second candidate action, and the aforementioned second candidate action is used to perform energy-saving actions on energy-related equipment in the target object.

[0164] In some embodiments, the above-described apparatus further includes:

[0165] The third monitoring data acquisition module is used to acquire the execution result obtained by the device at the target object at a historical time after executing the first candidate action, and to acquire the third monitoring data collected by the third sensor at the target object.

[0166] The aforementioned linkage scheme prediction module 430 is specifically used to input anomaly monitoring data, execution results, and third monitoring data into a preset deep learning model for linkage scheme prediction processing, and to determine at least one candidate linkage scheme corresponding to the target object.

[0167] It should be noted that the apparatus provided in this embodiment of the invention can implement all the method steps implemented in the above method embodiment and can achieve the same technical effect. Therefore, the parts and beneficial effects that are the same as those in the method embodiment will not be described in detail here.

[0168] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5 As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530, and a communications bus 540, wherein the processor 510, the communications interface 520, and the memory 530 communicate with each other through the communications bus 540. The processor 510 can call logic instructions in the memory 530 to execute a rail transit linkage scheme prediction method. This method includes: acquiring first monitoring data collected by a first sensor at a target object in the rail transit system; the target object includes a target area and / or a target vehicle; performing anomaly detection on the first monitoring data; if anomalies are detected in the first monitoring data, acquiring the abnormal monitoring data containing the anomalies; inputting the abnormal monitoring data into a preset deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object; each candidate linkage scheme includes multiple first candidate actions for linkage, each first candidate action corresponding to an adjustment action to address anomalies in the target object; the deep learning model is pre-trained under supervised conditions and / or unsupervised conditions based on multiple sample linkage schemes and the known operating conditions corresponding to each sample linkage scheme; and determining a target linkage scheme corresponding to the target object based on at least one candidate linkage scheme; the target linkage scheme includes multiple target actions for linkage and control parameters for each target action.

[0169] Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0170] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the rail transit linkage scheme prediction method provided by the above methods. The method includes: acquiring first monitoring data collected by a first sensor at a target object in the rail transit; the target object includes a target area and / or a target vehicle; performing anomaly detection on the first monitoring data, and if anomalies are detected in the first monitoring data, acquiring the abnormal monitoring data in the first monitoring data; inputting the abnormal monitoring data into a preset deep learning model for linkage scheme prediction processing, and determining at least one candidate linkage scheme corresponding to the target object; each candidate linkage scheme includes multiple first candidate actions for linkage, and each first candidate action corresponds to an adjustment action to adjust for the abnormal situation existing in the target object, wherein the deep learning model is obtained in advance through supervised training and / or unsupervised training based on multiple sample linkage schemes and the known working conditions corresponding to each sample linkage scheme; and determining the target linkage scheme corresponding to the target object based on at least one candidate linkage scheme of the target object; the target linkage scheme includes multiple target actions for linkage and control parameters for each target action.

[0171] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the rail transit linkage scheme prediction method provided by the above methods. The method includes: acquiring first monitoring data collected by a first sensor at a target object in the rail transit; the target object includes a target area and / or a target vehicle; performing anomaly detection on the first monitoring data, and if anomalies are detected in the first monitoring data, acquiring the abnormal monitoring data in the first monitoring data; inputting the abnormal monitoring data into a preset deep learning model for linkage scheme prediction processing, and determining at least one candidate linkage scheme corresponding to the target object; each candidate linkage scheme includes multiple first candidate actions for linkage, and each first candidate action corresponds to an adjustment action to adjust for the abnormal situation existing in the target object, wherein the deep learning model is obtained in advance through supervised training and / or unsupervised training based on multiple sample linkage schemes and the known working conditions corresponding to each sample linkage scheme; and determining a target linkage scheme corresponding to the target object based on at least one candidate linkage scheme for the target object; the target linkage scheme includes multiple target actions for linkage and control parameters for each target action.

[0172] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0173] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the linkage scheme of rail transit, characterized in that, include: Acquire the first monitoring data collected by the first sensor at the target object in the rail transit system; The target objects include target areas and / or target vehicles; Anomaly detection is performed on the first monitoring data. If anomalies are detected in the first monitoring data, the abnormal monitoring data containing the anomalies in the first monitoring data is obtained. The anomaly monitoring data is input into a preset deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object; each candidate linkage scheme includes multiple first candidate actions for linkage, and each first candidate action corresponds to an adjustment action to adjust the anomaly situation existing in the target object. The deep learning model is obtained in advance by supervised training and / or unsupervised training based on multiple sample linkage schemes and the known working conditions corresponding to each sample linkage scheme. Based on at least one candidate linkage scheme for the target object, determine the target linkage scheme corresponding to the target object; The target linkage scheme includes multiple linked target actions and control parameters for each target action; The system acquires the execution result obtained after the device at the target object performs the first candidate action at a historical time, and acquires the third monitoring data collected by the third sensor at the target object; the type of the third monitoring data is different from the type of the first monitoring data, and the third monitoring data includes the location data of the target area in the target object; The step of inputting the anomaly monitoring data into a preset deep learning model for linkage scheme prediction processing, and determining at least one candidate linkage scheme corresponding to the target object, includes: The anomaly monitoring data, the execution result, and the third monitoring data are input into a preset deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object.

2. The method for predicting the linkage scheme of rail transit according to claim 1, characterized in that, The step of determining the target linkage scheme corresponding to the target object based on at least one candidate linkage scheme of the target object includes: Based on the priority of each first candidate action in each candidate linkage scheme, an initial quantization value is determined for each first candidate action in each candidate linkage scheme; the priorities of each first candidate action in each candidate linkage scheme are not completely the same; Based on the initial quantization value corresponding to each first candidate action in each candidate linkage scheme, each candidate linkage scheme is quantized to determine the target quantization value corresponding to each candidate linkage scheme. Based on the target quantification value corresponding to each candidate linkage scheme, the target linkage scheme is determined from the candidate linkage schemes.

3. The method for predicting the linkage scheme of rail transit according to claim 1, characterized in that, The step of determining the target linkage scheme corresponding to the target object based on at least one candidate linkage scheme of the target object includes: Display the at least one candidate linkage scheme; Obtain the selection operation input by the user in the at least one candidate linkage scheme; In response to the selection operation, the candidate linkage scheme corresponding to the selection operation is determined as the target linkage scheme.

4. The method for predicting rail transit linkage schemes according to any one of claims 1 to 3, characterized in that, The method further includes: Acquire the performance data of the target object and / or the performance data of the target device within the target object; Based on the performance data of the target object and / or the performance data of the target device, the control parameters of at least one target linkage action in the target linkage scheme are adjusted.

5. The method for predicting the linkage scheme of rail transit according to claim 1, characterized in that, The training methods for the deep learning model include: Acquire the sample linkage schemes for multiple sample objects and the known operating conditions corresponding to each sample linkage scheme; the known operating conditions include sample anomaly monitoring data at the sample object. Each of the known working conditions is input into the initial deep learning model for linkage scheme prediction processing, and the predicted linkage scheme corresponding to the sample object for each of the known working conditions is determined. Calculate the loss between each of the predicted linkage schemes and the corresponding sample linkage schemes; The initial deep learning model is trained in a supervised manner according to each of the aforementioned losses, and / or trained in an unsupervised manner according to each of the aforementioned known working conditions to obtain the deep learning model; the supervised training is used to learn the mapping relationship between the known working conditions and the corresponding sample linkage scheme, and the unsupervised training is used to mine the inherent correlation relationship between various types of data in the known working conditions.

6. The method for predicting rail transit linkage schemes according to any one of claims 1 to 3, characterized in that, The method further includes: Acquire the second monitoring data collected by the second sensor at the target object; The step of inputting the anomaly monitoring data into a preset deep learning model for linkage scheme prediction processing, and determining at least one candidate linkage scheme corresponding to the target object, includes: The abnormal monitoring data and the second monitoring data are input into a preset deep learning model for linkage scheme prediction processing to determine at least one candidate linkage scheme corresponding to the target object; the candidate linkage scheme includes each of the first candidate action and the second candidate action, and the second candidate action is used to perform energy-saving actions on energy-related equipment in the target object.

7. A predictive device for linkage schemes in rail transit, characterized in that, include: The first monitoring data acquisition module is used to acquire the first monitoring data collected by the first sensor at the target object in the rail transit; The target objects include target areas and / or target vehicles; An anomaly detection module is used to perform anomaly detection on the first monitoring data. If an anomaly is detected in the first monitoring data, the abnormal monitoring data containing the anomaly in the first monitoring data is obtained. The linkage scheme prediction module is used to input the anomaly monitoring data into a preset deep learning model for linkage scheme prediction processing, and to determine at least one candidate linkage scheme corresponding to the target object; each candidate linkage scheme includes multiple first candidate actions for linkage, and each first candidate action corresponds to an adjustment action to adjust the anomaly situation existing in the target object. The deep learning model is obtained in advance by supervised training and / or unsupervised training based on multiple sample linkage schemes and the known working conditions corresponding to each sample linkage scheme. The target linkage scheme determination module is used to determine the target linkage scheme corresponding to the target object based on at least one candidate linkage scheme of the target object; The target linkage scheme includes multiple linked target actions and control parameters for each target action; The third monitoring data acquisition module is used to acquire the execution result obtained after the device at the target object performs the first candidate action at a historical time, and to acquire the third monitoring data collected by the third sensor at the target object; the type of the third monitoring data is different from the type of the first monitoring data, and the third monitoring data includes the location data of the target area in the target object; The linkage scheme prediction module is specifically used to input the anomaly monitoring data, the execution result and the third monitoring data into a preset deep learning model to perform linkage scheme prediction processing, and determine at least one candidate linkage scheme corresponding to the target object.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the rail transit linkage scheme prediction method as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the rail transit linkage scheme prediction method as described in any one of claims 1 to 6.

Images