A locomotive status data transmission method and related apparatus
By filtering and transmitting abnormal status data at the locomotive end and combining it with ground model data recovery, the problems of resource consumption and latency in locomotive data transmission were solved, achieving efficient and reliable data transmission and monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CRRC QISHUYAN CO LTD
- Filing Date
- 2026-04-30
- Publication Date
- 2026-07-31
AI Technical Summary
The full transmission of data generated by sensors in modern intelligent locomotives requires a large amount of wireless spectrum resources, which may lead to data congestion, loss, or transmission delay, especially in areas with dense train traffic or poor signal.
An abnormal state data is filtered using a time-series prediction model on the locomotive side. Only abnormal data is transmitted to the ground data center. The ground data center deploys the same model to recover the untransmitted data. The transmission process is optimized by combining confidence and caching strategies.
This significantly reduces the amount of data transmission, minimizes data congestion, loss, and transmission delays, and ensures that the ground data center obtains complete, continuous, and highly reliable locomotive status data to meet the needs of status monitoring and fault analysis.
Smart Images
Figure CN122496919A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of data transmission technology, and in particular to a method and related apparatus for transmitting locomotive status data. Background Technology
[0002] Modern intelligent locomotives are equipped with a large number of sensors (such as vibration, temperature, voltage, current, pressure, etc.) to monitor the operating status of key locomotive components (such as bogies, traction motors, braking systems, pantographs, etc.) in real time. These sensors generate massive amounts of data in real time, which need to be transmitted to ground data centers for analysis and early warning via wireless networks (such as 4G / 5G, satellite, trackside WLAN).
[0003] However, data transmission bandwidth is limited. Transmitting all data generated by all sensors requires a large amount of valuable wireless spectrum resources, which may lead to data congestion, loss or transmission delay, especially in areas with poor signal, such as dense train traffic or tunnels. Summary of the Invention
[0004] In view of the above problems, this application provides a locomotive status data transmission method and related apparatus to achieve the purpose of selectively transmitting abnormal status data in real time, while not transmitting normal status data in real time. The specific solution is as follows:
[0005] The first aspect of this application provides a method for transmitting locomotive status data, including:
[0006] The system acquires real-time status data collected by locomotive sensors and a sequence of normal status data collected prior to the real-time status data. The normal status data in the normal status data sequence refers to the operating status data collected when the locomotive is in normal operating condition.
[0007] The normal state data sequence is input into a pre-trained first time-series prediction model to obtain the state prediction value corresponding to the real-time state data. The state prediction value represents the running state data under the simulated normal operating state at the time of the real-time state data acquisition.
[0008] Based on the predicted state value, it is determined whether the real-time state data is abnormal state data. Abnormal state data refers to the operating state data collected when the locomotive is in an abnormal operating state.
[0009] If so, the real-time status data is transmitted to the ground data center; otherwise, the real-time status data is not transmitted. The ground data center is equipped with a second time-series prediction model with the same structure and parameters as the first time-series prediction model. The second time-series prediction model is used to recover the untransmitted real-time status data based on the normal status data pre-transmitted from the locomotive-end equipment.
[0010] In one possible implementation, the second time-series prediction model is trained based on normal state data samples received by the ground data center, and the parameters of the first time-series prediction model are updated along with the second time-series prediction model.
[0011] In one possible implementation, determining whether the real-time state data is abnormal state data based on the state prediction value includes:
[0012] Calculate the deviation between the predicted state value and the real-time state data;
[0013] If the deviation value is greater than the preset deviation threshold, the real-time status data is determined to be abnormal status data; otherwise, the real-time status data is determined not to be abnormal status data.
[0014] In one possible implementation, the first time-series prediction model outputs the confidence level of the state prediction value along with the state prediction value, wherein the confidence level of the state prediction value is used as the confidence level of the real-time state data corresponding to the state prediction value.
[0015] The locomotive status data transmission method further includes:
[0016] During the transmission of each of the real-time status data, if the confidence level corresponding to any of the real-time status data is less than the first threshold, the full data transmission mode is activated so that all subsequently collected real-time status data are transmitted to the ground data center until the preset full data transmission mode closing condition is met.
[0017] One possible implementation also includes:
[0018] During the transmission of each of the aforementioned real-time status data, the timer is reset every time a piece of real-time status data is transmitted to the ground data center;
[0019] If no real-time status data is transmitted within the target time period, a heartbeat packet indicating that the system is normal is sent to the ground data center when the target time period is reached, and the timer is restarted.
[0020] The target timing duration is initialized to a first duration value;
[0021] During transmission, if the confidence level of the real-time status data transmitted to the ground data center is within the intermediate confidence level range, the target timing duration is shortened. The intermediate confidence level range consists of the first threshold and a second threshold greater than the first threshold. If the confidence level of all untransmitted real-time status data within the target timing duration is greater than the second threshold, the target timing duration is restored to the first duration value.
[0022] One possible implementation also includes:
[0023] Before transmitting the real-time status data to the ground data center, the real-time status data is cached in the target buffer;
[0024] And / or, before transmitting all subsequently collected real-time status data to the ground data center, cache all subsequently collected real-time status data in the target buffer.
[0025] One possible implementation also includes:
[0026] Delete real-time status data in the target buffer that has been cached for a preset duration threshold.
[0027] A second aspect of this application provides a computer program product including computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the locomotive status data transmission method described in the first aspect or any implementation thereof.
[0028] A third aspect of this application provides an electronic device, comprising at least one processor and a memory connected to the processor, wherein:
[0029] The memory is used to store computer programs;
[0030] The processor is used to execute the computer program so that the electronic device can implement the locomotive status data transmission method of the first aspect or any implementation thereof.
[0031] A fourth aspect of this application provides a computer storage medium carrying one or more computer programs, which, when executed by an electronic device, enable the electronic device to implement the locomotive status data transmission method described in the first aspect or any implementation thereof.
[0032] By employing the above technical solution, the locomotive status data transmission method provided in this application acquires real-time status data collected by locomotive sensors and a sequence of normal status data collected prior to the real-time status data. The normal status data sequence is then input into a pre-trained first time-series prediction model to obtain the predicted status value corresponding to the real-time status data. Since the operating status data collected by the locomotive sensors is time-series data, and its fluctuations are relatively gentle under normal operating conditions, the first time-series prediction model, based on the normal status data sequence, can predict the normal status data at the same moment as the real-time status data, i.e., the predicted status value. Based on the predicted status value, it can be determined whether the real-time status data is abnormal. If so, the real-time status data is transmitted to the ground data center; otherwise, it is not transmitted. Therefore, this application can transmit real-time status data to the ground data center only when it is abnormal. Since the locomotive is rarely in an abnormal operating state, the number of abnormal status data is far less than the number of normal status data, thereby significantly reducing the amount of data transmitted and reducing the possibility of data blockage, loss, and transmission delay.
[0033] Meanwhile, this application deploys a second time-series prediction model in the ground data center with the same structure and parameters as the first time-series prediction model. This enables the ground data center to use the second time-series prediction model to recover the untransmitted real-time status data with high fidelity based on the normal status data pre-transmitted by the locomotive-end equipment. Combined with the received real-time status data, the complete original data can be accurately restored. Under the premise of significantly reducing the amount of data transmitted, the ground data center can still obtain a complete, continuous, and highly reliable locomotive status data sequence, which meets the needs of status monitoring, fault analysis, and trend prediction. Attached Figure Description
[0034] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.
[0035] Figure 1 A flowchart illustrating a locomotive status data transmission method provided in this application;
[0036] Figure 2 A schematic diagram of the structure of a locomotive status data transmission device provided in this application;
[0037] Figure 3 This is a schematic diagram of the structure of an electronic device provided in this application. Detailed Implementation
[0038] The embodiments of this application are described below with reference to the accompanying drawings. The terminology used in the implementation section of this application is for explaining specific embodiments only and is not intended to limit the scope of this application.
[0039] The embodiments of this application will now be described with reference to the accompanying drawings. Those skilled in the art will recognize that, with technological advancements and the emergence of new scenarios, the technical solutions provided in the embodiments of this application are equally applicable to similar technical problems.
[0040] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such terms are interchangeable where appropriate; this is merely a way of distinguishing objects with the same attributes in the embodiments of this application. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, so that a process, method, system, product, or apparatus that comprises a series of elements is not necessarily limited to those elements, but may include other elements not explicitly listed or inherent to those processes, methods, products, or apparatuses.
[0041] To enable those skilled in the art to better understand this application, the locomotive status data transmission method of this application embodiment will be described in detail below with reference to the accompanying drawings.
[0042] This application provides a locomotive status data transmission method, which can be applied to locomotive-side equipment. Preferably, the locomotive-side equipment can be a locomotive-mounted edge computing device.
[0043] For example, in one possible implementation, the locomotive-mounted edge computing device can be configured with a dedicated artificial intelligence acceleration module, on which a first time-series prediction model is deployed. When the locomotive-mounted edge computing device collects real-time status data from the locomotive's sensors through the onboard edge computing gateway, it can detect anomalies in the real-time status data based on the first time-series prediction model, and then decide whether to transmit the real-time status data.
[0044] Optionally, both the AI acceleration module and the vehicle edge computing gateway can use low-power, high-performance chips, such as ARM Cortex-A series CPUs (Central Processing Units), NPUs (Neural Processing Units), and FPGAs (Field Programmable Gate Arrays). Hardware-level optimization code can be written on the chip (e.g., using the ARM NEON instruction set and NPU-specific SDKs (Software Development Kits)) to implement the above-mentioned anomaly detection and data transmission based on the optimized code, maximizing the computing efficiency of the edge hardware.
[0045] Optionally, power management strategies can be designed to reduce the model inference frequency to save energy when the locomotive is in hibernation or running at low speed.
[0046] For a more detailed implementation process, please refer to Figure 1 , Figure 1 This is a flowchart illustrating a locomotive status data transmission method provided in an embodiment of this application, as follows: Figure 1 As shown, the locomotive status data transmission method may include:
[0047] Step S101: Obtain the real-time status data collected by the locomotive sensors and the normal status data sequence collected before the real-time status data.
[0048] In this embodiment, the locomotive sensors continuously sample the locomotive's operating status data at a preset fixed period, forming ordered time-series data. Here, each sampled data point collected by the locomotive sensors in real time (excluding the first sampled data point) is referred to as real-time status data.
[0049] Research has shown that when a locomotive is in normal operating condition (i.e., stable operating condition), the fluctuation of locomotive status data is small. Therefore, if a large fluctuation is found in the real-time status data collected at a certain sampling time, it can be determined that there is an anomaly in the real-time status data.
[0050] In order to accurately identify whether there are any anomalies in each real-time status data, this embodiment can obtain a sequence of normal status data collected before the real-time status data, so as to realize the anomaly identification of the real-time status data based on the normal status data sequence in the following text.
[0051] Here, the normal state data sequence consists of at least one normal state data, which refers to the operating state data collected when the locomotive is in normal operating condition.
[0052] The normal state data sequence is preferably a sequence that is as close as possible to the real-time state data in terms of acquisition time. For example, if the acquisition time of the real-time state data is t, then the acquisition time period of the normal state data sequence is selected as [t-10, t-1].
[0053] Step S102: Input the normal state data sequence into the pre-trained first time-series prediction model to obtain the state prediction value corresponding to the real-time state data.
[0054] As mentioned earlier, when the locomotive is in normal operation, the fluctuation of the locomotive status data is relatively small. In order to capture this relatively stable data fluctuation for predicting the normal status data at the same moment as the real-time status data, this embodiment can use normal status data samples to pre-train a time series prediction model, so that the time series prediction model can learn the variation pattern of sensor sampling data under normal operation, that is, learn the variation pattern of normal status data.
[0055] In this embodiment, the time-series prediction model deployed on the locomotive-side equipment is defined as the first time-series prediction model. Then, by inputting the normal state data sequence into the first time-series prediction model, the state prediction value corresponding to the real-time state data can be obtained. Here, the state prediction value represents the operating state data under simulated normal operating conditions at the time of real-time state data acquisition.
[0056] Considering the poor flexibility of model deployment and updates, deploying complex algorithms on vehicle-side equipment places extremely high demands on the computing power and power consumption of the onboard hardware, and model updates are difficult. Therefore, to optimize the first time-series prediction model, this embodiment can optionally use a lightweight model architecture with low computational cost and few parameters, such as TinyLSTM (Tiny Long Short-Term Memory), miniature Transformers (such as Mobile-Transformer), deep separable temporal convolutional networks, and GRU (Gated Recurrent Unit). Furthermore, techniques such as model pruning (i.e., removing redundant connections), quantization (such as INT8 (8-bit Integer Precision) quantization), and knowledge distillation can be employed to further compress the model size and computational requirements.
[0057] As described in the background section, locomotives have a large number of sensors, making it difficult to learn the variation patterns of all sensor sampling data under normal operating conditions using a single model. Therefore, alternatively, a first time-series prediction model can be trained for each locomotive sensor. In a more preferred implementation, locomotive operation fault types can be statistically analyzed, and a first time-series prediction model can be trained for each fault type; that is, a first time-series prediction model can be trained using data collected from multiple locomotive sensors corresponding to each fault type.
[0058] Therefore, the process of “inputting the normal state data sequence into the pre-trained first time-series prediction model” can specifically include: determining the first time-series prediction model corresponding to the locomotive sensor corresponding to the normal state data sequence, and inputting the normal state data sequence into the determined first time-series prediction model.
[0059] Of course, there are other implementations, such as designing a model that shares a feature extractor among multiple sensors to reduce the overall number of parameters.
[0060] Step S103: Based on the state prediction value, determine whether the real-time state data is abnormal state data.
[0061] In this embodiment, the predicted state value is the normal state data at the same moment as the real-time state data. If the real-time state data deviates too much from the predicted state value, it indicates that the real-time state data is abnormal, i.e., the real-time state data is abnormal state data. Here, abnormal state data refers to the operating state data collected when the locomotive is in an abnormal operating state.
[0062] Optionally, this embodiment can calculate the deviation between the predicted state value and the real-time state data. If the deviation value is greater than a preset deviation threshold, it indicates that the current real-time state data contains new information that the model failed to predict accurately. In this case, the real-time state data is likely to be abnormal or have a sudden change in state, and the real-time state data is determined to be abnormal state data. Conversely, if the deviation value is less than or equal to the deviation threshold, it indicates that the real-time state data can be accurately predicted by the model, and the real-time state data is determined not to be abnormal state data. Here, the deviation threshold can be a static threshold or a dynamic threshold, and no specific limitation is made here.
[0063] Optionally, the deviation value can be one of the absolute error value, relative error value, or mean square error value. Of course, the deviation value can also be other values, which are not specifically limited here.
[0064] Considering that single-point judgment may have noise errors, in order to avoid misjudgment, the real-time state data and multiple consecutive real-time state data in the preceding direction can be compared with their respective state prediction values. If the deviation values corresponding to the real-time state data and the multiple consecutive real-time state data in the preceding direction are all greater than the deviation threshold, then the real-time state data is determined to be abnormal state data.
[0065] Of course, there are other ways to implement it, which will not be specifically limited here.
[0066] Step S104: If yes, transmit the real-time status data to the ground data center; otherwise, do not transmit the real-time status data. The ground data center is equipped with a second time-series prediction model with the same structure and parameters as the first time-series prediction model. The second time-series prediction model is used to recover the untransmitted real-time status data based on the normal status data pre-transmitted from the locomotive-end equipment.
[0067] In this embodiment, transmitting all normally changing state data would result in a large amount of redundant data, which has limited value but consumes bandwidth. To avoid transmitting redundant data as much as possible, this embodiment only transmits real-time state data to the ground data center when it is determined that the real-time state data is in an abnormal state. When it is determined that the real-time state data is in a normal state, it is not transmitted to the ground data center. This can save a lot of bandwidth resources, reduce the possibility of data congestion, loss, and transmission delay, and improve transmission efficiency.
[0068] It should be noted that the real-time status data comes with a collection timestamp attribute, meaning that the collection timestamp (i.e., the collection time) is transmitted along with the real-time status data itself.
[0069] Preferably, in addition to transmitting real-time status data, metadata such as the sensor identifier where the real-time status data is located can also be packaged together and transmitted to the ground data center via wireless network.
[0070] In order for the ground data center to obtain the full data sequence even when only abnormal state data is transmitted, this embodiment can deploy a second time series prediction model with the same structure and parameters as the first time series prediction model in the ground data center. This allows the ground data center to use the second time series prediction model to recover the untransmitted real-time state data based on the normal state data pre-transmitted by the locomotive-end equipment.
[0071] For example, in one possible implementation, after establishing wireless communication (such as 4G, 5G, satellite, or trackside WLAN) between the locomotive-side equipment and the ground data center, and once the locomotive is running stably, the locomotive-side equipment transmits the first few sampled data points as normal state data to the ground data center. Then, following the steps described above, only the abnormal state data can be transmitted. This first few sampled data points are then input into a second time-series prediction model to obtain the predicted state values corresponding to the untransmitted real-time state data. Combined with the received abnormal state data, the complete original data sequence can be reconstructed with high fidelity. This process achieves lossy but high-fidelity compression, with bandwidth savings of 70%-95% (depending on system stability).
[0072] As mentioned earlier, training the model on the locomotive-side equipment has high hardware requirements. Therefore, alternatively, the second time-series prediction model can be trained in a ground data center, and the trained model parameters can be safely and efficiently distributed to the locomotive-side equipment via a differential update method through a wireless network, so that the parameters of the first time-series prediction model are updated with the second time-series prediction model.
[0073] The second time-series prediction model is trained based on normal state data samples received by the ground data center. That is, after the ground data center receives the normal state data samples transmitted from the locomotive-end equipment, it splits the normal state data samples into sample labels corresponding to the input data samples, and then trains the second time-series prediction model.
[0074] For example, if the normal state data samples transmitted from the locomotive-end equipment include normal state data at times t-10, t-9, ..., t, then the normal state data from times t-10 to t-1 can be used as input data samples, and the normal state data at time t can be used as sample labels to train the second time series prediction model.
[0075] The locomotive status data transmission method provided in this application acquires real-time status data collected by locomotive sensors and a sequence of normal status data collected prior to the real-time status data. The normal status data sequence is then input into a pre-trained first time-series prediction model to obtain the predicted status value corresponding to the real-time status data. Since the operating status data collected by the locomotive sensors is time-series data, and the fluctuations during normal operation are relatively smooth, the first time-series prediction model predicts the normal status data at the same moment as the real-time status data, i.e., the predicted status value. Based on the predicted status value, it can be determined whether the real-time status data is abnormal. If so, the real-time status data is transmitted to the ground data center; otherwise, it is not transmitted. Therefore, this application can transmit real-time status data to the ground data center only when it is abnormal. Since the locomotive is rarely in an abnormal operating state, the number of abnormal status data is far less than the number of normal status data, thus significantly reducing the amount of data transmitted and lowering the possibility of data congestion, loss, and transmission delay.
[0076] Meanwhile, this application deploys a second time-series prediction model in the ground data center with the same structure and parameters as the first time-series prediction model. This enables the ground data center to use the second time-series prediction model to recover the untransmitted real-time status data with high fidelity based on the normal status data pre-transmitted by the locomotive-end equipment. Combined with the received real-time status data, the complete original data can be accurately restored. Under the premise of significantly reducing the amount of data transmitted, the ground data center can still obtain a complete, continuous, and highly reliable locomotive status data sequence, which meets the needs of status monitoring, fault analysis, and trend prediction.
[0077] Through the above methods, this application realizes intelligent data filtering and adaptive transmission at the locomotive end, effectively reducing wireless bandwidth occupation and improving the real-time transmission of abnormal data. At the same time, it relies on the vehicle-to-ground model synchronization mechanism to ensure data integrity, so that the system achieves the optimal balance between bandwidth efficiency, transmission reliability and monitoring accuracy. It is especially suitable for rail transit scenarios with complex locomotive operating environments (such as limited power consumption, space and heat dissipation conditions), fluctuating wireless signals and limited bandwidth resources.
[0078] In one possible implementation, the first time-series prediction model outputs both the predicted state value and a confidence level for that predicted state value. For ease of description below, the confidence level of the predicted state value is taken as the confidence level of the real-time state data corresponding to the predicted state value. The confidence level can be obtained through internal model uncertainty estimation, ensemble learning, or historical statistics of prediction errors; this application does not impose specific limitations on this.
[0079] In this embodiment, the higher the confidence level of the real-time status data, the more accurate the anomaly determination result in step S103, and the more reliable the transmission method prioritizing abnormal status data in step S104. Conversely, if the confidence level of the real-time status data is very low, the transmission method in step S104 may be unreliable, meaning that abnormal status data may be misjudged as normal status data and not transmitted to the ground data center. To avoid this problem, optionally, in this embodiment, during the transmission of each real-time status data, if the confidence level of any real-time status data is less than a preset first threshold (i.e., a low confidence level), then a full data transmission mode is activated, so that all subsequently collected real-time status data are transmitted to the ground data center until the preset full data transmission mode shutdown condition is met.
[0080] Optionally, there are several ways to achieve the conditions for closing the full data transmission mode. The following methods are provided, but are not limited to:
[0081] The first method: Reaching the preset time. That is, during the transmission process, if the confidence level of a certain real-time status data is less than the first threshold, it may indicate an unknown operating condition or gigabit failure. In this case, the full data transmission mode is triggered, and all subsequent real-time status data is transmitted to the ground data center until the preset time is reached. Then, it switches back to the abnormal status data priority transmission mode mentioned above (hereinafter referred to as abnormal data transmission mode).
[0082] The second method involves collecting multiple consecutive real-time status data points (the number of data points can be determined by a preset duration or a preset data point quantity) whose confidence levels are all greater than or equal to the first threshold. That is, after triggering the full data transmission mode, the system monitors whether the confidence level corresponding to each real-time status data point is greater than or equal to the first threshold. The system switches back to the abnormal data transmission mode when multiple consecutive real-time status data points have confidence levels greater than or equal to the first threshold.
[0083] In this embodiment, by switching to the full data transmission mode, it can be ensured that the ground data center obtains the most complete, authentic, and accurate raw data sequence for subsequent in-depth analysis.
[0084] In another possible implementation, considering that most of the sensor sampling data is normal state data, if no data is transmitted, it may trigger abnormal alarms in the ground data center. To avoid this problem, alternatively, during the transmission of each real-time state data, the timer can be restarted after each real-time state data is transmitted to the ground data center. If no real-time state data is transmitted within the target time period, a heartbeat packet indicating that the system is normal is sent to the ground data center when the target time period is reached, and the timer is restarted. Here, system normality includes: normal locomotive operation and normal communication link between the locomotive-end equipment and the ground data center.
[0085] This embodiment helps maintain link activity by sending heartbeat packets to the ground data center, ensuring timely transmission to the ground data center in the event of an anomaly, and preventing the ground data center from misjudging the locomotive-side equipment as offline, faulty, or disconnected.
[0086] Optionally, the target timing duration can be a fixed duration value or a variable duration value.
[0087] In one possible implementation, in order to monitor system anomalies in a timely manner, this embodiment can initialize the target timing duration to a first duration value (such as 5 minutes) and change it during transmission according to the confidence level and the magnitude of the deviation value.
[0088] Specifically, during transmission, if the confidence level of the real-time status data transmitted to the ground data center is within the intermediate confidence level range, the target timing duration is shortened. The intermediate confidence level range consists of a first threshold and a second threshold greater than the first threshold. If the confidence level of all untransmitted real-time status data within the target timing duration is greater than the second threshold, the target timing duration is restored to the first duration value.
[0089] To explain in more detail, during transmission, when the deviation between a real-time status data (denoted as data a) and the corresponding predicted status value exceeds the deviation threshold, data a is transmitted to the ground data center. At the same time, if the confidence level corresponding to data a is detected to be within the intermediate confidence level range (i.e., medium confidence level, large deviation), it indicates that data a has undergone unpredictable changes, but it is uncertain whether it is noise. Therefore, a high-frequency monitoring mode can be activated to shorten the target timing duration, such as from 5 minutes to 3 minutes, in order to promptly detect locomotive operation anomalies.
[0090] If no real-time status data is transmitted within a target time period (such as 3 minutes or 5 minutes), and the confidence level of all untransmitted real-time status data within that target time period is greater than the second threshold (i.e., high confidence and small deviation), it indicates that the current data is highly predictable and the status is stable. At this time, a heartbeat packet can be sent when the target time period is reached, and the target time period can be restored to the first time period value.
[0091] Of course, during the transmission process, there may be situations where the confidence level of the real-time status data is greater than the second threshold, and the deviation between the real-time status data and the corresponding state prediction value is greater than the deviation threshold (i.e., high confidence and large deviation). In this case, the real-time status data can be transmitted, but the target timing duration will not be changed.
[0092] In addition, there may be cases where the confidence level of the real-time status data is within the intermediate confidence level range, and the deviation between the real-time status data and the corresponding state prediction value is less than or equal to the deviation threshold (i.e., medium confidence level, small deviation). In this case, the real-time status data can be left untransmitted, and the target timing duration can be left unchanged.
[0093] This embodiment can adjust the timing window in a timely manner based on the changes in confidence level corresponding to the real-time status data collected by the locomotive sensors, so that the heartbeat packets can be transmitted to the ground data center on demand and in a timely manner. The ground data center can detect locomotive operation abnormalities and take appropriate measures in a timely manner based on the heartbeat packet transmission interval.
[0094] In another possible implementation, in order to ensure that the real-time status data that needs to be transmitted can still be successfully transmitted to the ground data center under conditions such as insufficient bandwidth or network anomalies, this embodiment can deploy a target buffer on the locomotive-side equipment. Optionally, the target buffer can be a large-capacity circular data buffer.
[0095] Therefore, before transmitting the real-time status data to the ground data center in step S104 above, this embodiment can first cache the real-time status data in the target buffer. In addition, in the full data transmission mode described above, before transmitting all subsequently collected real-time status data to the ground data center, all subsequently collected real-time status data can also be cached in the target buffer to prevent data overflow.
[0096] By setting a target buffer, real-time status data that needs to be transmitted but cannot be transmitted to the ground data center due to network interruptions or other reasons can be temporarily cached. When the network recovers or a specific data request is received from the ground data center, the cached real-time status data can be retransmitted (i.e., resuming interrupted transmission), effectively solving the problem of data loss during network interruptions. Furthermore, the caching mode can more effectively support backtracking queries from the ground data center.
[0097] Considering that the amount of data to be transmitted to the ground data center will increase with the duration of data collection, a large amount of cache space will be occupied if the real-time status data in the target buffer is not deleted. To reduce the cache space usage, this embodiment can optionally delete real-time status data in the target buffer that has been cached for a preset duration threshold (such as 24 hours).
[0098] Optionally, the real-time status data that has been transmitted to the ground data center in the target buffer can also be deleted; this embodiment does not impose specific limitations on this.
[0099] In summary, this application provides a locomotive status data transmission mechanism based on a "prediction, comparison, and selection" process. This mechanism enables data analysis and decision transmission at the data source (i.e., locomotive-end equipment), allowing only high-value abnormal status data to be transmitted, avoiding redundant data transmission and saving bandwidth. Because abnormal status data is immediately identified and prioritized for transmission (normal status data can be transmitted later when bandwidth is sufficient), the ground data center receives abnormal information earlier and faster, facilitating early intervention in locomotive operation anomalies. Furthermore, the data transmitted to the ground data center is already pre-processed and of high value, reducing the workload of data cleaning and filtering at the ground data center. Simultaneously, this application provides a multi-level transmission and caching strategy based on confidence levels, enabling the system to adapt to changes in network quality and different operating conditions, balancing the conflict between real-time performance, reliability, and bandwidth consumption.
[0100] The above describes a locomotive status data transmission method provided by the embodiments of this application. The following will describe the apparatus for performing the above locomotive status data transmission method.
[0101] Please see Figure 2 , Figure 2 This is a schematic diagram of a locomotive status data transmission device provided in an embodiment of this application. Figure 2 As shown, the locomotive status data transmission device may include:
[0102] The data acquisition unit 201 is used to acquire real-time status data collected by locomotive sensors and a normal status data sequence collected before the real-time status data. The normal status data in the normal status data sequence refers to the operating status data collected when the locomotive is in normal operating condition.
[0103] Model prediction unit 202 is used to input the normal state data sequence into the pre-trained first time series prediction model to obtain the state prediction value corresponding to the real-time state data. The state prediction value represents the running state data under the simulated normal operating state at the time of real-time state data acquisition.
[0104] The anomaly detection unit 203 is used to determine whether the real-time status data is abnormal status data based on the status prediction value. Abnormal status data refers to the operating status data collected when the locomotive is in an abnormal operating state.
[0105] The data transmission unit 204 is used to transmit real-time status data to the ground data center when the real-time status data is abnormal, and not to transmit real-time status data when the real-time status data is not abnormal. The ground data center is equipped with a second time-series prediction model with the same structure and parameters as the first time-series prediction model. The second time-series prediction model is used to recover the untransmitted real-time status data based on the normal status data pre-transmitted from the locomotive-end equipment.
[0106] Each module in the aforementioned locomotive status data transmission device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.
[0107] This application also provides an electronic device, which may include at least one processor and a memory connected to the processor, wherein:
[0108] Memory is used to store computer programs;
[0109] The processor is used to execute computer programs to enable electronic devices to implement any of the locomotive status data transmission methods provided in the embodiments of this application.
[0110] refer to Figure 3 The diagram illustrates a structural schematic suitable for implementing the electronic device in the embodiments of this application. The electronic device in the embodiments of this application may include, but is not limited to, fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 3 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.
[0111] like Figure 3 As shown, the electronic device may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. When the electronic device is powered on, the RAM 603 also stores various programs and data required for the operation of the electronic device. The processing unit 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.
[0112] Typically, the following devices can be connected to I / O interface 605: input devices 606 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 607 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 608 including, for example, memory cards, hard drives, etc.; and communication devices 609. Communication device 609 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or have alternatively.
[0113] This application also provides a computer program product including computer-readable instructions, which, when executed on an electronic device, cause the electronic device to implement any of the locomotive status data transmission methods provided in this application.
[0114] This application also provides a computer-readable storage medium that carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any of the locomotive status data transmission methods provided in this application.
[0115] It should also be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and 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. In addition, in the device embodiment drawings provided in this application, the connection relationship between modules indicates that they have a communication connection, which can be implemented as one or more communication buses or signal lines.
[0116] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, 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 is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk, or optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, training equipment, or network device, etc.) to execute the methods described in the various embodiments of this application.
[0117] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product.
[0118] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, training device, or data center to another website, computer, training device, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).
Claims
1. A method of locomotive condition data transmission, characterized by, include: The system acquires real-time status data collected by locomotive sensors and a sequence of normal status data collected prior to the real-time status data. The normal status data in the normal status data sequence refers to the operating status data collected when the locomotive is in normal operating condition. The normal state data sequence is input into a pre-trained first time-series prediction model to obtain the state prediction value corresponding to the real-time state data. The state prediction value represents the running state data under the simulated normal operating state at the time of the real-time state data acquisition. Based on the predicted state value, it is determined whether the real-time state data is abnormal state data. Abnormal state data refers to the operating state data collected when the locomotive is in an abnormal operating state. If so, the real-time status data is transmitted to the ground data center; otherwise, the real-time status data is not transmitted. The ground data center is equipped with a second time-series prediction model with the same structure and parameters as the first time-series prediction model. The second time-series prediction model is used to recover the untransmitted real-time status data based on the normal status data pre-transmitted from the locomotive-end equipment.
2. The locomotive status data transmission method of claim 1 wherein, The second time-series prediction model is trained based on normal state data samples received by the ground data center, and the parameters of the first time-series prediction model are updated with the second time-series prediction model.
3. The locomotive status data transmission method of claim 1 wherein, The step of determining whether the real-time state data is abnormal state data based on the state prediction value includes: Calculate the deviation between the predicted state value and the real-time state data; If the deviation value is greater than the preset deviation threshold, the real-time status data is determined to be abnormal status data; otherwise, the real-time status data is determined not to be abnormal status data.
4. The locomotive status data transmission method according to any one of claims 1 to 3, characterized in that, The first time series prediction model outputs the state prediction value and the confidence level of the state prediction value at the same time, wherein the confidence level of the state prediction value is used as the confidence level of the real-time state data corresponding to the state prediction value. The locomotive status data transmission method further includes: During the transmission of each of the real-time status data, if the confidence level corresponding to any of the real-time status data is less than the first threshold, the full data transmission mode is activated so that all subsequently collected real-time status data are transmitted to the ground data center until the preset full data transmission mode closing condition is met.
5. The locomotive status data transmission method of claim 4 wherein, Also includes: During the transmission of each of the aforementioned real-time status data, the timer is reset every time a piece of real-time status data is transmitted to the ground data center; If no real-time status data is transmitted within the target time period, a heartbeat packet indicating that the system is normal is sent to the ground data center when the target time period is reached, and the timer is restarted. The target timing duration is initialized to a first duration value; During transmission, if the confidence level of the real-time status data transmitted to the ground data center is within the intermediate confidence level range, the target timing duration is shortened. The intermediate confidence level range consists of the first threshold and a second threshold greater than the first threshold. If the confidence level of all untransmitted real-time status data within the target timing duration is greater than the second threshold, the target timing duration is restored to the first duration value.
6. The locomotive status data transmission method of claim 4 wherein, Also includes: Before transmitting the real-time status data to the ground data center, the real-time status data is cached in the target buffer; And / or, before transmitting all subsequently collected real-time status data to the ground data center, cache all subsequently collected real-time status data in the target buffer.
7. The locomotive status data transmission method of claim 6 wherein, Also includes: Delete real-time status data in the target buffer that has been cached for a preset duration threshold.
8. A computer program product, characterised in that, It includes computer-readable instructions that, when executed on an electronic device, cause the electronic device to implement the locomotive status data transmission method as described in any one of claims 1 to 7.
9. An electronic device, characterized in that, It includes at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program to enable the electronic device to implement the locomotive status data transmission method as described in any one of claims 1 to 7.
10. A computer storage medium, characterized in that, The storage medium carries one or more computer programs, which, when executed by an electronic device, enable the electronic device to implement the locomotive status data transmission method as described in any one of claims 1 to 7.