Vehicle-mounted calculation unit of train operation environment sensing system
By designing a flexible combination of power supply boards and computing boards for the onboard computing unit of the train operation environment perception system, the problems of existing technologies being unable to identify foreign object intrusion in real time and computing equipment being unable to adapt to different vehicle models and speed levels have been solved, achieving the effect of real-time detection and stable operation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA ACADEMY OF RAILWAY SCI CORP LTD
- Filing Date
- 2025-12-08
- Publication Date
- 2026-04-28
AI Technical Summary
The existing obstacle detection systems for rail transit trains mainly rely on passive collision switches, which cannot identify foreign object intrusions in real time, posing a safety hazard. Furthermore, the data processing requirements of high-frequency sensors are complex, and existing computing power equipment cannot flexibly adapt to the differentiated needs of different train models and speed levels.
An onboard computing unit for a train operation environment perception system was designed. By combining power supply boards, computing boards, and gigabit Ethernet data exchange boards, the computing power can be flexibly covered to adapt to the perception needs of different train models and speed levels. It also interacts with the train control system through communication boards and uses a fan plate for precise heat dissipation.
It enables real-time detection of foreign object intrusion, improves train operation safety, meets the computing power requirements of different train models and speed levels, is compatible with mainstream train control system network interfaces, and ensures stable operation of the equipment under high load.
Smart Images
Figure CN121929211A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of assisted automatic driving of rail transit vehicles, specifically an onboard computing unit for a train operation environment perception system. Background Technology
[0002] Foreign object intrusion is a significant factor endangering the safety of rail transit trains. Such intrusions are generally random and sudden, occurring in various ways such as: rockfalls from landslides, unauthorized pedestrians entering open areas, track maintenance personnel and their equipment, and accidental vehicle intrusions. Traditional track inspection methods cannot eliminate these hazards. Currently, the rail transit industry primarily relies on drivers to manually observe and detect foreign objects within the operating clearance, assess their impact on train safety, and then manually execute emergency braking and other protective actions. This method heavily relies on the driver's expertise, and short-term human judgment and operation are prone to misjudgment and delays, posing certain safety risks. In addition, traditional rail transit train obstacle detection also includes a passive obstacle detection mode. This mode relies on multiple collision switches installed at the ends of the vehicle to feed back the status to the vehicle control system. When a collision occurs, the collision switches are activated, and the train network control system collects the collision signal, prompting the vehicle to perform emergency protective operations. Passive obstacle detection can only passively trigger system functions after a collision to weaken the longitudinal impact and mitigate its consequences.
[0003] With the continuous development and application of intelligent technologies such as LiDAR, image sensor, and machine learning, sensors can collect the three-dimensional environmental status within the clearance area in real time. Data modeling can then be used to establish a vehicle operation clearance model, and software algorithms can detect and identify information about the train's operating environment in real time. This provides crucial data to support decision-making for train assistance and automatic driving systems. Higher-level assistance / automatic driving systems also have corresponding requirements regarding the train's autonomous perception capabilities of the operating environment. For example, Level 6 assisted driving systems require trains to use intelligent sensing technology to perceive the color light status of ground signals, detect the presence of obstacles on the track ahead, and take appropriate safety measures to ensure safe operation.
[0004] However, massive amounts of high-frequency sensor data (video data, radar point cloud data, infrared thermal maps, inertial navigation data, etc.) require real-time decoding and processing. Large-scale and complex machine learning models require high-performance computing equipment. Furthermore, different types and speed levels of rail transit trains have varying timeliness requirements for environmental perception tasks. Therefore, designing an onboard computing unit for train environmental perception that can flexibly expand computing resources according to the needs of different application scenarios has significant practical value. Summary of the Invention
[0005] To address the problems in the existing technology, this application provides an on-board computing unit for a train operation environment perception system, which can adapt to the differentiated needs of different train models and speed levels in terms of computing power for operation environment perception. By adjusting the combination of power supply boards and computing power boards, flexible coverage of the overall computing power can be achieved, enabling data interaction between the perception system and the on-board control system.
[0006] To solve the above-mentioned technical problems, this application provides the following technical solution:
[0007] This application provides an onboard computing unit for a train operation environment perception system, including a power supply board, a communication board, a computing power board, and a gigabit Ethernet data exchange board.
[0008] The power supply board provides power to the computing board and the gigabit Ethernet data exchange board; the gigabit Ethernet data exchange board provides a multi-source sensor data access channel for the multi-source sensor group; the computing board generates intrusion detection control commands based on the multi-source sensor data received from the gigabit Ethernet data exchange board, and sends the intrusion detection control commands to the train control and management system through the communication board.
[0009] Furthermore, the Gigabit Ethernet data exchange board receives the multi-source sensor data via Gigabit Ethernet and transmits the multi-source sensor data to the computing power board via Gigabit Ethernet.
[0010] Furthermore, the computing power board includes a main computing power card; the main computing power card includes a sensor data acquisition and distribution module, an intrusion detection module, a decision control module, a network interaction module, and a human-computer interaction display module;
[0011] Specifically, the sensor data acquisition and publishing module acquires the multi-source sensor data and publishes it to the intrusion detection module; the intrusion detection module performs intrusion detection on the multi-source sensor data and sends the intrusion detection result to the decision control module; the decision control module makes a decision based on the intrusion detection result and sends the generated decision result and the intrusion detection control command to the human-machine interaction display module and the network interaction module, respectively; the network interaction module sends the intrusion detection control command to the train control and management system through the communication board.
[0012] Furthermore, the power supply board includes a main power supply card; the input terminal of the main power supply card includes two 110V DC power supplies; the output terminal of the main power supply card includes one 5V DC power supply and one 24V DC power supply; wherein, the 5V DC power supply is used to power the communication board; and the 24V DC power supply is used to power the computing board and the gigabit Ethernet data exchange board.
[0013] Furthermore, the computing power board also includes a first expansion computing power board; the power supply board also includes a first expansion power supply board; wherein, the input terminal of the first expansion power supply board includes two 110V DC power supplies; the output terminal of the first expansion power supply board includes one 24V DC power supply for powering the first expansion computing power board; the first expansion computing power board is used to perform intrusion detection on the multi-source sensor data using an intrusion detection model to enhance the computing power of the main computing power board.
[0014] Furthermore, the computing power board also includes a second expansion computing power board; the power supply board also includes a second expansion power supply board; wherein, the input terminal of the second expansion power supply board includes two 110V DC power supplies; the output terminal of the second expansion power supply board includes one 24V DC power supply for powering the second expansion computing power board; the second expansion computing power board is used to perform intrusion detection on the multi-source sensor data using an intrusion detection model to enhance the computing power of the main computing power board.
[0015] Furthermore, the main computing power card also includes an expansion card data management module; wherein, the expansion card data management module publishes the multi-source sensor data corresponding to the first expansion computing power card and the second expansion computing power card to the first expansion computing power card and the second expansion computing power card respectively; the first expansion computing power card and the second expansion computing power card send the intrusion detection results generated by each to the expansion card data management module.
[0016] Furthermore, the communication board includes a multi-functional vehicle bus communication board and a train real-time data protocol communication board; if the train control network is a multi-functional vehicle bus network, it is accessed through the multi-functional vehicle bus communication board; if the train control network is a train real-time data protocol communication network, it is accessed through the train real-time data protocol communication board.
[0017] Furthermore, the multi-source sensing data includes millimeter-wave radar data, visible light camera data, infrared camera data, lidar data, and combined inertial navigation data.
[0018] Furthermore, the onboard computing unit of the train operation environment perception system also includes a fan disk for providing precise heat dissipation support for the computing board.
[0019] As described above, the train operation environment perception computing unit with flexible expandable physical computing resources provided in this application can adapt to the differentiated computing power requirements of trains of different models and speed levels in terms of operation environment perception. It can achieve flexible coverage of the overall computing power by adjusting the combination of power supply boards and computing power boards, meeting the computing power requirements of most current train operation environment perception scenarios. Furthermore, the train operation environment perception computing unit with flexible expandable physical computing resources provided in this application is compatible with mainstream Train Control and Management System (TCMS) network interface types, including MVB and TRDP interfaces, allowing flexible access to the onboard network control system and easy data interaction between the perception system and the onboard control system. Addressing the issue of excessive heat generation from high-performance computing boards, a vertical fan plate is also designed, which uses active bottom-up air-cooled circulating airflow to precisely control the temperature of the computing unit within a reasonable range. Attached Figure Description
[0020] To more clearly illustrate the technical solutions in the embodiments of the present 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0021] Figure 1 The main view of the vehicle-mounted computing unit (250 TOPS computing power) in this embodiment of the application.
[0022] Figure 2 Main view of the vehicle-mounted computing unit (500 TOPS computing power) in this embodiment of the application.
[0023] Figure 3 Main view of the vehicle-mounted computing unit (750 TOPS computing power) in this embodiment of the application.
[0024] Figure 4 The power supply schematic diagram of the vehicle-mounted computing unit in this embodiment;
[0025] Figure 5 Communication topology diagram of the vehicle-mounted computing unit in this embodiment;
[0026] Figure 6 This application provides a schematic diagram of the data flow of the computing board in the embodiments of this application. Detailed Implementation
[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0028] In one embodiment, see Figure 1 , Figure 2 and Figure 3 In order to adapt to the different needs of trains of different models and speed levels in terms of computing power for sensing the operating environment, this application provides an on-board computing unit for a train operating environment sensing system, including a power board, a communication board, a computing board and a gigabit Ethernet data exchange board, by adjusting the combination of power board and computing board to achieve flexible coverage of the whole machine's computing power and realize data interaction between the sensing system and the on-board control system.
[0029] The power supply board provides power to the computing board and the gigabit Ethernet data exchange board; the gigabit Ethernet data exchange board provides a multi-source sensor data access channel for the multi-source sensor group; the computing board generates intrusion detection control commands based on the multi-source sensor data received from the gigabit Ethernet data exchange board, and sends the intrusion detection control commands to the train control and management system through the communication board.
[0030] It is understandable that, in order to solve the problems existing in the prior art, this application embodiment, combined with the actual application conditions of train operation environment perception, provides a train operation environment perception on-board computing unit (also called the train operation environment perception system on-board computing unit, or simply the computing unit) with flexible expandable computing resources, aiming to provide differentiated hardware computing power support for the application of larger-scale models. Physically, the computing unit includes three independently powered power cards PWR1-1, PWR1-2, and PWR2, an MVB communication board, a TRDP communication board, a main computing power board SL-1, a gigabit Ethernet data exchange board JH, two flexibly configurable computing power boards SL-2 and SL-3, and a 1U fan plate at the bottom of the host.
[0031] Power supply board PWR2: Input includes two redundant 110V DC power supplies. Output includes one 5V DC output and one 24V DC output. The 5V output powers the MVB and TRDP communication boards, while the 24V output powers the SL-1 board, JH board, and the fan plate at the bottom of the chassis. The front panel of the board connects to the train's 110V power supply via a power connector.
[0032] MVB and TRDP boards: This computing unit is compatible with most mainstream train control network types and can be flexibly connected based on the type of the train's control network bus: if the train control network is an MVB bus network, it can be connected via an MVB board; if the train control network is a TRDP network, it can be connected via a TRDP board. Both the MVB and TRDP boards interact with the main computing board SL-1 via the bus, communicating with the train network according to the standard MVB or TRDP communication protocols, respectively.
[0033] The SL-1 main computing board is the foundational computing resource within this computing unit. Its front panel includes two Gigabit Ethernet ports and one HDMI video signal port, facilitating the projection of sensor data onto the display screen. The board contains an independent CPU, memory, high-capacity storage media, and a highly integrated GPU module, providing parallel computing interface resources for train operation environment sensing.
[0034] JH Gigabit Ethernet Data Exchange Board: Its panel supports 7 independent access Gigabit Ethernet channels, which can provide multiple data access channels for train operation environment sensing sensor groups (such as radar, cameras, infrared, etc.), and connect each sensor to the main computing card and expansion computing card through the backplane Gigabit Ethernet, thereby realizing data interconnection.
[0035] Power supply card PWR1-1 and expansion computing card SL-3: As the first set of flexibly configurable computing resources, they can be selected according to computing power requirements. The power supply card PWR1-1 has two redundant 110V DC power inputs and one 24V DC power output. This power supply only powers the expansion computing card SL-3. The expansion computing card SL-3 includes two Gigabit Ethernet ports on the front panel and one on the back panel. The Gigabit Ethernet port on the back panel connects to the JH board, enabling data acquisition from sensors and interaction with the main computing card. The SL-3 board also carries the same hardware computing resources as the main computing card, allowing high-resource-intensive tasks from the main computing card to be distributed to this card, thereby improving the overall computing power and real-time performance of the computing unit.
[0036] Power card PWR1-2 and expansion computing power card SL-2: The working principle is the same as that of power card PWR1-1 and expansion computing power card SL-3. As a second set of computing power resources, they can be selected according to computing power requirements.
[0037] Fan plate: Provides precise heat dissipation support for the computing unit through bottom-up vertical airflow. This design effectively prevents heat buildup on high-frequency load modules such as computing boards due to heavy workloads, and prevents the overall performance of the computing unit from being affected by inadequate heat dissipation, ensuring that the device maintains a stable and efficient working state under high load.
[0038] As described above, the train operation environment perception computing unit with flexibly expandable physical computing resources provided in this application can adapt to the differentiated computing power requirements of trains of different models and speed levels in terms of operation environment perception. It can achieve flexible coverage of the overall computing power by adjusting the combination of power supply boards and computing power boards, meeting the computing power requirements of most current train operation environment perception scenarios. Furthermore, the train operation environment perception computing unit with flexibly expandable physical computing resources provided in this application is compatible with the mainstream Train Control and Management System (TCMS) network interface types in the industry, including MVB and TRDP interfaces, flexibly connecting to the onboard network control system and easily realizing data interaction between the perception system and the onboard control system. Addressing the issue of excessive heat generation from high-performance computing boards, it also features a vertical fan plate, which uses active bottom-up air-cooled circulating airflow to precisely control the temperature of the computing unit within a reasonable range.
[0039] In one embodiment, see [reference needed]. Figure 1 , Figure 2 and Figure 3 The Gigabit Ethernet data exchange board receives the multi-source sensor data via Gigabit Ethernet and transmits the multi-source sensor data to the computing power board via Gigabit Ethernet.
[0040] In one embodiment, see [reference needed]. Figure 1 , Figure 2 and Figure 3 The power supply board includes a main power supply card; the input terminal of the main power supply card includes two 110V DC power supplies; the output terminal of the main power supply card includes one 5V DC power supply and one 24V DC power supply; wherein, the 5V DC power supply is used to power the communication board; and the 24V DC power supply is used to power the computing board and the gigabit Ethernet data exchange board.
[0041] In one embodiment, see [reference needed]. Figure 1 , Figure 2 and Figure 3 The communication board includes a multi-functional vehicle bus communication board and a train real-time data protocol communication board; if the train control network is a multi-functional vehicle bus network, it is accessed through the multi-functional vehicle bus communication board; if the train control network is a train real-time data protocol communication network, it is accessed through the train real-time data protocol communication board.
[0042] In one embodiment, see [reference needed]. Figure 1 , Figure 2 and Figure 3The multi-source sensing data includes millimeter-wave radar data, visible light camera data, infrared camera data, lidar data, and combined inertial navigation data.
[0043] In one embodiment, see [reference needed]. Figure 1 , Figure 2 and Figure 3 The on-board computing unit of the train operation environment perception system also includes a fan disk for providing precise heat dissipation support for the computing board.
[0044] Understandable, Figure 1 , Figure 2 , Figure 3 These are three typical schematic diagrams of the on-board computing unit host according to embodiments of this application, which can adapt to application scenarios with computing power requirements of 250TOPS, 500TOPS, and 750TOPS, respectively. It can be installed in a standard 4U chassis with a width of 84HP. The chassis includes the enclosure and a 1U fan plate at the bottom, and contains seven 3U boards. The computing unit's enclosure protection level meets IP20, and its electromagnetic radiation complies with the relevant standards of GB / T 25119-2021. The device's start-up ambient temperature should be between -40℃ and 70℃ (short-term 85℃), and the device's operating temperature should also be between -40℃ and 70℃ (short-term 85℃), meeting the needs of trains operating in different physical environments.
[0045] The rated input voltage of this computing unit is 110V, and the power supply meets the requirements of DC77-DC137.5V. Depending on the computing power requirements, the typical power consumption of the host device is 100W, 200W, and 300W (corresponding to three implementation schemes with computing power of 250TOPS, 500TOPS, and 750TOPS respectively), and the cooling method is active fan cooling. The shock and vibration resistance of the host device meets the relevant standards of GB / T 21563-2018.
[0046] The characteristics of the fan plate and various main body circuit boards are as follows:
[0047] Fan plate:
[0048] 1) Input: Input voltage is DC24V, power consumption is 5W;
[0049] 2) The air supply volume is 100 CFM and the noise level is less than 44.5 dB.
[0050] Power card PWR2:
[0051] 1) Input: 2 redundant DC100V channels on the front panel; Output: Dual DC5V + DC24V outputs on the back panel, with a maximum power of 150W;
[0052] 2) The 5V output supplies power to the MVB board and the TRDP board respectively;
[0053] 3) The 24V output supplies power to the computing power card SL-1, the data exchange board JH, and the fan plate respectively.
[0054] Communication network card MVB:
[0055] 1) Input: Input voltage is DC5V, power consumption is 7.5W;
[0056] 2) The front panel has two standard DB9 MVB connectors, which can be connected to an MVB-type TCMS network to establish network communication between the on-board computing unit host and the train control network.
[0057] 3) The backplane uses a standard ISA interface to interact with the SL-1 board. Through this interface, standard MVB protocol data format communication can be realized between the computing unit host and the TCMS network port.
[0058] TRDP (Traffic Network Interface Card):
[0059] 1) Input: Input voltage is DC5V, power consumption is 7.5W;
[0060] 2) The front panel has two standard M12 Ethernet connectors, which can be connected to a TRDP type TCMS network to establish network communication between the on-board computing unit host and the train control network.
[0061] 3) The backplane uses a standard ISA interface to interact with the SL-1 board. Through this interface, standard TRDP protocol data format communication can be realized between the computing unit host and the TCMS network port.
[0062] Gigabit Ethernet Data Switching Card JH:
[0063] 1) Input: Input voltage is DC24V, power consumption is 10W;
[0064] 2) The front panel includes 7 1000M Ethernet interfaces, which can be connected to sensors such as visible light cameras, lidar, millimeter-wave radar, infrared, and inertial navigation, thereby establishing a communication interface between the computing unit host and the sensor group; the back panel includes 3 1000M Ethernet interfaces, which connect the main computing power card SL-1 and the expansion computing power cards SL-2 and SL-3 respectively, and can be used for business information transmission channels and data synchronization between different computing power nodes;
[0065] Power supply board PWR1-1 / PWR1-2:
[0066] 1) Input: 2 redundant DC100V channels on the front panel; Output: DC24V output from the back panel, with a maximum power of 100W.
[0067] 2) Power card PWR1-1 directly supplies power to expansion computing card SL-3, and power card PWR1-2 directly supplies power to expansion computing card SL-2.
[0068] Figure 4 This is a power supply schematic diagram of an embodiment of this application. To meet the requirement of flexible expansion of computing power, the computing power resources in this embodiment are divided into three parts: the main computing power node SL-1 and two expansion computing power nodes SL-2 and SL-3. SL-3 is independently powered by PWR1-1, and SL-2 is independently powered by PWR1-2. All three power supply boards are connected to the train's 110V DC power supply. PWR2 converts DC110V to two outputs, DC24V and DC5V, to power different components in the host computer. PWR1-1 and PWR1-2 convert DC110V to DC24V to independently power the two expansion computing power cards.
[0069] Figure 5 This is a communication topology diagram of an embodiment of this application. The MVB and TRDP communication network cards interact with the main computing card SL-1 via the ISA bus on the backplane, enabling flexible access to MVB or TRDP train control networks. Sensor data from sensors used to detect the train's operating environment, such as millimeter-wave radar, visible light cameras, infrared cameras, lidar, and integrated inertial navigation systems, are transmitted to the host via the gigabit network on the front panel of the JH card. The host supports simultaneous data access from up to seven devices. When the main computing card resources are insufficient and additional computing resources are needed, the main computing card and the expanded computing card are connected via the gigabit Ethernet on the backplane of the JH card, thereby enabling data sharing and information synchronization among the various computing nodes.
[0070] In one embodiment, see [reference needed]. Figure 1 , Figure 4 , Figure 6 The computing power board includes a main computing power card; the main computing power card includes a sensor data acquisition and publishing module, an intrusion detection module, a decision control module, a network interaction module, and a human-computer interaction display module;
[0071] Specifically, the sensor data acquisition and publishing module acquires the multi-source sensor data and publishes it to the intrusion detection module; the intrusion detection module performs intrusion detection on the multi-source sensor data and sends the intrusion detection result to the decision control module; the decision control module makes a decision based on the intrusion detection result and sends the generated decision result and the intrusion detection control command to the human-machine interaction display module and the network interaction module, respectively; the network interaction module sends the intrusion detection control command to the train control and management system through the communication board.
[0072] In one embodiment, see [reference needed]. Figure 1, Figure 4 , Figure 6 The main computing power card also includes an expansion card data management module; wherein, the expansion card data management module publishes the multi-source sensor data corresponding to the first expansion computing power card and the second expansion computing power card to the first expansion computing power card and the second expansion computing power card respectively; the first expansion computing power card and the second expansion computing power card send the intrusion detection results generated by each to the expansion card data management module.
[0073] Main computing power card SL-1:
[0074] 1) Input: Input voltage is DC24V, power consumption is 75W;
[0075] 2) The front panel includes one HDMI port, which can be connected to a display screen to realize human-computer interaction function; the front panel includes two 1000M Ethernet ports, which can be connected to the maintenance network for board maintenance and the business network for data transmission.
[0076] 3) The board uses an ARM 12-core processor with a 2.2GHz clock speed, 32GB of RAM, and 2TB of onboard storage;
[0077] 4) AI performance: 250 TOPS.
[0078] In one embodiment, see [reference needed]. Figure 2 , Figure 4 , Figure 6 The computing power board further includes a first expansion computing power board; the power supply board further includes a first expansion power supply board; wherein, the input terminal of the first expansion power supply board includes two 110V DC power supplies; the output terminal of the first expansion power supply board includes one 24V DC power supply for powering the first expansion computing power board; the first expansion computing power board is used to perform intrusion detection on the multi-source sensor data using an intrusion detection model to enhance the computing power of the main computing power board.
[0079] In one embodiment, see [reference needed]. Figure 3 , Figure 4 , Figure 6 The computing power board also includes a second expansion computing power board; the power supply board also includes a second expansion power supply board; wherein, the input terminal of the second expansion power supply board includes two 110V DC power supplies; the output terminal of the second expansion power supply board includes one 24V DC power supply for powering the second expansion computing power board; the second expansion computing power board is used to perform intrusion detection on the multi-source sensor data using an intrusion detection model to enhance the computing power of the main computing power board.
[0080] Expandable computing power cards SL-2 / SL-3:
[0081] 1) Input: Input voltage is DC24V, power consumption is 75W;
[0082] 2) The front panel includes two 1000M Ethernet interfaces, which can be connected to the maintenance network for board maintenance and the business network for data transmission;
[0083] 3) The board uses an ARM 12-core processor with a 2.2GHz clock speed, 32GB of RAM, and 2TB of onboard storage;
[0084] 4) AI performance: 250 TOPS.
[0085] Figure 6 This embodiment illustrates the data flow within the computing power board and between the main computing power card and the extended computing power cards. The main computing power card SL-1 collects, encodes, decodes, and publishes multi-source sensor data through a sensor data acquisition and publishing module. Subsequently, the detection model deployed on the computing power card subscribes to the required data from the published content of the sensor data publishing nodes according to input requirements and performs intrusion detection tasks. After the task is completed, the detection results are sent to the decision control module. The decision control module also obtains vehicle status information from the vehicle control network through an interaction module with the TCMS control network and sends control commands based on the comprehensive decision results to the TCMS network. Simultaneously, the decision control module sends the detection and decision results to the human-machine interaction display module, thereby enabling human-machine interaction between the system and the user. For real-time applications or applications with large-scale models, SL-2 and SL-3 expansion computing cards can be flexibly selected as needed to distribute time-consuming or resource-intensive tasks, thereby improving the overall processing power and response speed of the equipment. In this case, the expansion card data management module on the main computing card will publish the sensor data used by the expansion computing card to each task node of the expansion computing card. After completing the detection task, the detection module on the expansion computing card will send the detection results back to the expansion card data management module. This module will then send the detection results from the expansion computing card to the decision control module for overall decision-making.
[0086] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.
[0087] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.
[0088] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In the description of this specification, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the embodiments in this specification. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described can be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.
[0089] The above description is merely an embodiment of the present specification and is not intended to limit the embodiments of the present specification. For those skilled in the art, various modifications and variations can be made to the embodiments of the present specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the embodiments of the present specification should be included within the scope of the claims of the embodiments of the present specification.
Claims
1. An onboard computing unit for a train operation environment perception system, characterized in that, This includes power supply boards, communication boards, computing boards, and gigabit Ethernet data exchange boards; The power supply board provides power to the computing board and the gigabit Ethernet data exchange board; the gigabit Ethernet data exchange board provides a multi-source sensor data access channel for the multi-source sensor group; the computing board generates intrusion detection control commands based on the multi-source sensor data received from the gigabit Ethernet data exchange board, and sends the intrusion detection control commands to the train control and management system through the communication board.
2. The onboard computing unit of the train operation environment perception system according to claim 1, characterized in that, The Gigabit Ethernet data exchange board receives the multi-source sensor data via Gigabit Ethernet and transmits the multi-source sensor data to the computing power board via Gigabit Ethernet.
3. The onboard computing unit of the train operation environment perception system according to claim 1, characterized in that, The computing power board includes a main computing power card; the main computing power card includes a sensor data acquisition and publishing module, an intrusion detection module, a decision control module, a network interaction module, and a human-computer interaction display module; Specifically, the sensor data acquisition and publishing module acquires the multi-source sensor data and publishes it to the intrusion detection module; the intrusion detection module performs intrusion detection on the multi-source sensor data and sends the intrusion detection result to the decision control module; the decision control module makes a decision based on the intrusion detection result and sends the generated decision result and the intrusion detection control command to the human-machine interaction display module and the network interaction module, respectively; the network interaction module sends the intrusion detection control command to the train control and management system through the communication board.
4. The onboard computing unit of the train operation environment perception system according to claim 1, characterized in that, The power supply board includes a main power supply card; the input terminal of the main power supply card includes two 110V DC power supplies; the output terminal of the main power supply card includes one 5V DC power supply and one 24V DC power supply; wherein, the 5V DC power supply is used to power the communication board; the 24V DC power supply is used to power the computing board and the gigabit Ethernet data exchange board.
5. The onboard computing unit of the train operation environment perception system according to claim 3, characterized in that, The computing power board also includes a first expansion computing power board; the power supply board also includes a first expansion power supply board; wherein, the input terminal of the first expansion power supply board includes two 110V DC power supplies; the output terminal of the first expansion power supply board includes one 24V DC power supply for powering the first expansion computing power board; the first expansion computing power board is used to perform intrusion detection on the multi-source sensor data using an intrusion detection model to enhance the computing power of the main computing power board.
6. The onboard computing unit of the train operation environment perception system according to claim 5, characterized in that, The computing power board also includes a second expansion computing power board; the power supply board also includes a second expansion power supply board; wherein, the input terminal of the second expansion power supply board includes two 110V DC power supplies; the output terminal of the second expansion power supply board includes one 24V DC power supply for powering the second expansion computing power board; the second expansion computing power board is used to perform intrusion detection on the multi-source sensor data using an intrusion detection model to enhance the computing power of the main computing power board.
7. The onboard computing unit of the train operation environment perception system according to claim 6, characterized in that, The main computing power card also includes an expansion card data management module; wherein, the expansion card data management module publishes the multi-source sensor data corresponding to the first expansion computing power card and the second expansion computing power card to the first expansion computing power card and the second expansion computing power card respectively; the first expansion computing power card and the second expansion computing power card send the intrusion detection results generated by each to the expansion card data management module.
8. The onboard computing unit of the train operation environment perception system according to claim 1, characterized in that, The communication board includes a multi-functional vehicle bus communication board and a train real-time data protocol communication board. If the train control network is a multi-functional vehicle bus network, the train control network is accessed through the multi-functional vehicle bus communication board. If the train control network is a train real-time data protocol communication network, it is accessed through the train real-time data protocol communication board.
9. The onboard computing unit of the train operation environment perception system according to claim 1, characterized in that, The multi-source sensing data includes millimeter-wave radar data, visible light camera data, infrared camera data, lidar data, and combined inertial navigation data.
10. The onboard computing unit of the train operation environment perception system according to claim 1, characterized in that, It also includes a fan plate for providing precise heat dissipation support for the computing board.