Auxiliary system for safe driving of train

Abstract

The invention discloses an auxiliary system for safe driving of a train. The auxiliary system comprises an operation environment monitoring module, a self-state detection module, a decision analysis module and a man-machine interaction module. The four modules are based on different algorithms, monitoring of the surrounding environment and the state of the train is improved when the train travels, data information can be processed and fed back in time when the train travels at a high speed, and the automation performance is improved; multiple interaction modes in the man-machine interaction module improve cooperation between a driver and a train when the driver drives the train, facial expressions and gestures of the driver are analyzed based on an edge detection algorithm and a face detection algorithm of deep learning, and the algorithms cooperate with one another and work cooperatively, so that the driving safety of the driver is improved. The man-machine interaction module can accurately understand the intention of the driver, monitor the state of the driver and transmit information and suggestions of the system to the driver in a proper mode, and therefore the overall performance of the train safe driving auxiliary system and the man-machine cooperation efficiency are improved.

Description

Technical Field

[0001] The present invention relates to a train safety driving system, in particular to a train safety driving auxiliary system, and belongs to the technical field of train safety driving. Background Art

[0002] The train safety driving assistance system uses multi-sensor fusion technology to provide train safety driving assistance system solutions for situations such as poor lighting, bad weather, and obstacle intrusion that may be encountered during the normal operation of rail transit trains. The locomotive crew's compliance status plays an important role in train operation.

[0003] After searching, Chinese patent number CN116534080A discloses a train safe driving auxiliary reminder system, an image intelligent recognition processor equipped with a face recognition monitor, an infrared distance monitor and a shunting signal monitor, a phase zone automatic detection processor equipped with a speedometer and a real-time handle status monitor, and an audio line connected to a voice player. The analysis and processing of data by the on-board intelligent analysis processor avoids a series of safety problems that occur during train driving. However, the current design of the human-computer interaction interface may not be intuitive and humane enough, resulting in difficulties for drivers in obtaining and understanding auxiliary information, affecting the use of the system, and the overall response speed of the system in the above patent document may not be fast enough. From the detection of anomalies by the monitoring equipment to the reminder issued by the voice player in the cab, the signal processing, data transmission and decision-making judgment in this process all require a certain amount of time. If there is a delay in any of these links, the driver may not be able to obtain key information in time, thereby affecting the timeliness of his response measures. Summary of the invention

[0004] The purpose of the present invention is to provide a train safety driving assistance system in order to solve the above-mentioned problems.

[0005] The present invention achieves the above-mentioned purpose through the following technical solutions: a train safety driving assistance system, including an operating environment monitoring module, a self-state detection module, a decision analysis module and a human-computer interaction module, wherein the operating environment monitoring module is used to monitor the surrounding environment in real time when the train is running, and the operating environment monitoring module can ensure driving safety, improve the operating efficiency of the train, and enhance the emergency response capability of the train in the event of an unexpected situation, and the operating environment monitoring module includes a track condition monitoring unit, a surrounding obstacle detection unit and a meteorological condition perception unit;

[0006] The self-state monitoring module is used to monitor the train's own situation in real time, and can monitor the train's speed and position to ensure the running order, prevent the risk of failure, optimize driving operations, reduce unnecessary parking and maintenance, and more quickly locate the area and cause of the failure. The self-state monitoring module includes a speed and position monitoring unit, a vehicle key component monitoring unit, and a train tilt angle monitoring unit;

[0007] The decision analysis module is used to integrate a large amount of heterogeneous data in the operating environment monitoring module and the self-state detection module collected during train driving, and to sort them out according to unified data format and timestamp rules so that they can be effectively used by subsequent analysis and processing processes, and based on the integrated data, use the risk assessment model to analyze and quantitatively evaluate various potential risks faced by train operation. The decision analysis module includes a data collection and integration submodule, a risk assessment submodule, a decision-making submodule, a strategy execution and feedback submodule, a plan management submodule, and a learning and optimization submodule;

[0008] The human-computer interaction module is used to convey the warning information, operation suggestions and various abnormal conditions monitored by the system generated by the decision analysis module to the driver in an intuitive manner in a timely manner to ensure that the driver's operations are accurately performed. The human-computer interaction module includes voice interaction, in-cabin display screen interaction, control handle and button interaction, touch screen interaction, camera monitoring interaction, biosensor monitoring interaction, smart watch / bracelet interaction and gesture interaction.

[0009] Preferably, the track condition monitoring unit includes sensor applications and algorithm analysis, the sensor applications include laser displacement sensors, ultrasonic sensors and electromagnetic induction sensors, and the algorithm analysis is based on a track defect identification algorithm and a track geometry parameter calculation algorithm based on machine learning.

[0010] Preferably, the surrounding obstacle detection unit includes sensor applications and algorithm analysis, the sensor applications include millimeter wave radar sensors, lidar sensors and high-definition cameras, and the algorithm analysis is based on a weighted average fusion algorithm.

[0011] Preferably, the meteorological condition sensing unit includes a wind speed and direction sensor, a rainfall sensor and a visibility sensor, and the meteorological condition sensing unit and the surrounding obstacle detection unit perform calculations based on a Kalman filter fusion algorithm.

[0012] Preferably, the speed and position monitoring unit includes a photoelectric speed sensor and a magnetoelectric speed sensor, and its position is monitored and located in real time based on a global satellite positioning system.

[0013] Preferably, the vehicle key component monitoring unit includes a vibration sensor, a temperature sensor, a pressure sensor and a brake pad wear detection.

[0014] Preferably, the train tilt angle monitoring unit includes a gyroscope sensor, an acceleration sensor, an inclination sensor and a dual-axis or tri-axis electronic compass, based on an angle solution algorithm, a trigonometric function calculation algorithm, a temperature compensation and nonlinear correction algorithm, and a magnetic field component calculation and angle estimation algorithm.

[0015] Preferably, the decision analysis module transmits information to the human-computer interaction module, and the human-computer interaction module feeds back information to the decision analysis module.

[0016] Preferably, the voice interaction is based on a statistical n-gram language model, the gesture interaction is based on an edge detection algorithm, and the camera monitoring interaction collects the driver's facial image through a camera, based on a deep learning-based face detection algorithm.

[0017] The present invention has the following beneficial effects:

[0018] 1. The operating environment detection module is used to detect when the train is running. The laser displacement sensor detects whether the track is sunken, convex or partially worn, resulting in an uneven surface. The ultrasonic sensor determines whether there are hidden cracks, holes and other defects inside the track. The electromagnetic induction sensor detects whether the track is broken, poorly connected, and other problems that affect the electromagnetic circuit. This can better ensure driving safety, improve the operating efficiency of the train, and enhance the emergency response capabilities of the train in the event of unexpected situations.

[0019] 2. The position and speed of the train during driving are monitored in real time through its own status monitoring module, and the presence of fault frequency components is determined through laser displacement sensors, ultrasonic sensors and electromagnetic induction sensors;

[0020] 3. Multiple modules are based on different algorithms to improve the monitoring of the surrounding environment and the train's own status when the train is running, and can process and feedback data information in time when the train is running at high speed, thereby improving automation performance;

[0021] 4. Various interaction methods in the human-computer interaction module improve the coordination between the driver and the train when driving the train. The driver's facial expressions and gestures are analyzed based on edge detection algorithms and deep learning face detection algorithms (such as MTCNN). These algorithms cooperate with each other and work together to enable the human-computer interaction module to accurately understand the driver's intentions, monitor the driver's status, and convey system information and suggestions to the driver in an appropriate manner, thereby improving the overall performance of the train safety driving assistance system and the efficiency of human-computer collaboration. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A logic block diagram of a train safety driving assistance system proposed by the present invention;

[0023] Figure 2 This is a logic block diagram of an environment detection module in a train safety driving assistance system proposed by the present invention;

[0024] Figure 3 This is a logic block diagram of the self-state monitoring in a train safety driving assistance system proposed by the present invention;

[0025] Figure 4 This is a logic block diagram of a decision analysis module and a human-computer interaction module in a train safety driving assistance system proposed by the present invention. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0027] Embodiment 1:

[0028] Reference Figure 1-4 A train safety driving assistance system includes an operating environment monitoring module, a self-state detection module, a decision analysis module and a human-computer interaction module. The operating environment monitoring module is used to monitor the surrounding environment in real time when the train is running. The operating environment monitoring module can ensure driving safety, improve the operating efficiency of the train and enhance the emergency response capability of the train in the event of unexpected situations. The operating environment monitoring module includes a track condition monitoring unit, a surrounding obstacle detection unit and a meteorological condition perception unit.

[0029] The track condition monitoring unit includes sensor application and algorithm analysis. The sensor application includes laser displacement sensor, ultrasonic sensor and electromagnetic induction sensor. The algorithm analysis is based on machine learning track disease identification algorithm and track geometry parameter calculation algorithm.

[0030] The surrounding obstacle detection unit includes sensor application and algorithm analysis. The sensor application includes millimeter wave radar sensor, lidar sensor and high-definition camera. The algorithm analysis is based on weighted average fusion algorithm.

[0031] The meteorological condition perception unit includes a wind speed and direction sensor, a rainfall sensor and a visibility sensor. The meteorological condition perception unit and the surrounding obstacle detection unit perform calculations based on the Kalman filter fusion algorithm.

[0032] In this embodiment, it should be noted that:

[0033] The four modules are based on different algorithms to improve the train's monitoring of the surrounding environment and its own status while it is running. They can process and feedback data information in a timely manner when the train is running at high speed, thereby improving automation performance.

[0034] The monitoring of geometric parameters such as the flatness and gauge of the track is often calculated using data obtained from measuring equipment such as laser displacement sensors and total stations. For example, the distance values ​​measured by laser displacement sensors at different positions on the track are used to fit the curve of the track surface using data fitting algorithms such as the least squares method, and then the flatness deviation of the track is calculated; for the gauge, it can be obtained through simple difference calculation based on the data measured at specific positions on both sides of the track, and compared with the standard gauge value to determine whether the gauge meets the requirements. These algorithms can monitor in real time whether the geometric state of the track is within the safety standard range, ensuring the stability and safety of train travel.

[0035] Millimeter-wave radar sensors convert time-domain radar echo signals into frequency-domain signals based on the fast Fourier transform (FFT) algorithm and clustering algorithms (such as DBSCAN), thereby analyzing the frequency changes of the reflected waves, and then calculating the distance of the obstacle (based on the relationship between the speed of electromagnetic wave propagation and the round-trip time), relative speed (through the Doppler frequency shift principle) and azimuth (based on information such as the beam pointing of the radar antenna and the phase difference between multiple antennas). Then, through a clustering algorithm (such as DBSCAN), multiple reflection points belonging to the same obstacle are clustered and integrated to determine the approximate outline and accurate location of the obstacle, providing a basis for the train to judge the risk of collision.

[0036] Image recognition algorithms are used to analyze images collected by high-definition cameras installed around the train. Traditional image processing methods include edge detection (such as the Canny edge detection algorithm, which extracts the edge contour of an object by calculating the gradient change of pixels in the image), feature extraction (such as HOG features, which are used to describe the shape and texture information of an object), etc., combined with machine learning classifiers (such as support vector machines SVM) to identify obstacles in the image. Nowadays, target detection algorithms in deep learning (such as the YOLO series, FasterRCNN, etc.) are more commonly used. They can directly locate and identify different types of obstacles in the image, and pass their location, category and other information to the train safety driving assistance system.

[0037] For data collected by wind speed and direction sensors installed at locations such as the top of a train, a vector decomposition algorithm is often used to calculate the wind speed and direction.

[0038] Embodiment 2:

[0039] Different from the first embodiment, Figure 1-2This embodiment also has the following further contents: the self-state monitoring module is used to monitor the train's own conditions in real time, and can enable the train speed and position monitoring to ensure the running order, prevent the risk of failure, optimize the driving operation, reduce unnecessary parking and maintenance, and locate the fault area and cause more quickly. The self-state monitoring module includes a speed and position monitoring unit, a vehicle key component monitoring unit and a train tilt angle monitoring unit.

[0040] The speed and position monitoring unit includes a photoelectric speed sensor and a magnetoelectric speed sensor, which monitor and locate their positions in real time based on the global satellite positioning system. The vehicle key component monitoring unit includes a vibration sensor, a temperature sensor, a pressure sensor, and a brake pad wear detection.

[0041] The train tilt angle monitoring unit includes a gyroscope sensor, an acceleration sensor, an inclination sensor and a dual-axis or tri-axis electronic compass, based on an angle solution algorithm, a trigonometric function calculation algorithm, a temperature compensation and nonlinear correction algorithm, and a magnetic field component calculation and angle estimation algorithm.

[0042] In this embodiment, it should be noted that:

[0043] Based on the principle of conservation of angular momentum, the gyroscope sensor can measure the angular velocity changes of the train around different coordinate axes (usually the roll axis, pitch axis and yaw axis). According to the angular velocity value output by the gyroscope, at discrete time intervals (for example, every certain sampling period, such as 10 milliseconds, etc.), the angular velocity in each time interval is integrated and the accumulated amount of angular velocity over time is converted into the angle change. For example, if within a certain time period, the gyroscope measures the angular velocity around the roll axis as ω (usually in radians / second), and the sampling period is Δ t (seconds), then the change in the roll angle during this period of time Δ θ The formula Δ θ =ω×Δ t To calculate, by successively adding up the angle changes in each time period, we can get the train's tilt angle value relative to the initial state.

[0044] When the train is in a horizontal and stationary state, it is only affected by gravity in the vertical direction and outputs a fixed signal corresponding to the gravity acceleration; when the train tilts, the gravity acceleration components on each measuring axis of the sensor will change. Taking a two-dimensional plane as an example (assuming that the train only has pitch or roll angle changes), let the gravity acceleration be g, and the sensor has measurement axes in the horizontal and vertical directions respectively. When the train tilts at an angle of θ, the acceleration component on the vertical measuring axis is a, and the acceleration component on the horizontal measuring axis is a. y =g×cosθ. By measuring a x and ay The value of the inverse trigonometric function (such as θ = arctan (a x / a y ), and make corresponding adjustments according to the specific quadrant conditions) to calculate the tilt angle θ.

[0045] Embodiment three:

[0046] Reference Figure 1-4 Compared with the first and second embodiments, in this embodiment: the decision analysis module is used to integrate a large amount of heterogeneous data in the operating environment monitoring module and the self-state detection module collected during train driving, and to sort them out according to unified data formats and timestamps and other rules, so that they can be effectively used by subsequent analysis and processing processes, and based on the integrated data, the risk assessment model is used to analyze and quantitatively evaluate various potential risks faced by train operation. The decision analysis module includes a data collection and integration sub-module, a risk assessment sub-module, a decision-making sub-module, a strategy execution and feedback sub-module, a plan management sub-module, and a learning and optimization sub-module.

[0047] The decision analysis module transmits information to the human-computer interaction module, and the human-computer interaction module feeds back information to the decision analysis module.

[0048] In this embodiment, it should be noted that:

[0049] Gesture interaction is based on edge detection algorithms (such as Canny edge detection, local binary pattern LBP, etc.), and camera monitoring interaction collects the driver's facial image through the camera, based on deep learning face detection algorithms (such as MTCNN, etc.).

[0050] The environment monitoring module and the self-state monitoring module monitor and analyze the external and self-state of the train in real time and transmit the data information to the decision analysis module for processing. In the decision analysis template, the data collection and integration submodule is responsible for collecting data from the operating environment monitoring module of the train safety driving assistance system and the train self-state monitoring module. These data cover many aspects of the train operation process, including but not limited to track condition data (such as track flatness, presence of foreign objects, turnout status, etc.), surrounding obstacle information (position, distance, speed, size, etc.), train real-time speed, position, working parameters of key components (such as running gear, braking system, traction system, etc.) and meteorological conditions (wind speed, wind direction, rainfall, visibility, etc.). Then, the large amount of heterogeneous data collected is integrated and sorted according to unified data format and timestamp rules so that it can be effectively used by subsequent analysis and processing processes. For example, the train speed data and track humidity data collected by different sensors at the same time are integrated together to form a complete train real-time operation status data packet.

[0051] The data collection and integration submodule is based on data acquisition interface (serial communication interface (such as RS232, RS485, etc.), Ethernet interface (such as RJ45 interface), CAN bus interface, wireless communication interface (such as GSM-R, LTE-R, etc.)), data caching technology and data preprocessing algorithm (data cleaning, data normalization).

[0052] Z-score normalization (standard score normalization): This method is based on the mean μ and standard deviation σ of the data for normalization, and the calculation formula is Z = (X-μ) / σ.

[0053] Based on the integrated data, the risk assessment submodule uses a risk assessment model to analyze and quantitatively evaluate the various potential risks faced by train operation. The model will comprehensively consider the relationship between different factors and their impact on train safety. For example, when analyzing the risk of train speeding, not only will the difference between the current speed of the train and the speed limit be paid attention to, but also the line position of the train (such as special sections such as curves and ramps), the friction coefficient of the track (affected by weather and track surface conditions), the performance status of the braking system and other factors will be combined to accurately judge the degree of danger that may be caused by speeding; in the case of foreign objects on the track, the size of the collision risk will be evaluated based on the size, material, distance from the train and the current speed of the train. According to the results of the risk assessment, the risks are divided into different levels, such as low risk, medium risk, high risk, etc., so that targeted countermeasures can be taken later.

[0054] The decision-making submodule uses decision trees, expert systems, intelligent optimization algorithms and other technologies to formulate corresponding decision strategies based on the risk level and specific risk conditions obtained by the risk assessment submodule, combined with the train's operating objectives (such as arriving at the station on time, ensuring passenger safety, etc.) and actual operating conditions (such as current line resources, the position relationship of adjacent trains, etc.). For example, when the risk assessment shows that there are small foreign objects on the track ahead of the train and the risk level is low, the decision may be to remind the driver to slow down appropriately and continue driving after observation; if a fault is detected in the train's braking system and the risk level is high, the decision may be to immediately trigger emergency braking and send a fault alarm message to the dispatching center to request support and coordinate subsequent operation arrangements; when facing complex road conditions (such as heavy rain causing water accumulation on the line, construction of the road ahead, etc.), the train's speed, load, remaining braking distance and other factors will be comprehensively considered to formulate a reasonable driving plan, such as selecting a suitable deceleration point, changing the operating line (if there is a feasible backup line), etc., to ensure the safe passage of the train through the relevant area.

[0055] The strategy execution and feedback submodule is responsible for converting the decision strategy determined by the decision-making submodule into specific executable instructions with the help of communication interfaces, instruction encoding and decoding technology, and real-time monitoring technology, and issuing them to the relevant control systems of the train (such as traction system, braking system, signal system, etc.) to ensure that the decision can be effectively implemented. For example, if the decision is to require the train to slow down, an instruction to reduce power will be sent to the traction system, and an instruction to apply appropriate brakes will be sent to the braking system to coordinate the various systems to complete the deceleration operation. During the execution of the instructions, the feedback information of each system is continuously monitored, such as the actual speed change, the change of braking pressure, the response of the signal system, etc., to determine whether the decision execution has achieved the expected effect. If it does not meet the expectations, the decision strategy will be adjusted in time according to the feedback, and the instructions will be re-issued to form a closed-loop decision execution and feedback mechanism to ensure that the train is always in a safe operating state.

[0056] The plan management submodule stores various pre-made emergency response plans, which cover a variety of emergencies that may be encountered during train operation, such as earthquakes, fires, major equipment failures, severe weather disasters, etc. When the train safety driving assistance system detects that the relevant abnormal situation triggers the corresponding plan conditions, the submodule will quickly call out the corresponding plan content and provide detailed response measures and operation instructions to assist drivers, train staff, and dispatching centers and other relevant personnel to quickly and orderly carry out emergency response work, minimize accident losses, and ensure the safety of trains and personnel. For example, when encountering an earthquake warning, the plan will clearly stipulate that the train should immediately take emergency braking measures, open emergency escape passages, and convey shock-absorbing information to passengers through the broadcasting system. A series of operating procedures.

[0057] The learning and optimization submodule records and analyzes the data collected, decisions made, and final operation results during each operation of the train, summarizes lessons learned, and finds deficiencies in the decision analysis module. For example, the reason for a certain decision leading to train delays is analyzed, whether it is inaccurate risk assessment, unreasonable decision strategy, or problems in the execution link. Then, using machine learning, data analysis and other technical means, combined with newly acquired data and actual operation feedback, various links such as risk assessment models and decision-making algorithms are continuously optimized and improved, so that the decision analysis module can better adapt to different operating conditions, continuously improve the accuracy and effectiveness of decisions, and ensure the safe and efficient operation of trains.

[0058] Embodiment 4:

[0059] Reference Figure 1-4Compared with Embodiment 1, Embodiment 2 and Embodiment 3, in this embodiment: the human-computer interaction module is used to convey the warning information, operation suggestions and various abnormal conditions detected by the system generated by the decision analysis module to the driver in an intuitive way in a timely manner to ensure that the driver's operations are accurately performed. The human-computer interaction module includes voice interaction, in-cabin display screen interaction, control handle and button interaction, touch screen interaction, camera monitoring interaction, biosensor monitoring interaction, smart watch / bracelet interaction and gesture interaction.

[0060] Voice interaction is based on the statistical n-gram language model, gesture interaction is based on the edge detection algorithm, and camera monitoring interaction collects the driver's facial image through the camera and is based on the deep learning face detection algorithm.

[0061] In this embodiment, it should be noted that:

[0062] The decision analysis module transmits information to the human-computer interaction module: Based on the analysis and evaluation of various data on train operation, the decision analysis module will transmit the corresponding warning information to the human-computer interaction module once it determines that the train is facing potential risks or abnormal conditions (such as foreign objects detected on the track, failure of key train components, overspeeding, etc.). This information includes details such as the specific content and severity of the risk, such as "stones appear on the track 100 meters ahead, the risk of collision is high, please brake immediately". After receiving it, the human-computer interaction module will display it to the driver in an appropriate way (such as popping up a striking warning window on the display screen in the cab, voice broadcasting in a hurried tone to convey the information, etc.), to ensure that the driver can know it in time and take countermeasures.

[0063] When the decision analysis module develops a decision strategy for specific situations (such as suggesting that the driver slow down to a safe speed due to heavy rain and water accumulation on the road ahead, or suggesting that the driver plan to stop for maintenance in advance due to the degradation of the braking system performance), these decision suggestions will be passed to the human-computer interaction module. The human-computer interaction module will present them to the driver in an intuitive and easy-to-understand form (such as presenting specific operation steps with pictures and texts on the display screen, explaining the suggestions in detail through voice, etc.), assisting the driver to understand and refer to the corresponding decisions to ensure the safe operation of the train.

[0064] A camera is installed in the cab to capture the driver's facial expressions, eyes, body movements and other images in real time, and use image analysis technology to determine whether the driver is in a state of fatigue (such as frequent yawning, long-term eye closure, lack of concentration, etc.) or whether there are abnormal operating behaviors (such as leaving the hands on the control device for too long). Once these abnormal conditions are detected, the system will remind the driver to pay attention to his or her own status through pop-up reminder boxes on the display screen, voice broadcast prompts (such as "You have shown signs of fatigue, please rest and adjust as soon as possible"), etc., and will also record relevant data for subsequent analysis and management.

[0065] Biosensors (such as heart rate sensors, pressure sensors, grip sensors, etc.) installed on seats, steering wheels, etc. are used to detect the driver's physiological data in real time and analyze whether his physical condition is suitable for continuing to drive. For example, when the driver's heart rate is too high or too low, or there are abnormal changes in grip strength, the system will determine that there may be health problems that affect driving safety, and display the corresponding health prompt information on the display screen (such as "Your heart rate is abnormal, please pay attention to your physical condition"), and remind the driver through voice broadcast.

[0066] The human-computer interaction module feeds back information to the decision analysis module: the driver operates through the operation control devices (such as joysticks, buttons, etc.) provided by the human-computer interaction module, or responds using voice commands (for example, the driver confirms that the warning information has been received and the braking operation has been performed, the train speed has been adjusted, etc.), and the human-computer interaction module will feed back the actual operation of these drivers to the decision analysis module in real time. Based on this feedback information, the decision analysis module further judges the execution effect of the decision. If it is found that the actual execution situation does not meet expectations, the subsequent decision strategy can be adjusted in time to form a closed-loop control to ensure that the train operation status meets safety requirements.

[0067] The human-computer interaction module supports the driver to personalize the system settings (such as adjusting the volume of voice broadcasts, selecting the information category to be displayed first, etc.), and these settings will also be passed to the decision analysis module. The decision analysis module will appropriately adjust the priority and presentation of information output according to the driver's personalized needs in the subsequent data analysis, risk assessment and decision-making process, so that the system is more in line with the driver's usage habits and improves the efficiency and effectiveness of human-computer collaboration.

[0068] The decision analysis module quickly analyzes and processes train operation data and generates decision results, while the human-computer interaction module must quickly and accurately convey these contents to the driver. The two work closely together to minimize the time delay of information transmission and processing, ensuring that the driver can obtain key information and respond correctly at the first time. For example, when encountering an emergency (such as a pedestrian suddenly appearing in front of the train), the decision analysis module determines the risk and makes a braking decision in a very short time, and the human-computer interaction module immediately prompts the driver to take emergency braking through sound and light alarms and clear voice broadcasts to ensure that the train can stop in time and avoid accidents.

[0069] At the same time, the human-computer interaction module feeds back the driver's understanding and operation of the information to the decision analysis module, so that it can verify whether the decision is accurately executed. If the driver misunderstands a certain warning information or operates improperly, the decision analysis module can again issue supplementary prompts or adjusted decision suggestions through the human-computer interaction module, further enhancing the ability of the entire system to ensure safe train driving.

Claims

1. A train safety driving assistance system, comprising an operating environment monitoring module, a self-state detection module, a decision analysis module and a human-computer interaction module, characterized in that: The operating environment monitoring module is used to monitor the surrounding environment in real time when the train is running. The operating environment monitoring module can ensure driving safety, improve the operating efficiency of the train, and enhance the emergency response capability of the train in the event of unexpected situations. The operating environment monitoring module includes a track condition monitoring unit, a surrounding obstacle detection unit, and a meteorological condition perception unit; The self-state monitoring module is used to monitor the train's own situation in real time, and can monitor the train's speed and position to ensure the running order, prevent the risk of failure, optimize driving operations, reduce unnecessary parking and maintenance, and more quickly locate the area and cause of the failure. The self-state monitoring module includes a speed and position monitoring unit, a vehicle key component monitoring unit, and a train tilt angle monitoring unit; The decision analysis module is used to integrate a large amount of heterogeneous data in the operating environment monitoring module and the self-state detection module collected during train driving, and to sort them out according to unified data format and timestamp rules so that they can be effectively used by subsequent analysis and processing processes, and based on the integrated data, use the risk assessment model to analyze and quantitatively evaluate various potential risks faced by train operation. The decision analysis module includes a data collection and integration submodule, a risk assessment submodule, a decision-making submodule, a strategy execution and feedback submodule, a plan management submodule, and a learning and optimization submodule; The human-computer interaction module is used to convey the warning information, operation suggestions and various abnormal conditions monitored by the system generated by the decision analysis module to the driver in an intuitive manner in a timely manner to ensure that the driver's operations are accurately performed. The human-computer interaction module includes voice interaction, in-cabin display screen interaction, control handle and button interaction, touch screen interaction, camera monitoring interaction, biosensor monitoring interaction, smart watch / bracelet interaction and gesture interaction.

2. A train safety driving assistance system according to claim 1, characterized in that: The track condition monitoring unit includes sensor applications and algorithm analysis. The sensor applications include laser displacement sensors, ultrasonic sensors, and electromagnetic induction sensors. The algorithm analysis is based on a machine learning track defect identification algorithm and a track geometry parameter calculation algorithm.

3. A train safety driving assistance system according to claim 2, characterized in that: The surrounding obstacle detection unit includes sensor applications and algorithm analysis, the sensor applications include millimeter wave radar sensors, lidar sensors and high-definition cameras, and the algorithm analysis is based on a weighted average fusion algorithm.

4. A train safety driving assistance system according to claim 3, characterized in that: The meteorological condition sensing unit includes a wind speed and direction sensor, a rainfall sensor, and a visibility sensor. The meteorological condition sensing unit and the surrounding obstacle detection unit perform calculations based on a Kalman filter fusion algorithm.

5. A train safety driving assistance system according to claim 1, characterized in that: The speed and position monitoring unit includes a photoelectric speed sensor and a magnetoelectric speed sensor, and its position is monitored and positioned in real time based on a global satellite positioning system.

6. A train safety driving assistance system according to claim 5, characterized in that: The vehicle key component monitoring unit includes a vibration sensor, a temperature sensor, a pressure sensor and a brake pad wear detection.

7. A train safety driving assistance system according to claim 6, characterized in that: The train tilt angle monitoring unit includes a gyroscope sensor, an acceleration sensor, an inclination sensor and a dual-axis or tri-axis electronic compass, and is based on an angle calculation algorithm, a trigonometric function calculation algorithm, a temperature compensation and nonlinear correction algorithm, and a magnetic field component calculation and angle estimation algorithm.

8. A train safety driving assistance system according to claim 1, characterized in that: The decision analysis module transmits information to the human-computer interaction module, and the human-computer interaction module feeds back information to the decision analysis module.

9. A train safety driving assistance system according to claim 8, characterized in that: The voice interaction is based on a statistical n-gram language model, the gesture interaction is based on an edge detection algorithm, and the camera monitoring interaction collects the driver's facial image through a camera, based on a deep learning face detection algorithm.

Images