Safety system based on train autonomous perception and train safety protection method

Abstract

The invention provides a safety system based on train autonomous perception and a train safety protection method, and relates to the technical field of rail transit, the system comprises a sensor assembly, a calculation main control assembly and a logic main control assembly; each sensor module in the sensor assembly is used for sensing real-time operation parameters and real-time environment parameters of the train to obtain sensing parameters; each calculation main control board in the calculation main control assembly is used for carrying out obstacle detection, vehicle-mounted annunciator lamp position identification, speed measurement and positioning on a train according to sensing parameters sensed by the sensor modules correspondingly connected with the calculation main control boards, and according to an obstacle detection result, a vehicle-mounted annunciator lamp position identification result, a speed measurement result and a positioning result, a vehicle-mounted annunciator lamp position identification result and a vehicle-mounted annunciator lamp position identification result are obtained. Calculating to obtain a safety protection decision result; and the logic main control component is used for outputting a target safety protection decision result of the train according to the safety protection decision result calculated by each calculation main control board. According to the invention, accidents can be effectively prevented, and the safety and reliability of train operation are ensured.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, and in particular to a safety system and a train safety protection method based on train autonomous perception. Background Technology

[0002] Train equipment malfunctions mainly include signal failures, power supply failures, track failures, and vehicle malfunctions. Among these, track failures (such as obstacles on the track) have the greatest impact on train safety. A collision between a train and an obstacle can cause a major train accident and result in huge economic losses. Therefore, effectively detecting obstacles in the track area to ensure train safety is an important issue that urgently needs to be studied.

[0003] In related technologies, in scenarios involving manned trains, the detection of obstacles in the track area typically relies on the driver's personal judgment to ensure train safety. However, in fully automated train operation scenarios, since there is no resident driver, the detection of obstacles on the track is currently mainly passive, meaning that emergency braking is only triggered to protect the train in the event of a collision. Therefore, these technologies cannot effectively prevent accidents and cannot effectively guarantee the safe operation of trains.

[0004] Therefore, there is an urgent need to propose a safety system and train safety protection method based on train autonomous perception to solve the above-mentioned technical problems. Summary of the Invention

[0005] This invention provides a safety system and a train safety protection method based on train autonomous perception, which solves the shortcomings of the existing technology that passively detects track obstacles, cannot effectively prevent accidents, and cannot effectively ensure the safe operation of trains. It realizes automatic detection and early warning of track obstacles to improve the safety of train operation.

[0006] This invention provides a safety system based on train autonomous perception, including sensor components, computing main control components, and logic main control components; Each sensor module in the sensor assembly is used to sense the real-time operating parameters and real-time environmental parameters of the train to obtain the sensing parameters of the train. Each computing main control board in the computing main control component is used to perform obstacle detection, on-board signal light position recognition, speed measurement and positioning of the train based on the sensing parameters obtained by the sensor modules connected to each computing main control board, and to calculate the safety protection decision result of the train based on the obstacle detection result, on-board signal light position recognition result, speed measurement result and positioning result; The logic main control component is used to output the target safety protection decision result of the train based on the safety protection decision result calculated by each of the computing main control boards. The computing main control component includes a first computing main control board and a second computing main control board; The sensor assembly includes a first sensor module and a second sensor module; The first sensor module is connected to the first computing main control board, and the second sensor module is connected to the second computing main control board. The location of at least one sensor in the first sensor module is different from the location of at least one sensor in the second sensor module.

[0007] According to the present invention, a safety system based on train autonomous perception is provided, wherein the logic main control component includes a first logic main control board and a second logic main control board; The first logic main control board is used to determine a first candidate security protection decision result based on the comparison result between the security protection decision result calculated by the first computing main control board and the security protection decision result calculated by the second computing main control board, and to compare the first candidate security protection decision result with the second candidate security protection decision result in the second logic main control board to determine the target security protection decision result. The second candidate security protection decision result is obtained by the second logic main control board based on the comparison between the security protection decision result calculated by the first computing main control board and the security protection decision result calculated by the second computing main control board.

[0008] According to the present invention, a safety system based on train autonomous perception is provided, wherein the first computing main control board includes a first operating system, a first processing platform and a first deep learning module, and the second computing main control board includes a second operating system, a second processing platform and a second deep learning module; The first operating system and the second operating system are configured with different system architecture types, the first processing platform and the second processing platform are configured with different platform architecture types, and the first deep learning module and the second deep learning module are configured with different learning environments, learning frameworks and learning models.

[0009] According to a safety system based on train autonomous perception provided by the present invention, the first computing main control board is used to call the first deep learning module on the basis of the first operating system and the first processing platform, and perform obstacle detection, on-board signal light position recognition, speed measurement and positioning of the train according to the perception parameters obtained by the sensor module connected to the first computing main control board, and calculate the safety protection decision result of the train based on the obstacle detection result, the on-board signal light position recognition result, the speed measurement result and the positioning result.

[0010] According to the present invention, a safety system based on train autonomous perception is provided, wherein the first sensor module includes a first speed sensor, a first positioning sensor, a first environmental sensor disposed at the front end of the train, and a second environmental sensor disposed at the rear end of the train. The second sensor module includes a second speed sensor, a second positioning sensor, a third environmental sensor disposed at the rear end of the train, and a fourth environmental sensor disposed at the front end of the train. The first environmental sensor and the third environmental sensor are configured with different device models, and the second environmental sensor and the fourth environmental sensor are configured with different device models.

[0011] According to the present invention, a safety system based on train autonomous perception is provided, the safety system further comprising a power supply component; The power supply assembly includes multiple first power supply units and multiple second power supply units; The first power supply unit includes two power supplies that are mutually backups of each other, and the second power supply unit includes one independent power supply. Each of the first power supply units is connected to each of the logic main control boards in the logic main control assembly; Each of the second power supply units is connected to a corresponding computing main control board in the computing main control assembly.

[0012] According to the present invention, a safety system based on train autonomous perception is provided, the safety system further comprising a distance extension unit; The distance extension unit is used to extend the communication and power supply connections between each sensor in the sensor assembly and each computing main control board in the computing main control assembly.

[0013] According to the present invention, a safety system based on train autonomous perception is provided, the safety system further includes an interface configuration unit and multiple interface units; The interface configuration unit is used to determine a target interface unit among a plurality of interface units according to the communication type between the components in the security system, and to establish a communication connection between the components according to the target interface unit.

[0014] The present invention also provides a train safety protection method, the method being applied to a train autonomous perception-based safety system as described in any of the preceding claims, the method comprising: By using the various sensor modules in the sensor assembly, the real-time operating parameters and real-time environmental parameters of the train are sensed respectively, and the sensing parameters of the train are obtained. Using each main control board in the main control component, based on the sensing parameters obtained by the sensor modules connected to each main control board, the train is subjected to obstacle detection, on-board signal light position recognition, speed measurement and positioning. Based on the obstacle detection results, on-board signal light position recognition results, speed measurement results and positioning results, the safety protection decision results of the train are calculated. Using the logic main control component, based on the safety protection decision results calculated by each of the computing main control boards, the target safety protection decision result of the train is output. The computing main control component includes a first computing main control board and a second computing main control board; The sensor assembly includes a first sensor module and a second sensor module; The first sensor module is connected to the first computing main control board, and the second sensor module is connected to the second computing main control board. The location of at least one sensor in the first sensor module is different from the location of at least one sensor in the second sensor module.

[0015] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the train safety protection method as described above.

[0016] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the train safety protection method as described above.

[0017] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the train safety protection method as described above.

[0018] The safety system and train safety protection method based on autonomous train perception provided by this invention comprehensively perceive the real-time operating parameters and real-time environmental parameters of the train through multiple sensor modules. Multiple computing main control boards autonomously calculate the perceived parameters obtained by the sensor modules connected to each main control board. Furthermore, a logic main control component intelligently makes decisions based on the safety protection decision results calculated by the multiple computing main control boards. This effectively meets the SIL4 requirements for onboard active obstacle detection, onboard signal light position recognition, autonomous train speed measurement, and autonomous train positioning. It effectively overcomes the limitations of individual driver judgment and can achieve active obstacle detection and safety protection even without a resident driver, thereby effectively preventing accidents and ensuring the safety and reliability of train operation. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is one of the structural schematic diagrams of the safety system based on train autonomous perception provided by the present invention.

[0021] Figure 2 This is the second structural schematic diagram of the safety system based on train autonomous perception provided by the present invention.

[0022] Figure 3 This is a flowchart illustrating the train safety protection method provided by the present invention.

[0023] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention.

[0024] Figure label: 110: Sensor component; 120: Computational main control component; 130: Logic main control component. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0026] Fully automated train operation is a complex system engineering project involving various factors such as signaling, stations, platform screen doors, train operation organization, and vehicle depot configuration. It places high demands on the system's safety, reliability, and practicality. Train equipment failures mainly include signal failures, power supply failures, track failures, and vehicle failures. Among these, track failures (such as obstacles on the track) have the greatest impact on train safety. A collision between a train and an obstacle can cause a major train accident, resulting in significant economic losses. Therefore, effectively detecting obstacles in the track area to ensure train safety is a crucial issue that urgently needs to be addressed.

[0027] In related technologies, in scenarios involving manned trains, obstacle detection in the track area relies on the driver's personal judgment to achieve train safety. This method requires the driver to constantly monitor the track conditions and apply emergency braking upon detecting an emergency. Consequently, dangerous events outside the signal system's protection boundary depend entirely on the driver's emergency response, which is ineffective in preventing accidents and ensuring safe train operation. In scenarios where GoA3-4 level fully automated trains operate without a resident driver, obstacle detection is currently primarily passive, meaning emergency braking is only triggered upon impact. This also fails to effectively prevent accidents and ensure safe train operation.

[0028] Therefore, existing obstacle detection methods cannot prevent accidents and are one of the main safety hazards for trains operating in DTO (GoA3 level) and above modes. How to safely and effectively detect obstacles on long-distance tracks without a resident driver, in order to formulate corresponding safety protection decision-making strategies and achieve early warning or braking, is an urgent need for the safe operation of trains in DTO (Driverless Train Operation) and above modes.

[0029] In response, this embodiment provides a safety system based on train autonomous perception. By autonomously perceiving the real-time operating parameters and environmental parameters of the train, and autonomously calculating and making intelligent decisions based on the perceived information, it effectively meets the requirements of Safety Integrity Level 4 (SIL4) functions such as onboard active obstacle detection, onboard signal light position recognition, autonomous train speed measurement, and autonomous train positioning. It effectively eliminates the limitations of the driver's personal judgment, and can also achieve active obstacle detection and safety protection even when there is no resident driver on the fully automated train, thereby effectively preventing accidents and ensuring the safety and reliability of train operation. It effectively meets the urgent needs of safe train operation in DTO and above modes.

[0030] Figure 1 This is one of the structural schematic diagrams of the safety system based on train autonomous perception provided by the present invention; for example... Figure 1 As shown, the system includes a sensor assembly 110, a computing main control assembly 120, and a logic main control assembly 130; Each sensor module in the sensor assembly 110 is used to sense the real-time operating parameters and real-time environmental parameters of the train to obtain the sensing parameters of the train. Each computing main control board in the computing main control component 120 is used to perform obstacle detection, on-board signal light position recognition, speed measurement and positioning of the train based on the sensing parameters obtained by the sensor modules connected to each computing main control board, and to calculate the safety protection decision result of the train based on the obstacle detection result, on-board signal light position recognition result, speed measurement result and positioning result; The logic main control component 130 is used to output the target safety protection decision result of the train based on the safety protection decision result calculated by each of the computing main control boards.

[0031] like Figure 1 As shown, the safety system based on train autonomous perception includes at least a sensor assembly 110 and a motherboard assembly; the main control assembly adopts a two-layer design, which includes a computing main control assembly 120 and a logic main control assembly 130 respectively.

[0032] The sensor assembly is used for real-time parameter sensing of the train. It includes at least two sensor modules for sensing real-time operating parameters and real-time environmental parameters of the train from different locations. For example, it may include two sensor modules, one for sensing parameters at the front end and the other at the rear end of the train. This embodiment does not specifically limit this. Each sensor module includes multiple sensors for sensing different types of parameters, such as at least a speed sensor and a positioning sensor for measuring real-time operating parameters of the train, and an environmental sensor for measuring real-time operating environmental parameters of the train. This embodiment does not specifically limit this.

[0033] The main computing control component carries a SIL2-level safety function module, which is responsible for processing and analyzing data from sensor components and running train autonomous perception algorithms, including but not limited to onboard active obstacle detection, onboard signal light position recognition, train autonomous speed measurement, train autonomous positioning, and safety protection decisions. This embodiment does not specifically limit these aspects. The main computing control component consists of multiple main computing control boards, such as two series of main computing control boards, namely Series I and Series II. The main computing control boards of different series do not communicate with each other, but different sensor modules of the main computing control boards of different series are connected accordingly. The main computing control boards of different series adopt a heterogeneous design, such as different system structure types, platform structure types, and deep learning modules configured in different series of main computing control boards.

[0034] The logic master control component carries a SIL4 level safety function module, responsible for performing binary decision-making logic and information output on the safety protection decision results calculated by multiple computing master control boards. The logic master control component can include one or more systems of logic master control boards. When the logic master control component contains only one system of logic master control boards, the computing master control boards of different systems communicate with each other to input the safety protection decision results calculated by the computing master control boards of different systems into the logic master control board for final binary decision-making, thereby obtaining the target safety protection decision result for the train. When the logic master control component also contains multiple systems of logic master control boards, the computing master control boards of different systems communicate with each other to input the safety protection decision results calculated by the computing master control boards of different systems into each system's logic master control board for final binary decision-making, thereby obtaining the target safety protection decision result for the train.

[0035] Optionally, in the process of autonomous train perception for safety protection, the sensor modules in the sensor assembly can first perceive the real-time operating parameters and real-time environmental parameters of the train from different directions to obtain the perception parameters of the train from different directions.

[0036] Subsequently, each main control board in the main control assembly autonomously performs obstacle detection, onboard signal light position recognition, speed measurement, and positioning of the train based on the perception parameters obtained from the sensor modules connected to each main control board. Based on the obstacle detection results, onboard signal light position recognition results, speed measurement results, and positioning results, the train's safety protection decision is calculated. The obstacle detection, onboard signal light position recognition, speed measurement, positioning, and safety protection decision-making can be implemented by calling the train autonomous perception algorithm configured in each main control board. This train autonomous perception algorithm can be implemented using a deep learning algorithm model.

[0037] Subsequently, the logic control component, employing a binary decision logic, makes a final decision based on the safety protection decision results calculated by all the main control boards, outputting the train's final safety protection decision result, i.e., the target safety protection decision result. Based on this target safety protection decision result, the train is controlled to avoid obstacles or adjust its operating status in advance, thereby ensuring safe train operation. This achieves a system structure design with lower complexity, fully utilizing cutting-edge technologies such as deep learning, and meeting the SIL4 level requirements for onboard active obstacle detection, onboard signal light position recognition, autonomous train speed measurement, and autonomous train positioning. It effectively solves the safety hazards of existing passive obstacle detection methods being unable to prevent accidents, enabling the detection of long-distance track obstacles without a resident driver, so as to provide early warning or braking, thereby better improving the safety of train operation.

[0038] The system provided in this embodiment autonomously perceives the real-time operating parameters and environmental parameters of the train through multiple sensor modules. It also autonomously calculates the perceived parameters obtained by the sensor modules connected to each main control board through multiple computing main control boards. Furthermore, it intelligently makes decisions based on the safety protection decision results calculated by multiple computing main control boards through a logic main control component. This effectively meets the SIL4 requirements for onboard active obstacle detection, onboard signal light position recognition, autonomous train speed measurement, and autonomous train positioning. It effectively eliminates the limitations of the driver's personal judgment and can achieve active obstacle detection and safety protection even without a resident driver, thereby effectively preventing accidents and ensuring the safety and reliability of train operation.

[0039] In some embodiments, the logic master control component includes a first logic master control board and a second logic master control board; The computing main control component includes a first computing main control board and a second computing main control board; The first logic main control board is used to determine a first candidate security protection decision result based on the comparison result between the security protection decision result calculated by the first computing main control board and the security protection decision result calculated by the second computing main control board, and to compare the first candidate security protection decision result with the second candidate security protection decision result in the second logic main control board to determine the target security protection decision result. The second candidate security protection decision result is obtained by the second logic main control board based on the comparison between the security protection decision result calculated by the first computing main control board and the security protection decision result calculated by the second computing main control board.

[0040] Figure 2 This is the second structural schematic diagram of the safety system based on train autonomous perception provided by the present invention; as shown. Figure 2 As shown, the safety system based on train autonomous perception can adopt a two-out-of-two safety architecture. Accordingly, the logic main control component includes a first logic main control board (also known as the I-series logic main control board, or logic main control board A) and a second logic main control board (also known as the II-series logic main control board, or logic main control board B). The computing main control component includes a first computing main control board (also known as the I-series computing main control board, or computing main control board A) and a second computing main control board (also known as the II-series computing main control board, or computing main control board B).

[0041] Both logic control board A and logic control board B carry SIL4 level safety function modules, responsible for binary decision-making logic and information output. Both logic control boards A and B adopt the same design, such as using an x86 architecture processing platform and the Vxworks secure real-time operating system. Logic control boards A and B communicate with each other and with the two main computing control boards to obtain the train autonomous perception calculation results (i.e., safety protection decision results) calculated by the two main computing control boards. Logic control boards A and B each perform binary decision-making on the train autonomous perception calculation results obtained from the two main computing control boards, and then communicate with each other again to perform binary decision-making. Only when the results are consistent can the target safety protection decision result of the train be output externally. The specific implementation steps are as follows: The logic control board A determines the first candidate security protection decision based on the comparison between the security protection decision results calculated by the computing control board A and the security protection decision results calculated by the computing control board B. It then interacts with the logic control board B to obtain the second candidate security protection decision based on the comparison between the security protection decision results calculated by the logic control board A and the computing control board B. Finally, it compares the first candidate security protection decision with the second candidate security protection decision and outputs the target security protection decision.

[0042] Alternatively, the logic control board B compares the security protection decision results calculated by the computing control board A and the security protection decision results calculated by the computing control board B to determine a second candidate security protection decision result. It then interacts with the logic control board A to obtain the first candidate security protection decision result based on the comparison between the security protection decision results calculated by the logic control board A and the security protection decision results calculated by the computing control board B. Finally, it compares the first candidate security protection decision result with the second candidate security protection decision result to output the target security protection decision result.

[0043] The logic control board A and the computing control board A can be set on two different control boards, or they can be combined and set on the same control board. Similarly, the logic control board B and the computing control board B can be set on two different control boards, or they can be combined and set on the same control board.

[0044] For example, in a combined configuration, for each logic control board in logic control board A and logic control board B, a secure microcontroller unit (MCU) within its corresponding computing control board can be implemented internally. All MCUs utilize an Advanced RISC Machine (ARM) architecture processing platform and the QNX secure real-time operating system. Specifically, logic control board A implements the secure microcontroller unit (MCU) within computing control board A, and logic control board B implements the secure microcontroller unit (MCU) within computing control board B. Furthermore, the MCUs within both logic control boards utilize an Advanced RISC Machine (ARM) architecture processing platform and the QNX secure real-time operating system to further reduce system complexity.

[0045] For example, in the combined configuration, logic control board A and logic control board B can be implemented by sharing the core functions of the two computing control boards. That is, logic control board A integrates an ARM architecture processing platform such as Orin and an insecure real-time operating system such as Ubuntu to realize the functions of the secure microcontroller unit within computing control board A. Logic control board B integrates an ARM architecture processing platform such as KA200 and an insecure real-time operating system such as SylixOS to realize the functions of the secure microcontroller unit within computing control board B, thereby further reducing system complexity.

[0046] The system provided in this embodiment adopts a two-out-of-two security architecture, enabling the logic main control component and the computing main control component to work together to perform autonomous train perception and early warning protection, thus achieving high reliability and low complexity of the autonomous train perception safety system.

[0047] In some embodiments, the first computing main control board includes a first operating system, a first processing platform, and a first deep learning module, and the second computing main control board includes a second operating system, a second processing platform, and a second deep learning module; The first operating system and the second operating system are configured with different system architecture types, the first processing platform and the second processing platform are configured with different platform architecture types, and the first deep learning module and the second deep learning module are configured with different learning environments, learning frameworks and learning models.

[0048] like Figure 2As shown, both main control boards A and B carry SIL2-level safety function modules, responsible for processing sensor data and calculating train autonomous perception algorithms (including onboard active obstacle detection, onboard signal light position recognition, train autonomous speed measurement, and train autonomous positioning). To improve the system's fault tolerance, enhance its safety, and optimize its performance, main control boards A and B can adopt a heterogeneous design. This means that main control boards A and B have different operating system architectures, different processing platform architectures, and different learning environments, frameworks, and models for their deep learning modules.

[0049] For example, computing control board A uses an ARM architecture processing platform such as Orin and an insecure real-time operating system like Ubuntu; computing control board B uses an ARM architecture processing platform such as KA200 and an insecure real-time operating system like SylixOS. The two computing control boards do not communicate with each other, but each communicates with its respective logical control board to transmit the train's autonomous sensing calculation results data, avoiding data interference and conflicts. Furthermore, the operating environments of computing control boards A and B adopt a heterogeneous design. Operating environment A on computing control board A and operating environment B on computing control board B are configured separately, each using different types of deep learning environments and frameworks, and inferring different types of deep learning models. For example, operating environment A on computing control board A is configured with deep learning environments such as CUDA and TensorRT, and deep learning frameworks such as TensorFlow, and infers deep learning models such as YOLO; operating environment B on computing control board B is configured with deep learning environments such as OpenCL and Vulkan, and deep learning frameworks such as PyTorch, and infers deep learning models such as Faster R-CNN.

[0050] Accordingly, in some embodiments, the first computing main control board is used to call the first deep learning module on the basis of the first operating system and the first processing platform, and perform obstacle detection, on-board signal light position recognition, speed measurement and positioning of the train according to the perception parameters perceived by the sensor module connected to the first computing main control board, and calculate the safety protection decision result of the train based on the obstacle detection result, the on-board signal light position recognition result, the speed measurement result and the positioning result.

[0051] Optionally, the main control board A can call the deep learning model configured in the first deep learning module based on the first operating system and the first processing platform, so as to perform obstacle detection, on-board signal light position recognition, speed measurement and positioning of the train according to the perception parameters obtained by the sensor module connected to the main control board A. Based on the obstacle detection results, on-board signal light position recognition results, speed measurement results and positioning results, the safety protection decision results of the train can be calculated.

[0052] Similarly, the main control board B can call the deep learning model configured in the second deep learning module based on the second operating system and the second processing platform. According to the perception parameters obtained by the sensor module connected to the main control board B, it can perform obstacle detection, on-board signal light position recognition, speed measurement and positioning of the train. Based on the obstacle detection results, on-board signal light position recognition results, speed measurement results and positioning results, it can calculate the safety protection decision results of the train.

[0053] The system provided in this embodiment adopts a low-complexity system architecture design and makes full use of cutting-edge technologies such as deep learning to achieve functions such as onboard active obstacle detection, onboard signal light position recognition, autonomous train speed measurement, and autonomous train positioning that meet the SIL4 level requirements. By using a heterogeneous two-system computing main control board, it is equipped with operating systems of different system architecture types, processing platforms of different platform architecture types, and deep learning modules with different learning environments, learning frameworks, and learning models. Furthermore, the two systems are configured to not communicate with each other, which effectively improves the fault tolerance, safety, and performance of the train autonomous perception computing system, effectively ensuring the safety and reliability of train operation, and thus meeting the urgent needs of safe train operation in DTO and above modes.

[0054] In some embodiments, the sensor assembly includes a first sensor module and a second sensor module; The first sensor module is connected to the first computing main control board, and the second sensor module is connected to the second computing main control board. The location of at least one sensor in the first sensor module is different from the location of at least one sensor in the second sensor module.

[0055] like Figure 2 As shown, to enable the system to more comprehensively cover the monitoring area while ensuring data diversity to improve the overall system stability and accuracy, the sensor component may include a first sensor module and a second sensor module. The first sensor module is connected to the first main control board to provide corresponding sensing parameters to the first main control board. The second sensor module is connected to the second main control board to provide corresponding sensing parameters to the second main control board.

[0056] The first sensor module and the second sensor module each contain multiple sensors, such as at least a speed sensor and a positioning sensor for measuring real-time operating parameters of the train, and an environmental sensor for measuring real-time operating environmental parameters of the train. This embodiment does not specifically limit these.

[0057] In the first and second sensor modules, some sensors are distributed in the same physical space, while others are distributed differently. This distribution difference ensures that the system can acquire sensing parameters from multiple angles and locations, thereby providing more comprehensive monitoring coverage, improving the system's fault tolerance, and ultimately enhancing the safety of train operation.

[0058] For example, in some embodiments, the first sensor module includes a first speed sensor, a first positioning sensor, a first environmental sensor disposed at the front end of the train, and a second environmental sensor disposed at the rear end of the train. The second sensor module includes a second speed sensor, a second positioning sensor, a third environmental sensor disposed at the rear end of the train, and a fourth environmental sensor disposed at the front end of the train. The first environmental sensor and the third environmental sensor are configured with different device models, and the second environmental sensor and the fourth environmental sensor are configured with different device models.

[0059] like Figure 2 As shown, the first speed sensor (also called the head speed sensor) and the second speed sensor (also called the tail speed sensor) are two speed sensors with the same structure, which are located at different positions. That is, the head speed sensor and the tail speed sensor contain speed sensors of the same device model. The speed sensors here are not limited to millimeter-wave radar and inertial measurement units. This embodiment does not specifically limit them.

[0060] The first positioning sensor (also known as the head positioning sensor) and the second positioning sensor (also known as the tail positioning sensor) are two positioning sensors with the same structure, located at different positions. That is, the head positioning sensor and the tail positioning sensor contain positioning sensors of the same device model. The positioning sensors here include, but are not limited to, Beidou satellite positioning sensors. This embodiment does not specifically limit them.

[0061] The first environmental sensor (also known as the first-end environmental sensor A) and the third environmental sensor (also known as the first-end environmental sensor B) are two environmental sensors located at the same position and have heterogeneous structures. That is, the first-end environmental sensor A and the first-end environmental sensor B contain environmental sensors of different device models. For example, the first-end environmental sensor A includes, but is not limited to, visible light camera A, lidar A, and infrared camera A, while the first-end environmental sensor B includes, but is not limited to, visible light camera B, lidar B, and infrared camera B. Among them, the device models of visible light camera A and visible light camera B are different, the device models of lidar A and lidar B are different, and the device models of infrared camera A and infrared camera B are different.

[0062] Similarly, the second environmental sensor (also known as tail-end environmental sensor A) and the fourth environmental sensor (also known as tail-end environmental sensor B) are two environmental sensors located in the same position and have heterogeneous structures. That is, tail-end environmental sensor A and tail-end environmental sensor B contain environmental sensors of different device models. For example, tail-end environmental sensor A includes, but is not limited to, visible light camera A, lidar A, and infrared camera A, while tail-end environmental sensor B includes, but is not limited to, visible light camera B, lidar B, and infrared camera B. Among them, the device models of visible light camera A and visible light camera B are different, the device models of lidar A and lidar B are different, and the device models of infrared camera A and infrared camera B are different.

[0063] Accordingly, the first computing main control board communicates with the head-end speed sensor combination, the head-end positioning sensor combination, the head-end environmental sensor A combination, and the tail-end environmental sensor A combination, so as to use these sensors to perceive the real-time operating parameters and real-time environmental parameters of the train and obtain the train's multi-dimensional perception parameters.

[0064] The second main control board communicates with the tail-end speed sensor, the tail-end positioning sensor, the head-end environmental sensor B, and the tail-end environmental sensor B to use these sensors to sense the train's real-time operating parameters and real-time environmental parameters, thereby obtaining the train's multi-dimensional sensing parameters.

[0065] The system provided in this embodiment achieves comprehensive and multi-dimensional perception of train operation and environmental parameters by setting up homogeneous speed and positioning sensors and heterogeneous environmental sensors at the head and tail of the train, which significantly improves the safety and reliability of train operation.

[0066] In some embodiments, the security system further includes a power supply component; The power supply assembly includes multiple first power supply units and multiple second power supply units; The first power supply unit includes two power supplies that are mutually backups of each other, and the second power supply unit includes one independent power supply. Each of the first power supply units is connected to each of the logic main control boards in the logic main control assembly; Each of the second power supply units is connected to a corresponding computing main control board in the computing main control assembly.

[0067] Optionally, the security system also includes a power supply component; the number of first power supply units included in the power supply component is the same as the number of logic main control boards in the logic main control component, and the number of second power supply units included in the power supply component is the same as the number of computing main control boards in the computing main control component.

[0068] Each first power supply unit contains two power supplies that are mutually redundant, providing two redundant power supplies to the logic main control board connected to each first power supply unit; each second power supply unit contains one independent power supply, providing one independent power supply to the computing main control component connected to each second power supply unit.

[0069] The system provided in this embodiment adds a power supply component to the safety system to provide two redundant power supplies to the logic main control board and one independent power supply to the computing main control board. This improves system reliability, enhances fault tolerance, and meets high safety requirements, while also simplifying system design and optimizing energy utilization, thereby enhancing the overall system safety and operational stability.

[0070] In some embodiments, the safety system further includes a distance-extending unit; The distance extension unit is used to extend the communication and power supply connections between each sensor in the sensor assembly and each computing main control board in the computing main control assembly.

[0071] Optionally, a range extender unit can be further introduced into the security system to meet the needs of long-distance communication and power supply.

[0072] As a crucial component of the safety system, the distance extender unit's core function is to enhance and extend the connectivity between the sensor assembly and the main computing control assembly. Specifically, it can effectively extend the communication connection between the sensors located at the beginning and end of the sensor assembly and the various main control boards in the main computing control assembly through its built-in local communication module or an external communication module. This means that even if the sensors are deployed relatively far from the main control board, data transmission between them can remain stable and efficient with the assistance of the distance extender unit.

[0073] Besides extending communication connections, the range extender also plays a crucial role in extending power supply connections. In sensor networks, the stability and reliability of power supply are paramount. Through its built-in local power supply module or an external power supply module, the range extender can provide the necessary power support to sensors located far from the power interface, ensuring the stable operation of the entire sensor network.

[0074] Therefore, in practical applications, the introduction of the distance extension unit not only solves the communication and power supply problems caused by the excessive distance between the sensor and the main control board, but also further improves the overall performance and reliability of the safety system.

[0075] In some embodiments, the security system further includes an interface configuration unit and multiple interface units; The interface configuration unit is used to determine a target interface unit among a plurality of interface units according to the communication type between the components in the security system, and to establish a communication connection between the components according to the target interface unit.

[0076] The multiple interface units mentioned here include, but are not limited to, conventional interface units such as USB2.0, RS232, and RS485, as well as CAN, MVB, and TRDP interface units that can be expanded as needed. This embodiment does not specifically limit these.

[0077] Optionally, the security system also includes an interface configuration unit and multiple interface units; wherein the interface configuration unit can determine the interface unit that matches the communication type between each component in the security system as the target interface unit, and can establish communication connections between each component based on the target interface unit, so as to improve the connection efficiency and accuracy between each component and enhance the compatibility and maintainability of the system.

[0078] The train safety protection method provided by the present invention is described below. The train safety protection method described below can be referred to in correspondence with the safety system based on train autonomous perception described above.

[0079] Figure 3 This is a flowchart illustrating the train safety protection method provided by the present invention.

[0080] The main implementer of this method is the train autonomous perception-based safety system provided in the above embodiments; such as Figure 3 As shown, the method includes steps 310, 320 and 330.

[0081] Step 310: Using the sensor modules in the sensor assembly, the real-time operating parameters and real-time environmental parameters of the train are sensed respectively to obtain the sensing parameters of the train. Step 320: Using each main control board in the main control assembly, based on the sensing parameters obtained by the sensor modules connected to each main control board, the train is subjected to obstacle detection, on-board signal light position recognition, speed measurement and positioning. Based on the obstacle detection results, on-board signal light position recognition results, speed measurement results and positioning results, the safety protection decision results of the train are calculated. Step 330: Using the logic main control component, based on the safety protection decision results calculated by each of the computing main control boards, the target safety protection decision result of the train is output.

[0082] Optionally, in the process of autonomous train perception for safety protection, the sensor modules in the sensor assembly can first perceive the real-time operating parameters and real-time environmental parameters of the train from different directions to obtain the perception parameters of the train from different directions.

[0083] Subsequently, each main control board in the main control assembly autonomously detects obstacles, identifies onboard signal light positions, measures speed, and locates the train based on the sensing parameters obtained from the sensor modules connected to each main control board. Based on the obstacle detection results, onboard signal light position identification results, speed measurement results, and location results, the train's safety protection decision is calculated. The obstacle detection, onboard signal light position identification, speed measurement and location, as well as the safety protection decision, can be implemented by calling the train autonomous perception algorithm configured in each main control board. This train autonomous perception algorithm can be implemented using a deep learning algorithm model.

[0084] Subsequently, the logic control component, employing a binary decision logic, makes a final decision based on the safety protection decision results calculated by all the main control boards, outputting the train's final safety protection decision result, i.e., the target safety protection decision result. Based on this target safety protection decision result, the train is controlled to avoid obstacles or adjust its operating status in advance, thereby ensuring safe train operation. This achieves a system structure design with lower complexity, fully utilizing cutting-edge technologies such as deep learning to realize functions such as onboard active obstacle detection, onboard signal light position recognition, autonomous train speed measurement, and autonomous train positioning that meet SIL4 level requirements. It effectively solves the safety hazards of existing passive obstacle detection methods being unable to prevent accidents, and enables the detection of long-distance track obstacles without a resident driver, so as to provide early warning or braking, and better improve the safety of train operation.

[0085] The method provided in this embodiment autonomously perceives the real-time operating parameters and real-time environmental parameters of the train through multiple sensor modules. It also autonomously calculates the perceived parameters obtained by the sensor modules connected to each main control board through multiple computing main control boards. Furthermore, it intelligently makes decisions based on the safety protection decision results calculated by the multiple computing main control boards through a logic main control component. This effectively meets the SIL4 requirements for onboard active obstacle detection, onboard signal light position recognition, autonomous train speed measurement, and autonomous train positioning. It effectively overcomes the limitations of the driver's personal judgment and can achieve active obstacle detection and safety protection even without a resident driver, thereby effectively preventing accidents and ensuring the safety and reliability of train operation.

[0086] The method provided by this invention is executed based on the above-described system embodiments. For specific processes and details, please refer to the above embodiments, which will not be repeated here.

[0087] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4 As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other through the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute a train safety protection method, which includes: using each sensor module in the sensor assembly to sense the real-time operating parameters and real-time environmental parameters of the train to obtain the perceived parameters of the train; using each main control board in the main control assembly to perform obstacle detection, on-board signal light position recognition, speed measurement, and positioning of the train based on the perceived parameters sensed by the sensor modules connected to each main control board, and calculating the safety protection decision result of the train based on the obstacle detection result, on-board signal light position recognition result, speed measurement result, and positioning result; and using the logic main control assembly to output the target safety protection decision result of the train based on the safety protection decision result calculated by each main control board.

[0088] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0089] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the train safety protection method provided by the above methods. The method includes: using each sensor module in the sensor assembly to sense the real-time operating parameters and real-time environmental parameters of the train to obtain the sensing parameters of the train; using each main control board in the main control assembly to perform obstacle detection, on-board signal light position recognition, speed measurement, and positioning of the train based on the sensing parameters sensed by the sensor modules connected to each main control board, and calculating the safety protection decision result of the train based on the obstacle detection result, on-board signal light position recognition result, speed measurement result, and positioning result; and using a logic main control assembly to output the target safety protection decision result of the train based on the safety protection decision result calculated by each main control board.

[0090] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the train safety protection method provided by the above-described methods. The method includes: using each sensor module in the sensor assembly to sense the real-time operating parameters and real-time environmental parameters of the train to obtain the perceived parameters of the train; using each main control board in the main control assembly to perform obstacle detection, on-board signal light position recognition, speed measurement, and positioning of the train based on the perceived parameters sensed by the sensor modules connected to each main control board, and calculating the safety protection decision result of the train based on the obstacle detection result, on-board signal light position recognition result, speed measurement result, and positioning result; and using a logic main control assembly to output the target safety protection decision result of the train based on the safety protection decision result calculated by each main control board.

[0091] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0092] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A safety system based on train autonomous perception, characterized in that, It includes sensor components, computing main control components, and logic main control components; Each sensor module in the sensor assembly is used to sense the real-time operating parameters and real-time environmental parameters of the train to obtain the sensing parameters of the train. Each computing main control board in the computing main control component is used to perform obstacle detection, on-board signal light position recognition, speed measurement and positioning of the train based on the sensing parameters obtained by the sensor modules connected to each computing main control board, and to calculate the safety protection decision result of the train based on the obstacle detection result, on-board signal light position recognition result, speed measurement result and positioning result; The logic main control component is used to output the target safety protection decision result of the train based on the safety protection decision result calculated by each of the computing main control boards. The computing main control component includes a first computing main control board and a second computing main control board; The sensor assembly includes a first sensor module and a second sensor module; The first sensor module is connected to the first computing main control board, and the second sensor module is connected to the second computing main control board. The location of at least one sensor in the first sensor module is different from the location of at least one sensor in the second sensor module.

2. The safety system based on train autonomous perception according to claim 1, characterized in that, The logic master control component includes a first logic master control board and a second logic master control board; the first logic master control board is used to determine a first candidate security protection decision result based on the comparison result between the security protection decision result calculated by the first computing master control board and the security protection decision result calculated by the second computing master control board, and to compare the first candidate security protection decision result with the second candidate security protection decision result in the second logic master control board to determine the target security protection decision result. The second candidate security protection decision result is obtained by the second logic main control board based on the comparison between the security protection decision result calculated by the first computing main control board and the security protection decision result calculated by the second computing main control board.

3. The safety system based on train autonomous perception according to claim 2, characterized in that, The first computing main control board includes a first operating system, a first processing platform, and a first deep learning module; the second computing main control board includes a second operating system, a second processing platform, and a second deep learning module. The first operating system and the second operating system are configured with different system architecture types, the first processing platform and the second processing platform are configured with different platform architecture types, and the first deep learning module and the second deep learning module are configured with different learning environments, learning frameworks and learning models.

4. The safety system based on train autonomous perception according to claim 3, characterized in that, The first computing main control board is used to call the first deep learning module on the basis of the first operating system and the first processing platform, and to perform obstacle detection, on-board signal light position recognition, speed measurement and positioning of the train according to the perception parameters obtained by the sensor module connected to the first computing main control board. Based on the obstacle detection results, on-board signal light position recognition results, speed measurement results and positioning results, the safety protection decision results of the train are calculated.

5. The safety system based on train autonomous perception according to claim 1, characterized in that, The first sensor module includes a first speed sensor, a first positioning sensor, and a first environmental sensor disposed at the front end of the train, and a second environmental sensor disposed at the rear end of the train. The second sensor module includes a second speed sensor, a second positioning sensor, a third environmental sensor disposed at the rear end of the train, and a fourth environmental sensor disposed at the front end of the train. The first environmental sensor and the third environmental sensor are configured with different device models, and the second environmental sensor and the fourth environmental sensor are configured with different device models.

6. The safety system based on train autonomous perception according to any one of claims 1-5, characterized in that, The safety system also includes a power supply component; The power supply assembly includes multiple first power supply units and multiple second power supply units; The first power supply unit includes two power supplies that are mutually backups of each other, and the second power supply unit includes one independent power supply. Each of the first power supply units is connected to each of the logic main control boards in the logic main control assembly; Each of the second power supply units is connected to a corresponding computing main control board in the computing main control assembly.

7. The safety system based on train autonomous perception according to any one of claims 1-5, characterized in that, The safety system also includes a distance-extending unit; The distance extension unit is used to extend the communication and power supply connections between each sensor in the sensor assembly and each computing main control board in the computing main control assembly.

8. The safety system based on train autonomous perception according to any one of claims 1-5, characterized in that, The security system also includes an interface configuration unit and multiple interface units; The interface configuration unit is used to determine a target interface unit among a plurality of interface units according to the communication type between the components in the security system, and to establish a communication connection between the components according to the target interface unit.

9. A train safety protection method, characterized in that, The method is applied to the train autonomous perception-based safety system as described in any one of claims 1 to 8, and the method includes: By using the various sensor modules in the sensor assembly, the real-time operating parameters and real-time environmental parameters of the train are sensed respectively, and the sensing parameters of the train are obtained. Using each main control board in the main control component, based on the sensing parameters obtained by the sensor modules connected to each main control board, the train is subjected to obstacle detection, on-board signal light position recognition, speed measurement and positioning. Based on the obstacle detection results, on-board signal light position recognition results, speed measurement results and positioning results, the safety protection decision results of the train are calculated. Using the logic main control component, based on the safety protection decision results calculated by each of the computing main control boards, the target safety protection decision result of the train is output. The computing main control component includes a first computing main control board and a second computing main control board; The sensor assembly includes a first sensor module and a second sensor module; The first sensor module is connected to the first computing main control board, and the second sensor module is connected to the second computing main control board. The location of at least one sensor in the first sensor module is different from the location of at least one sensor in the second sensor module.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the train safety protection method as described in claim 9.

Images