A novel rail transit HMI display screen full-plane key system and device

Abstract

The application discloses a novel rail transit HMI display screen full-plane key system and device, and belongs to the technical field of software. The application aims to solve the technical problem that the existing rail transit HMI key feedback strategy is fixed and it is difficult to balance the avoidance of accidental touch and the guarantee of perceptibility. To this end, the application provides a self-adaptive feedback decision system, which comprises a multi-source data acquisition and preprocessing module, an operation scene recognition module, a key task risk assessment module, a self-adaptive feedback strategy generation module, an operation closed-loop monitoring and effect evaluation module and a strategy model online updating module. Compared with the prior art, the application dynamically evaluates the key task risk, generates a self-adaptive feedback strategy in combination with the driver preference and the environmental constraint, and performs closed-loop monitoring and online updating, and has the advantages of significantly improving operation safety, optimizing energy consumption efficiency and enhancing man-machine ergonomics.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction technology in rail transit, and more specifically, to a novel all-flat button system and device for rail transit HMI displays. Background Technology

[0002] In the rail transit sector, button operations on the Human-Machine Interface (HMI) display screen are the core method for drivers to interact with the train control system. The button logic and operational requirements for the HMI interface vary significantly depending on the different stages of train operation, such as entering and leaving stations, cruising within sections, turning back, and troubleshooting. For example, the importance, frequency of operation, and potential risks of regular operation buttons, one-time confirmation buttons, safety-critical buttons, and frequently used buttons all differ.

[0003] In low-light environments such as at night or in tunnels, drivers may rely more on tactile and optical feedback to determine whether an operation was successful. However, existing rail transit HMI button systems generally suffer from a fixed feedback strategy. That is, the vibration intensity, vibration mode, and LED brightness / flickering strategy of all buttons are uniformly configured, making it difficult for the system to strike a balance between avoiding accidental touches and ensuring perceptibility.

[0004] To address the above problems, this invention proposes a solution. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a novel all-flat button system and device for rail transit HMI display screens to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A novel all-flat button system for rail transit HMI displays includes: The multi-source data acquisition and preprocessing module is used to acquire key event data, train operation data, environmental data and driver identity information, and to perform normalization and time alignment processing on the data. The running scene recognition module is connected to the multi-source data acquisition and preprocessing module and is used to identify the current running scene of the train based on the preprocessed multi-source data. The button task risk assessment module is connected to the running scenario identification module and is used to dynamically assess the risk score of each button at the current moment by comprehensively considering the button function category, security level, historical misoperation records and the running scenario. An adaptive feedback strategy generation module, connected to the button task risk assessment module, is used to generate a feedback strategy for each button based on the risk score, preset driver preference parameters, and environmental constraints, including the vibration feedback intensity, vibration frequency, LED brightness, LED color, LED flashing mode, and whether to enable a special confirmation mode. The operation closed-loop monitoring and effect evaluation module is connected to the adaptive feedback strategy generation module and is used to monitor the actual effect after the key operation and quantify the comprehensive effect of the feedback strategy. The strategy model online update module is connected to the operation closed-loop monitoring and effect evaluation module, and is used to update the model parameters in the key task risk assessment module and the adaptive feedback strategy generation module online based on the comprehensive effect evaluation results.

[0007] In a preferred embodiment, the multi-source data acquisition and preprocessing module is specifically used to acquire button ID, press timestamp, pressure time sequence, current LED configuration, current vibration spring configuration, current speed, current position and platform distance, current operating mode, train control / signal status, cab illuminance, temperature, humidity, and driver ID; The operation scenario identification module is specifically used to classify the train operation status into at least one of the following scenarios: station entry scenario, station exit scenario, section cruise scenario, turnaround / shunting scenario, and fault / abnormal handling scenario.

[0008] In a preferred embodiment, the running scenario identification module specifically uses a combination of decision trees, random forests, and rules and models to identify the running scenario.

[0009] In a preferred embodiment, the button task risk assessment module is specifically used to combine the static attributes, dynamic scene labels, and behavioral history of the button into a feature vector, and to calculate the risk score using a logistic regression or shallow neural network model.

[0010] In a preferred embodiment, the static attributes of the button include function category, security level, and recommended basic feedback level; the dynamic scene label is the current scene label identified by the running scene recognition module; the behavior history includes the number of times the button has been pressed repeatedly, the number of times the button has been undone, and the correlation with alarms or abnormal events in a recent period.

[0011] In a preferred embodiment, the adaptive feedback strategy generation module is specifically used to divide the risk score interval into multiple feedback levels, and calculate the vibration feedback intensity, vibration frequency, LED brightness, and LED flashing frequency based on the feedback level, the driver preference parameters, and the environmental constraints.

[0012] In a preferred embodiment, the driver preference parameter includes the driver's sensitivity to vibration intensity. And preferences for LED brightness The environmental constraints include a temperature correction function. and illuminance correction function .

[0013] In a preferred embodiment, the vibration feedback intensity LED brightness Vibration frequency and LED blinking frequency The calculation method is as follows: in, , , , The preset base values ​​are for each feedback level. The feedback level is described above.

[0014] In a preferred embodiment, the strategy model online update module is specifically used to update the weights in the key task risk assessment module using gradient descent or rule-based discrete adjustment based on the strategy effect score quantified by the operation closed-loop monitoring and effect evaluation module, adaptively adjust the driver preference parameters based on long-term driver usage data, and fine-tune the environment correction function based on accidental / missed touches and feedback effects under different environmental conditions.

[0015] A novel all-flat button device for rail transit HMI displays includes a glass touch panel located above a pressure sensor, and a signal acquisition circuit board. The signal acquisition circuit board is equipped with a pressure sensor and a vibration spring, and a support plate is provided at the bottom of the signal acquisition circuit board.

[0016] The technical effects and advantages of this invention, a novel all-flat button system and device for rail transit HMI displays: This invention reduces the accidental touch rate of high-risk buttons and improves the success rate of safety-critical operations through dynamic risk assessment and adaptive feedback. The system can provide differentiated feedback based on real-time operating conditions and button importance, effectively avoiding misoperations and repetitive operations, and greatly improving the safety of rail transit operations.

[0017] This invention takes into account individual driver differences and environmental factors, and can provide personalized, scenario-based feedback experiences. For example, it can automatically enhance LED brightness and vibration feedback in low-light environments, or adjust the feedback intensity according to driver preferences, thereby reducing the driver's operational burden and cognitive confusion, and improving driving comfort and efficiency.

[0018] This invention employs an environmental adaptive compensation mechanism, enabling the system to dynamically adjust LED brightness and vibration intensity based on actual environmental conditions (such as illuminance and temperature), avoiding unnecessary excessive feedback. This reduces LED and vibration energy consumption, improves system endurance, and achieves energy conservation and environmental protection. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a novel all-flat button device for a rail transit HMI display screen according to the present invention; Figure 2 This is a schematic diagram of a novel all-flat button system for rail transit HMI displays. Figure labeling: 1. Glass touch panel; 2. Signal acquisition circuit board; 3. LED indicator; 4. Pressure sensor; 5. Vibration spring; 6. Support plate. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0021] This invention provides a novel all-flat button system and device for rail transit HMI displays, aiming to solve the technical problem that existing rail transit HMI button feedback strategies are fixed, making it difficult to balance avoiding accidental touches and ensuring perceptibility. This invention introduces an adaptive feedback decision system, enabling dynamic, personalized, and scenario-based adjustments to button feedback, thereby significantly improving operational safety, optimizing ergonomics and energy efficiency, and possessing long-term self-optimization capabilities.

[0022] Example 1, refer to Figure 2 This invention provides a novel all-flat button system for a rail transit HMI display screen, comprising: a multi-source data acquisition and preprocessing module, an operation scenario recognition module, a button task risk assessment module, an adaptive feedback strategy generation module, an operation closed-loop monitoring and effect evaluation module, and a strategy model online update module. These modules work together to form a complete closed-loop feedback system. Specifically: The multi-source data acquisition and preprocessing module is responsible for collecting all necessary information and processing it uniformly. Specifically, it retrieves the following receipts: Data Collection Target: This module can collect various heterogeneous data in real time, including: Key-side data: Key ID ( ), Press timestamp ( ), pressure time series (used to analyze the pressure and duration of the press), current LED configuration (color, brightness, flashing frequency), and current vibration spring configuration (drive voltage, vibration duration, vibration frequency).

[0023] Train operation data: Current speed ( ), current location and distance from the platform ( ), current operating mode (such as manual / automatic / turnaround / shunting within the section, etc.), train control / signal status (such as signal protection level, target speed, speed limit section ahead).

[0024] Environmental data: Cab illuminance ( Used to determine day / night, tunnel / open-air, and temperature. ),humidity( ).

[0025] Driver status data: Driver ID ( ), current shift working hours, previous error records, etc.

[0026] Normalization and Time Alignment: To ensure that subsequent modules can make judgments based on consistent data, this module unifies data with different sampling frequencies to a fixed time step (e.g., 50ms or 100ms) through interpolation or downsampling. Simultaneously, it performs normalization or standardization on numerical data, such as velocity normalization. Illuminance normalization For discrete variables (such as operating mode and signal state), one-hot encoding or integer encoding is used.

[0027] By comprehensively collecting and uniformly processing information from multiple sources, this module ensures the integrity and consistency of the data foundation for subsequent scene identification, risk assessment, and strategy generation, thus avoiding decision-making biases caused by data mismatch.

[0028] The operational scenario recognition module, connected to the multi-source data acquisition and preprocessing module, is used to abstract complex train operation states into manageable operational scenarios. Specifically, it includes: Scenario Definition: Several typical operating scenarios are predefined, such as: Station Approach Scenario: The train approaches the platform, speed decreases, and braking operations are frequent. Station Departure Scenario: The train leaves the platform, speed increases. Section Cruise Scenario: Train speed is basically stable, with signal constraints as the primary factor. Turnaround / Shunting Scenario: The train operates at low speed, with numerous operations and limited space. Fault / Anomaly Handling Scenario: The system generates alarms or alerts.

[0029] Scene recognition model: This module uses feature sequences from a recent period of time. As input, lightweight machine learning models (such as decision trees or random forests) can be used for classification, or a combination of rules and models can be used. For example, rules can be used to first determine whether an alarm sign exists to identify the fault scenario, and then a classifier can be used to distinguish between entering / leaving / cruising. Output: Current scene label .

[0030] Thus, this invention compresses multi-source information into a single scene label. This module provides a clear contextual dimension for subsequent risk assessment, enabling the system to make intelligent decisions based on the actual operating conditions of the train.

[0031] The button task risk assessment module, connected to the runtime scenario recognition module, dynamically assesses the operational risk of each button at the current moment. Specifically, each button... Static attributes are configured at the factory, including: Functional categories: such as emergency braking, mode switching, door control, lighting control, etc. Safety level: For example, 3 represents safety criticality, and 1 represents normal functionality. Recommended basic feedback level: These are factory specifications, used as an initial reference.

[0032] It also includes dynamic information: combined with the current scene tags. Historical behavior data of buttons, including: The number of times this button has been pressed recently (in a short time window) (According to the number of times).

[0033] Undo operation count: The number of times the operation is reversed immediately after pressing the button.

[0034] The degree of correlation with alarms or abnormal events.

[0035] Finally, a feature vector and risk scoring model are constructed, which combines the above static and dynamic information into a feature vector. Risk scores are calculated using logistic regression or shallow neural network models. ,in For the Sigmoid function or normalization function, map the result to , This refers to the weight vector obtained through offline training and online fine-tuning. A rule-model hybrid approach can also be used, for example, depending on the security level. And the current scenario For entry or failure scenarios, set The lower limit is 0.7.

[0036] This module enables dynamic assessment of button risks, allowing feedback strategies to be adjusted in real time based on the inherent attributes of the buttons, the current operating scenario, and the driver's operation history, thereby effectively reducing the risk of misoperation.

[0037] An adaptive feedback strategy generation module, connected to the key-press task risk assessment module 130, is used to transform abstract risk scores into specific key-press feedback parameters and confirmation modes. Specifically, it includes: Feedback level mapping: Dividing the risk scoring range into multiple feedback levels. ,For example: in To configure thresholds. A comprehensive adjustment of driver preferences and environmental constraints: Driver preference parameters: for each driver Maintain a preference vector ,For example This indicates the driver's sensitivity to vibration intensity. It should be noted that a value >1 indicates a preference for strong vibrations. Indicates a preference for brightness. This indicates tolerance for long-press confirmation.

[0038] Environmental constraints: if illuminance In low temperatures (nighttime / tunnel), the system should enhance LED brightness and vibration feedback. If the temperature... If the LED is already close to its thermal design limit, then limit the LED brightness and increase the vibration weight.

[0039] Specific feedback parameters are calculated based on vibration intensity. Vibration frequency LED brightness flicker frequency For example, the calculation method is as follows: in, , , , These are the preset base values ​​for each feedback level. For example, a temperature correction function: For example, an illuminance correction function to enhance brightness in low illuminance conditions: Confirmation mode decision: When the feedback level When the button is a safety-critical function, enable long-press confirmation or double-click confirmation mode. Long-press threshold. It will be dynamically adjusted based on the risk score.

[0040] This module enables refined, personalized, and environmentally adaptive adjustment of feedback strategies, ensuring that buttons provide the most suitable and effective feedback under different scenarios, drivers, and environmental conditions, thus balancing operational safety and human-machine efficiency.

[0041] The operation closed-loop monitoring and effect evaluation module, connected to the adaptive feedback strategy generation module, is used to monitor and quantify the actual effect of the feedback strategy. This includes: Monitoring metrics: For each button press and each scenario, the following metrics will be calculated: Repeat operation rate: .

[0042] Cancel rate: The percentage of times that the action is immediately reversed or canceled after the action is pressed.

[0043] Accidental touch probability approximation: based on statistics of the combination of immediate cancellation after pressing and alarm after operation.

[0044] Driver feedback: Subjective scoring of the clarity of feedback through maintenance terminals.

[0045] Overall Strategy Effectiveness Rating: For each button and scenario, construct a strategy effectiveness rating. ,in The average energy consumption index for LED + vibration. These are weighting coefficients. Beneficial effects: This module provides an objective and measurable basis for subsequent strategy optimization by quantitatively evaluating the effectiveness of the strategy, ensuring continuous system improvement.

[0046] The strategy model online update module, connected to the operation closed-loop monitoring and effect evaluation module, is responsible for updating model parameters online based on evaluation results, enabling long-term self-optimization of the system. This includes: Risk model weight update: Under the premise of ensuring safety, the weights in the risk scoring model are updated using a small-step online learning approach. The goal is to reduce the rate of repeated operations and undo operations without increasing accidental or missed touches. Gradient descent or rule-based discrete adjustments can be used. For example, if a button has a high rate of repeated operations in a certain scenario, the weight corresponding to that scenario feature can be appropriately increased to improve the risk score. This improves the feedback level.

[0047] Driver preference parameter adaptive: Based on long-term driver usage data, if a driver frequently presses a button 2-3 times after pressing it, it indicates that the current feedback may be weak, and the system can automatically increase the sensitivity. , If a driver frequently shows dissatisfaction with long-press confirmation, the long-press threshold can be appropriately lowered. The coefficient.

[0048] Environmental adaptive parameter optimization: Analyzes the effects of false / missed touches and feedback under different environmental conditions, and automatically fine-tunes the temperature correction function. and illuminance correction function Shape (e.g., the magnitude of correction for increasing / decreasing temperature and illuminance).

[0049] This module allows the system to be gradually optimized as the route is operated, drivers' habits change, and hardware ages, rather than remaining at the factory settings forever, thus ensuring the long-term robustness and optimal performance of the system.

[0050] Example 2, refer to Figure 2 This invention provides a novel method for using full-plane buttons on a rail transit HMI display screen, which includes the following steps: Step S1: Multi-source Data Acquisition and Normalization Preprocessing. This step corresponds to the multi-source data acquisition and preprocessing module in the system. It involves real-time acquisition of button-side data such as button ID, press timestamp, pressure time series, current LED configuration, and current vibration spring configuration; train operation data such as train speed, position, operating mode, and signal status; environmental data such as cab illuminance, temperature, and humidity; and driver status data such as driver ID. Data from different sampling frequencies are then unified to a fixed time step (e.g., 50ms) through interpolation or downsampling. Numerical data is then normalized or standardized, such as… and The discrete variables are encoded using one-hot encoding or integer encoding. This step does not consider whether a button is pressed in isolation, but simultaneously incorporates the train's operating status, environmental status, and driver identification, and ensures that any subsequent judgment is based on multi-source information from the same moment by using a unified timeline.

[0051] Step S2: Scenario Identification. This step corresponds to the scenario identification module in the system. To summarize the complex and ever-changing operational states into a finite number of operational scenarios, providing a clear scenario dimension for subsequent risk assessment, Step S2 will use preprocessed feature sequences from a recent period (e.g., the past 5 seconds). This serves as input to the scene recognition model. Pre-trained decision trees, random forests, or a combination of rules and models are employed. For example, a rule can be set: if the train speed is below 5 km / h and the distance to the platform is less than 100m, it is initially judged as either an entry or exit scenario; then, combined with signal status and speed change trends, a classifier further subdivides it into entry or exit. If an emergency braking command or system alarm is present, it is directly identified as a fault / abnormal handling scenario. Finally, the current scene label is output. Therefore, the scene recognition of this invention does not rely solely on speed or location for judgment, but simultaneously uses multi-source information such as speed change trend, distance from the station, signal status, and illumination, and compresses this information into a scene label through a model.

[0052] Step S3: Dynamic Assessment of Key Pressing Task Risk. This step corresponds to the key pressing task risk assessment module in the system. This is achieved by acquiring key press data. Preset function categories and security levels And recommended basic feedback level Get the current scene label. Count the number of times this button has been pressed repeatedly within a recent period. Number of times to cancel operation And the correlation with alarms or abnormal events. The above information is then combined into a feature vector. Finally, a pre-trained logistic regression model or a shallow neural network model is used to calculate the risk score. For example, if (Safety is critical) and For fault scenarios, the model will assign higher weights, making Significantly improved. If If it is abnormally high in a short period of time, it should also be increased. This step does not only consider the safety level of the button function or a single aspect of the current scenario, but also takes into account three dimensions: static safety level, current scenario label, and recent behavior history. The risk score is a dynamic quantity that is updated in real time as the scenario changes and the driver's behavior changes.

[0053] Step S4: Adaptive Feedback Strategy Generation Based on Risk Score. This step corresponds to the adaptive feedback strategy generation module in the system. It transforms abstract risk scores and scenario information into specific vibration and LED feedback parameters, and determines whether to enable a special confirmation mode. This includes generating the adaptive feedback strategy based on the risk score. Map it to a four-level feedback system. .For example, Level 0 (low risk). It is classified as Level 1 (low to medium risk). It is classified as Level 2 (medium to high risk). Level 3 (High Risk). Loading the current driver. Preference parameters And based on current environmental data (temperature) Illuminance ), calculate the temperature correction function and illuminance correction function Then, based on the feedback level... Driver preference parameters and environmental correction functions are used to calculate the specific vibration intensity. Vibration frequency LED brightness flicker frequency .For example: in, It is a preset base value table, based on Retrieve from table. If If the button is for a safety-critical function (such as emergency braking), then a long-press confirmation mode will be enabled. Long-press threshold. Based on risk score Dynamic adjustment, for example The higher the risk, the longer the press and hold time. The feedback in this invention is not solely determined by the risk score, but rather by a combination of driver preference parameters and environmental constraints. For high-risk buttons, in low-light conditions and when the driver prefers strong feedback, the vibration and LED intensity will be automatically amplified, and more stringent confirmation will be required.

[0054] Step S5: Closed-Loop Monitoring and Strategy Effectiveness Quantification of Key Operations. This step corresponds to the closed-loop monitoring and effectiveness evaluation module in the system. It involves real-time statistics of the repetitive operation rate of each key in different scenarios. cancellation rate For example, calculations can be made by recording whether the same button is pressed again or the opposite operation occurs within one second after a button is pressed. Furthermore, average energy consumption is calculated based on LED brightness and vibration intensity. Finally, the overall strategy effectiveness score is calculated. .

[0055] Step S6: Online Update and Long-Term Evolution of the Policy Model. This step corresponds to the online update module of the policy model in the system. Based on the policy performance score... The weights in the risk model are updated using either small-step gradient descent or rule-based discrete adjustment. For example, if a certain button is used in a certain scenario... If it remains low, then adjust. The weights of corresponding features are assigned to improve the risk score of the button in that scenario, thereby prompting the system to generate stronger feedback. Furthermore, the system analyzes drivers' operating habits based on long-term (e.g., one month) usage data. If a driver frequently triggers a long press to confirm but releases it near the threshold, it indicates a low tolerance for long presses, and the system's tolerance for long presses is appropriately reduced. The parameters were adjusted to shorten the long-press threshold. Finally, the false / missed touch rate and feedback effect were analyzed under different temperature and illuminance conditions. If excessive LED brightness in a high-temperature environment leads to an increased failure rate, fine-tuning was performed. The function makes it more effective at suppressing LED brightness at high temperatures.

[0056] Example 3, as Figure 1 As shown, this invention also discloses a novel all-flat button device for a rail transit HMI display screen, specifically comprising: 1. Glass touch panel, 2. Signal acquisition circuit board, 3. LED indicator, 4. Pressure sensor, 5. Vibration spring, 6. Support plate.

[0057] The glass touch panel 1 is located above the pressure sensor 4. Different button symbols can be printed on the surface of the glass panel 1, and the glass touch panel 1 can be tempered to achieve explosion-proof and scratch-resistant properties. It has advantages such as long service life and easy cleaning. The pressure sensor 4 and vibration spring 5 are installed on the signal acquisition circuit board 2, and a support plate 6 is installed at the bottom of the signal acquisition circuit board 2.

[0058] When a finger presses on the glass touch panel 1, the pressure sensor 4 can collect the pressure value of the glass panel 1 in real time. The signal acquisition circuit board 2 converts the pressure value into an electrical signal and feeds the signal back to the vibration spring 5 and the LED indicator 3. After receiving the signal feedback, the vibration spring 5 vibrates, providing tactile feedback to the operator, and simultaneously illuminating the LED indicator 3 below the button. With the dual function of tactile feedback and the LED indicator, the operator can clearly perceive whether the button has been triggered.

[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended 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. Such 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 novel all-flat button system for rail transit HMI displays, characterized in that, include: The multi-source data acquisition and preprocessing module is used to acquire key event data, train operation data, environmental data and driver identity information, and to perform normalization and time alignment processing on the data. The running scene recognition module is connected to the multi-source data acquisition and preprocessing module and is used to identify the current running scene of the train based on the preprocessed multi-source data. The button task risk assessment module is connected to the running scenario identification module and is used to dynamically assess the risk score of each button at the current moment by comprehensively considering the button function category, security level, historical misoperation records and the running scenario. An adaptive feedback strategy generation module, connected to the button task risk assessment module, is used to generate a feedback strategy for each button based on the risk score, preset driver preference parameters, and environmental constraints, including the vibration feedback intensity, vibration frequency, LED brightness, LED color, LED flashing mode, and whether to enable a special confirmation mode. The operation closed-loop monitoring and effect evaluation module is connected to the adaptive feedback strategy generation module and is used to monitor the actual effect after the key operation and quantify the comprehensive effect of the feedback strategy. The strategy model online update module is connected to the operation closed-loop monitoring and effect evaluation module, and is used to update the model parameters in the key task risk assessment module and the adaptive feedback strategy generation module online based on the comprehensive effect evaluation results. The button task risk assessment module is specifically used to combine the static attributes, dynamic scene labels and behavioral history of the button into a feature vector, and to calculate the risk score using logistic regression or a shallow neural network model. The static attributes of the button include function category, security level, and recommended basic feedback level; the dynamic scene label is the current scene label identified by the running scene recognition module; the behavior history includes the number of times the button has been pressed repeatedly, the number of times the operation has been undone, and the correlation with alarms or abnormal events in a recent period of time. The adaptive feedback strategy generation module is specifically used to divide the risk score interval into multiple feedback levels, and calculate the vibration feedback intensity, vibration frequency, LED brightness, and LED flashing frequency based on the feedback level, the driver preference parameters, and the environmental constraints.

2. The novel all-flat button system for rail transit HMI displays according to claim 1, characterized in that: The multi-source data acquisition and preprocessing module is specifically used to acquire button ID, press timestamp, pressure time sequence, current LED configuration, current vibration spring configuration, current speed, current position and distance from the platform, current operating mode, train control / signal status, cab illuminance, temperature, humidity, and driver ID; The operation scenario identification module is specifically used to classify the train operation status into at least one of the following scenarios: station entry scenario, station exit scenario, section cruise scenario, turnaround and shunting scenario, and fault and anomaly handling scenario.

3. A novel all-flat button system for rail transit HMI displays according to claim 2, characterized in that: The operation scenario identification module specifically uses a combination of decision trees, random forests, and rules and models to identify the operation scenario.

4. The novel all-flat button system for rail transit HMI display screen according to claim 1, characterized in that: The driver preference parameters include the driver's sensitivity to vibration intensity. And preferences for LED brightness The environmental constraints include a temperature correction function. and illuminance correction function .

5. A novel all-flat button system for rail transit HMI displays according to claim 4, characterized in that: The vibration feedback intensity LED brightness Vibration frequency and LED blinking frequency The calculation method is as follows: , , , ;in, , , , The preset base values ​​are for each feedback level. The feedback level is described above.

6. A novel all-flat button system for rail transit HMI displays according to claim 1, characterized in that: The online update module of the strategy model is specifically used to update the weights in the button task risk assessment module by using gradient descent or rule-based discrete adjustment methods based on the strategy effect score quantified by the operation closed-loop monitoring and effect evaluation module, adaptively adjust the driver preference parameters based on the driver's long-term usage data, and fine-tune the environment correction function based on the accidental / missed touches and feedback effects under different environmental conditions.

7. The button device of a novel all-flat button system for rail transit HMI displays according to claim 1, characterized in that; It includes a glass touch panel located above the pressure sensor, a signal acquisition circuit board with a pressure sensor and a vibration spring, and a support plate at the bottom of the signal acquisition circuit board.

Images