Rail transit multi-system fusion platform

Abstract

The invention discloses a rail transit multi-system fusion platform, and relates to the technical field of intelligent operation and maintenance of rail transit vehicles, and the platform comprises a fusion platform host, a communication switch, an application system and terminal equipment thereof. The core of the system is that logic functions of all application systems are packaged in an SDK mode, interaction with a fusion platform host is achieved through a unified API interface, computing power sharing of hardware resources such as an AI module and an MPU module in the host is achieved through a 10-gigabit PCIE bus, a SWITCH switch module dispatches data transmission in a unified mode, the independent computing capacity of a subsystem is reserved, and the data transmission efficiency is improved. By means of the framework, the newly-added system does not need to additionally deploy a host, the function can be expanded only by accessing terminal equipment meeting the standard, the limitation that in a distributed framework, hardware needs to be added to the newly-added system is effectively solved, and elastic reuse of the hardware resources is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent operation and maintenance of rail transit vehicles, in particular to a rail transit multi-system fusion platform. BACKGROUND

[0002] With the rapid expansion of urban rail transit networks, the demand for intelligent operation and maintenance of trains is increasingly urgent. In order to ensure train safety and operational efficiency, subway trains are gradually equipped with pantograph monitoring systems, running gear state monitoring systems, active obstacle warning systems, passenger behavior monitoring systems and other application systems. These systems collect vehicle state data, fault information and environmental parameters to support operation and maintenance decisions. Each system is equipped with a dedicated host, terminal equipment and communication cables, and it is urgent to reduce the life cycle cost through equipment simplification and standardization.

[0003] In the prior art, various monitoring systems exist in the form of independent complete equipment, each system is equipped with a dedicated host, terminal equipment, power cable and communication interface. For example, the pantograph monitoring system and the running gear monitoring system communicate with the train TCMS through independent hosts, using different wiring schemes and interface standards. The AI computing module of the obstacle warning system and the sensor processing module of the running gear each occupy installation space, resulting in repeated configuration of similar ports and complex cable routing paths. This "decentralized" architecture lacks a unified hardware platform and data interaction standard, making it impossible to share computing power and storage resources, difficult to solve subsystem compatibility problems, and unable to meet the high real-time requirements of unmanned projects for multi-system collaborative response.

[0004] In summary, the current independent configuration mode of rail transit monitoring systems has gradually exposed problems such as equipment redundancy, high energy consumption and complex maintenance, which is significantly different from the development requirements of smart rail, therefore, developing a fusion platform that can integrate multi-system functions, unify interface standards and optimize resource allocation is the key to solving the above problems, and is of great significance for promoting the intelligent upgrading of rail transit. SUMMARY

[0005] The track traffic multi-system fusion platform is provided to make up for the shortage of the prior art, and the multi-system fusion is realized through a "Lego type" modular hardware architecture, the core of which is that the logical functions of each application system are encapsulated in the form of SDK, and the AI module, the MPU module and other hardware resources in the host are shared in computing power through a gigabit PCIE bus, and the SWITCH switch module uniformly schedules data transmission, thereby retaining the independent computing capability of the subsystems and avoiding the repeated configuration of hardware resources, so that the newly added system does not need to be additionally deployed, and only needs to access the terminal equipment conforming to the standard to extend the function, thereby effectively solving the limitation of "newly added system must increase hardware" in the decentralized architecture, and realizing the elastic reuse of hardware resources.

[0006] To solve the above technical problems, the application provides the following technical solutions: on the one hand, a track traffic multi-system fusion platform, the composition of the platform comprises a fusion platform host, a communication switch, an application system and a terminal device thereof; The fusion platform host adopts a standard case structure, and is internally integrated with a PSU power module, an FSP processor, a SWITCH switch module, an HDD high-speed mechanical hard disk, an NVR super capacitor module, an AP wireless communication module, an AI intelligent operation module, an MPU controller module and a TRDP communication module. The communication switch is connected with the terminal device of each application system and the fusion platform host through an Ethernet. The application system and the terminal device thereof comprise at least two of a train active obstacle warning system, a fire alarm system, a driver behavior monitoring system, a stable and unstable state monitoring system, a running part monitoring system, a passenger behavior monitoring system and a storage battery state monitoring system. The terminal device of each application system collects target monitoring data and transmits the target monitoring data to the fusion platform host through an Ethernet. The SWITCH switch module realizes the data exchange between the modules in the fusion platform host through a gigabit bandwidth PCIE bus. The AI intelligent operation module and the MPU controller module in the fusion platform host cooperatively process data, and the processing result is uniformly reported to a train TCMS system through the TRDP communication module. The logical functions of each application system are encapsulated in the form of SDK, and the communication with the fusion platform host is realized through API interface calling, thereby realizing the isolated running of the functions.

[0007] Further, the PSU power module is a redundant power supply design, receiving a train 110VDC input, with a voltage range of DC 77V-137.5V, outputting DC 12V and DC 53V voltage, wherein the DC 12V powers other modules in the host, and the DC 53V output POE power supplies each application system.

[0008] Further, the SWITCH switch module is a gigabit Ethernet management switch, with a quantity of at least 2, and a backplane interface using PCIE aggregation technology to achieve a gigabit bandwidth, and a panel setting 12 gigabit Ethernet interfaces.

[0009] Further, the NVR super capacitor module provides ≥30 seconds of backup power supply, covering the AI intelligent operation module, the MPU controller module, and the power failure protection of the HDD high-speed mechanical hard disk.

[0010] Further, the FSP processor is based on FPGA technology, which collects vehicle speed signals in real time, sends the speed signals to the terminal devices of each application system through Ethernet as a synchronization reference for multi-system data acquisition, and simultaneously reserves signal output to the interface of the vehicle.

[0011] Further, at least one independent switch is further included, which is used for transmitting sensor data of the running gear monitoring system, and the switch is directly connected to the fusion platform host through a gigabit Ethernet.

[0012] Further, the MPU controller module integrates and processes data of the running gear monitoring system, the stability / instability monitoring system, and the fire alarm system, and triggers the AP wireless communication module to send an alarm when an anomaly is detected.

[0013] Further, the TRDP communication module uses a M12D type coded connector in accordance with IEC62076-2-101, with a hole type on the device end and a needle type on the cable end, and the key subsystem nodes are configured with 2 independent Ethernet controllers corresponding to 2 M12 connectors to achieve redundant and reliable communication.

[0014] On the other hand, a multi-system data collaborative processing method based on a rail transit multi-system fusion platform, the method comprising the following steps: Time synchronization: the FSP processor of the fusion platform host collects real-time train speed signals based on FPGA technology, generates a timestamp, and broadcasts it to the terminal devices of each application system through Ethernet as a synchronization reference for data acquisition by the terminal devices; Data acquisition and transmission: the terminal devices of each application system collect target monitoring data according to the synchronization reference, the monitoring data including image data, sensor signals, and bus data, and carrying a timestamp through a unified Ethernet interface to the fusion platform host; Parallel processing and packaging: the AI intelligent operation module of the fusion platform host calls the corresponding SDK to process image data, including obstacle warning, passenger behavior, and driver behavior image / point cloud data, the MPU controller module calls the corresponding SDK to process sensor data, including walking part vibration, stability, and battery state signal, both realize data interaction and parallel operation through the 10G PCIE bus of the SWITCH switch module, integrate the processing results, and package them into a unified data frame according to the communication protocol of the TRDP communication module; Unified reporting: the packaged data frame is transmitted to the train TCMS system through the M12D type coding connector of the TRDP communication module, and the original data and processing results are stored in the HDD high-speed mechanical hard disk.

[0015] Compared with the prior art, the rail transit multi-system fusion platform has the following beneficial effects: I. The application realizes multi-system fusion through a "Lego" modular hardware architecture, which is characterized by encapsulating the logical functions of each application system in the form of SDK, interacting with the fusion platform host through a unified API interface, and sharing computing power through the 10G PCIE bus of the AI module, MPU module, etc. in the host, and the SWITCH switch module uniformly schedules data transmission, which not only retains the independent computing capability of the subsystem, but also avoids the repeated configuration of hardware resources. This architecture allows new systems to be added without additional host deployment, and only needs to connect standard terminal devices to expand functions, effectively solving the limitation of "new system requires additional hardware" in the decentralized architecture, and realizing flexible reuse of hardware resources.

[0016] II. The application adopts the cooperative design of FSP processor based on FPGA and TRDP protocol, and builds a high-precision data synchronization and transmission mechanism. The speed synchronization signal generated by the FSP processor provides a unified time scale for each terminal device, ensuring the time alignment of image data and sensor signals. The TRDP communication module interacts with TCMS through the M12D type coding connector, and realizes redundant reliability of data transmission combined with the dual-homed network design. Compared with the problem of independent time service and incompatible communication protocols in traditional systems, this mechanism controls the time deviation of multi-source data to the microsecond level, and reduces the integration difficulty of different devices through standardized interfaces, providing high-consistency data source support for intelligent operation and maintenance.

[0017] Other advantages, objects, and features of the application will be set forth in part in the following specification, and in part will become apparent to those skilled in the art from a study of the following specification, or can be learned from practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to make the technical solutions in the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort on the premise that they do not pay creative effort. Figure 1 System architecture diagram of the rail transit multi-system fusion platform of embodiment one; Figure 2 Host layout schematic diagram of the rail transit multi-system fusion platform of embodiment one. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be described clearly and completely below. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present application.

[0020] Embodiment one This embodiment describes in detail the specific composition and working principle of a rail transit multi-system fusion platform. The platform integrates multiple types of core modules through a standardized chassis structure, takes unified Ethernet as a communication carrier, and realizes the collaborative operation of at least two types of application systems such as train active obstacle warning, fire alarm, and driver behavior monitoring. The platform uses FSP processors based on FPGA technology to provide a data synchronization reference, realizes high-speed data interaction between internal modules through a gigabit PCIE bus, and finally reports data to the train TCMS system in TRDP protocol through the collaborative processing of AI intelligent operation and MPU controllers. At the same time, the design of redundant power supply, super capacitor power failure protection, and functional isolation mechanism ensures the stability and safety of the platform, effectively solves the problems of equipment redundancy and poor compatibility of traditional distributed monitoring systems, and provides an efficient and integrated solution for rail transit intelligent operation and maintenance.

[0021] The hardware construction of the rail transit multi-system fusion platform is the basis for realizing the collaboration of multiple systems. Its core is to integrate various functional modules through a standardized chassis structure and complete the initial configuration to meet the needs of the train operation environment. The host of the fusion platform adopts a standard chassis structure, such as Figure 2As shown, the internal functional partitioning integrates nine core modules: PSU power module, FSP processor, SWITCH switch module, HDD high-speed mechanical hard disk, NVR super capacitor module, AP wireless communication module, AI intelligent operation module, MPU controller module, and TRDP communication module. The modules are physically connected through the guide rail and interface inside the case, wherein the SWITCH switch module is connected with other modules through the PCIE bus to ensure the high speed of data transmission; the PSU power module is connected with other modules inside the host and external application system terminal device through internal power supply lines to establish a power supply link.

[0022] The PSU power module adopts a dual-redundancy design, i.e., two identical power modules are installed in parallel in the case, when one module fails, the other module can automatically switch and undertake all power supply tasks, avoiding the failure of the platform due to power interruption. The module receives a train 110VDC input, its voltage withstand range covers DC 77V-137.5V, and can adapt to the fluctuation of the power supply voltage during train operation (such as voltage drop during start-up or voltage surge during braking), the module outputs two voltages: DC 12V is used to power the FSP processor, SWITCH switch module, AI intelligent operation module, etc. inside the host, and DC 53V is used to power the terminal devices of the external application system through the POE (Power over Ethernet) technology, such as the camera of the train active obstacle warning system and the sensor of the driver behavior monitoring system, without the need to lay independent power supply cables for the terminal devices, simplifying the train wiring.

[0023] The SWITCH switch module selects at least two gigabit Ethernet managed switches, the backplane interface of each switch adopts PCIE aggregation technology, combining multiple PCIE channels into a gigabit bandwidth link, ensuring that the data transmission rate between the modules inside the host reaches 10Gbps, and 12 gigabit Ethernet interfaces are arranged on the panel, of which 8 are used to connect the terminal devices of the external application system (such as the vibration sensor of the running gear monitoring system and the infrared detector of the passenger behavior monitoring system), 2 are reserved as expansion interfaces, and the other 2 are used to communicate with independent switches (when the running gear monitoring system is connected, the independent switch specially transmits the sensor data, and then directly connects to the SWITCH switch module through the gigabit Ethernet).

[0024] The NVR super capacitor module is connected with the AI intelligent operation module, the MPU controller module and the HDD high-speed mechanical hard disk in the cabinet through a power supply line. The capacity design ensures that ≥30 seconds of standby power supply can be provided when power is off. The purpose of this design is to prevent data loss caused by sudden power failure. When the train power supply is accidentally interrupted, the super capacitor can maintain the short-time operation of the above-mentioned modules, so that the AI intelligent operation module and the MPU controller module can temporarily store the data being processed to the HDD high-speed mechanical hard disk, complete the last data packaging and reporting, and avoid the loss of key monitoring data (such as abnormal vibration signal of running part and fire alarm information).

[0025] The TRDP communication module adopts a M12D type coded connector conforming to the IEC62076-2-101 standard. The device end (host side) is a hole type interface, and the cable end (connected to the TCMS system side) is a pin type interface, which ensures the anti-misplug and sealing performance of the connection (adapted to the vibration and dust environment during train operation). In order to improve the communication reliability, the module is configured with 2 independent Ethernet controllers, which correspond to 2 M12 connectors respectively, and are connected to the train TCMS system through double lines. When one of the lines fails, the other line can automatically take over the data transmission to realize communication redundancy.

[0026] As shown in Figure 1 The terminal devices of the train active obstacle early warning system, the fire alarm system, the driver behavior monitoring system and the running part monitoring system are distributed in different positions of the train according to the functional requirements, and are connected to the SWITCH switch module of the fusion platform host through Ethernet.

[0027] The terminal device of the train active obstacle early warning system is a high-definition camera (resolution 1920x1080) and a laser radar installed on the head of the train. The camera is responsible for collecting image data of the front track, and the laser radar is used to generate three-dimensional point cloud data to judge the distance of the obstacle. Both of them are connected to the No. 1 and No. 2 interfaces of the SWITCH switch module through a gigabit Ethernet cable, and are powered by the DC53VPOE power supply of the PSU power supply module.

[0028] The terminal device of the fire alarm system includes smoke sensors and temperature sensors distributed on the top of the car. Each car is equipped with 2 smoke sensors (to detect smoke concentration) and 1 temperature sensor (to monitor the environment temperature). The sensors are connected to the No. 3 interface of the SWITCH switch module after being gathered through a bus, and are also powered by POE.

[0029] The terminal device of the driver behavior monitoring system is an infrared camera and a heart rate sensor installed in the cockpit. The infrared camera is used to capture the driver's behaviors such as closing his eyes and lowering his head. The heart rate sensor collects the driver's physiological signals through a wristband or a seat patch. The data of the two devices are processed and transmitted to the host computer through the No. 4 interface.

[0030] The terminal device of the running part monitoring system is a vibration sensor and a temperature sensor (for monitoring bearing temperature) installed on the train bogie. Due to the large amount of data (1000 times of vibration signal collection per second), a separate switch is used for transmission. The separate switch is connected to the No. 5 interface of the SWITCH switch module through a 10-gigabit Ethernet network, ensuring that data transmission does not occupy the bandwidth of other systems.

[0031] After the platform is powered on, each module completes initialization in the preset order: The PSU power module is started first, outputs DC12V and DC53V voltages, and provides power for the host module and terminal devices. At the same time, it confirms the working status of the dual-redundant power supply through the internal self-checking program. If there is a fault, it will alarm through the indicator light. The FSP processor starts based on FPGA technology, starts real-time collection of train wheel speed signals (obtained through the train shaft end speed sensor), and converts them into digital signals to generate timestamps. The timestamp contains year, month, day, hour, minute, second, and microsecond information, which serves as the synchronization reference for all subsequent data collection. After the SWITCH switch module is started, it automatically identifies the type of terminal device connected and assigns a unique IP address to each device. It also enables the VLAN (Virtual Local Area Network) function, which divides the data of different systems into different virtual networks, achieving preliminary functional isolation. After the AI intelligent operation module and the MPU controller module are started, they load the preset algorithm model (such as the AI module loads the obstacle recognition algorithm, and the MPU module loads the vibration signal analysis algorithm), and establish communication with the SWITCH switch module through the SDK interface, waiting to receive data from the terminal device. When the TRDP communication module is initialized, it automatically establishes a connection with the train TCMS system, negotiates communication protocol parameters (such as data frame format and transmission rate), and sends test frames through a double line to confirm the smoothness of the communication link.

[0032] In rail transit monitoring, the monitoring data of different systems needs to be based on the same time reference, otherwise it will cause deviation in data analysis (such as the time of obstacle appearance not matching the train speed, which may misjudge the obstacle position). Therefore, the timestamp synchronization function of the FSP processor is the key to the efficient operation of the platform.

[0033] The FSP processor is connected with the speed sensor of the train through a special interface, and collects the speed signal of the vehicle in real time. Based on the parallel processing capability of the FPGA technology, the FSP processor can complete the analog-digital conversion of the speed signal within microseconds, and generates a time stamp in combination with an internal high-precision clock. The FSP processor broadcasts a synchronization frame containing the time stamp and the speed signal to the terminal equipment of each application system through Ethernet. After receiving the synchronization frame, the terminal equipment aligns the local clock thereof with the time stamp of the synchronization frame through a built-in clock calibration module, so as to ensure that the collection time deviation of all the equipment is not more than 5 microseconds.

[0034] For the camera and the laser radar of the active obstacle warning system, the synchronization reference ensures that the two collect data at the same time. For the vibration sensor of the running gear monitoring system, the synchronization reference makes the vibration signal collected by the vibration sensor be associated with the train speed. If the terminal equipment does not receive the synchronization frame for three times in succession, the terminal equipment automatically enters the local clock operation mode, and marks the state of “synchronization loss” in the data. At the same time, the terminal equipment sends an alarm signal to the SWITCH exchange module through Ethernet. After receiving the alarm, the MPU controller module triggers the AP wireless communication module to send a wireless alarm to the train control room, so as to remind the maintenance personnel to troubleshoot the fault. After the synchronization frame is restored, the terminal equipment automatically recalibrates the clock, so as to ensure the continuity of the data.

[0035] The terminal equipment of each application system collects data under the control of the synchronization reference, and transmits the data to the fusion platform host through Ethernet. The data collected by the terminal equipment is transmitted to the SWITCH exchange module through a gigabit Ethernet cable. The exchange module performs priority scheduling according to the data type and the emergency degree. For example: The warning data of the fire alarm system and the abnormal signal of the driver behavior monitoring system are marked as the highest priority, and the exchange module preferentially allocates bandwidth, so as to ensure that the data is transmitted to the host within 10 milliseconds.

[0036] The vibration data of the running gear monitoring system and the real-time image data of the active obstacle warning system are of medium priority, and the transmission delay is controlled within 50 milliseconds.

[0037] The conventional state data is of low priority, and can be transmitted when the bandwidth is idle.

[0038] In addition, the SWITCH switch module supports the IEEE 802.1p protocol, isolates the data of different systems through VLAN, prevents data conflicts, and before the terminal device transmits data, the terminal device performs CRC encoding on the data packet, generates a check value and appends the check value at the end of the data packet, after the SWITCH switch module receives the data packet, the SWITCH switch module recalculates the CRC value and compares it with the appended check value, if they are inconsistent, it is determined that the data transmission is incorrect, and immediately sends a retransmission request to the terminal device, and for data that fails to be retransmitted for 3 times, the switch will record the error log (including time, device IP, error type) and store it to the HDD high-speed mechanical hard disk for subsequent maintenance analysis.

[0039] After the fusion platform host receives the data of each terminal device, the AI intelligent operation module and the MPU controller module are cooperatively processed, high-speed data interaction is realized through the gigabit PCIE bus, and finally the analysis and integration of data are completed, the AI intelligent operation module is mainly responsible for processing the image / point cloud data of the active obstacle warning system and the image data of the driver behavior monitoring system, after receiving the video stream of the camera and the point cloud data of the laser radar, the AI module calls the obstacle recognition SDK, detects the target in the video frame, calculates the distance between the obstacle and the train in combination with the point cloud data, and sends the warning information to the MPU controller module in real time through the PCIE bus, at the same time, the image data of the infrared camera is extracted, the driver behavior is identified, and the behavior analysis result and the original image are stored in the HDD high-speed mechanical hard disk.

[0040] The MPU controller module focuses on processing the sensor data of the fire alarm system and the running part monitoring system, and cooperatively responding to the warning information with the AI module, after receiving the data of the smoke sensor and the temperature sensor, the MPU module calls the fire analysis SDK, immediately triggers the AP wireless communication module to send an audible and visual alarm to the cab and the train control room, the AI module and the MPU module realize data interaction through the gigabit PCIE bus of the SWITCH switch module, after the AI module and the MPU module are processed, all results (including warning signals, state evaluation, original data index) are transmitted to the SWITCH switch module, which integrates them into a unified data frame according to the preset format, the data frame includes the following fields: frame header (identifies the data type), timestamp (synchronization reference from FSP processor), device ID (distinguishes different systems), processing result (warning level, state value, etc.), check code (ensures data integrity).

[0041] The packaged TRDP data frame is transmitted to the train TCMS system through the TRDP communication module, at the same time, the original data and the processing result are stored in the HDD high-speed mechanical hard disk, forming a "real-time reporting + offline backup" dual-path data management mode, which not only meets the real-time monitoring needs of the train in operation, but also guarantees the long-term traceability of the data.

[0042] The RDP communication module sends data frames to the TCMS system through two independent M12D type coding connectors (corresponding to dual Ethernet controllers). The two-way transmission adopts a "parallel sending + priority confirmation" mechanism: the main link (No. 1 M12 connector) preferentially sends data, and the slave link (No. 2 M12 connector) synchronously sends the same content. After receiving, if the main link data is complete, the TCMS system only confirms the main link, and if the main link data is lost (such as transmission interruption caused by cable loosening), the slave link immediately requests retransmission, ensuring the reliability of data transmission.

[0043] During operation, the platform uses multiple levels of monitoring and protection mechanisms to deal with possible power interruptions, module failures, communication link abnormalities, etc., to ensure the continuity of core functions and data security. The PSU power module uses dual redundancy design, real-time monitoring of input voltage (train 110VDC) and output state (DC12V / DC53V). When the main power module fails due to voltage fluctuations, the standby power module automatically switches and takes over all power supply tasks. At the same time, it sends a "power module switching alarm" (including switching time and fault module number) to the train control room through the AP wireless communication module. For external application system power supply, if one POE output is short-circuited, the PSU module automatically cuts off the power supply and records the fault location. After the fault is eliminated, the power supply is manually restored.

[0044] For Ethernet links (including between terminal devices and host computers, and between host computers and TCMS systems), the SWITCH switch module and the TRDP communication module monitor the link quality (such as packet loss rate and delay) in real time. When the packet loss rate of the link between the terminal device and the host computer exceeds 5%, the SWITCH module automatically attempts to switch to the standby interface. If the switching fails, it sends a "link failure alarm" through the AP module to remind maintenance personnel to check the cable or interface. For the TRDP link between the host computer and the TCMS system, if the main link fails to retransmit for three consecutive times, the slave link immediately upgrades to the main link and marks the "link switching" status in the data frame, ensuring that the TCMS system knows the current communication path.

[0045] When the train is powered off as a whole, the NVR super capacitor module immediately starts to provide ≥30 seconds of backup power for the AI intelligent operation module, the MPU controller module, and the HDD high-speed mechanical hard disk. During this period, the two modules stop receiving new data, preferentially write the processing results in the memory that have not been saved to the HDD hard disk, and generate a "power failure protection record". The TRDP communication module sends the last frame of "system power failure alarm" to the TCMS system to complete the final synchronization of data. After the power is restored, the platform restarts and first reads the power failure protection record in the HDD hard disk to confirm the data integrity before normally starting each module, avoiding data chain breakage caused by power failure.

[0046] In summary, the embodiment details the hardware architecture of the rail transit multi-system fusion platform, the time synchronization mechanism, data acquisition and transmission, parallel processing, packaging and reporting, and emergency handling, etc. The platform integrates core modules in a standardized chassis, builds a high-speed communication network through a gigabit PCIE bus and a unified Ethernet, and solves the problems of device redundancy, data asynchronization, and poor compatibility in traditional decentralized systems through the collaborative operation of AI and MPU and the standardized output of TRDP protocol. At the same time, mechanisms such as redundant power supply, super capacitor protection, and link self-healing ensure the stability and reliability of the platform in complex train environments.

[0047] Embodiment Two Based on Embodiment One, the embodiment provides an operation process of the rail transit multi-system fusion platform in rail transit multi-system fusion. The specific steps of the process are as follows: I. Time Synchronization Phase FSP Processor Initialization: After the fusion platform host is powered on, the FSP processor starts and completes self-checking, initializes the internal clock module, and establishes a signal connection with the train speed sensor.

[0048] Vehicle Speed Signal Acquisition: The FSP processor receives the pulse signal of the train axle speed sensor in real time, and converts the pulse signal into real-time vehicle speed data through FPGA parallel operation.

[0049] Timestamp Generation: In combination with the internal high-precision clock, a timestamp containing year, month, day, hour, minute, second, and microsecond is generated for each set of vehicle speed data.

[0050] Synchronization Reference Broadcast: The FSP processor encapsulates the "vehicle speed data + timestamp" as a synchronization frame through Ethernet and broadcasts it to all terminal devices of the application system in UDP protocol.

[0051] Terminal Device Clock Calibration: After receiving the synchronization frame, each terminal device parses the timestamp and calibrates the local clock to ensure that the deviation between the local clock and the FSP processor timestamp is ≤5μs, completing the unification of data acquisition reference.

[0052] II. Data Acquisition and Transmission Phase Terminal Device Sampling Trigger: The terminal device triggers the acquisition action according to the timestamp of the synchronization reference; Running Part Vibration Sensor: The vibration acceleration is acquired every 1ms, and the sampling point is one-to-one corresponding to the microsecond level scale of the timestamp.

[0053] Monitoring Data Generation: The terminal device generates data by type: Video: Camera output encoded video frames, laser radar output point cloud binary data; Sensor type: vibration sensor outputs analog signals, smoke sensor outputs concentration values; Bus type: battery monitoring system outputs voltage / current bus data.

[0054] Data timestamping: the terminal device adds a timestamp in the synchronization reference to each packet of collected data.

[0055] Data encapsulation into Ethernet frames: the terminal device encapsulates "timestamp + data content + device ID + check code" into an Ethernet data frame.

[0056] Transmission path selection: normal sensor data (e.g., driver behavior camera): transmitted to the SWITCH module of the fusion platform host through the communication switch; running part monitoring data: directly connected to the SWITCH module through an independent switch (gigabit Ethernet) to avoid occupying normal bandwidth.

[0057] Data priority marking: the terminal device marks the priority according to the urgency of the data: Data integrity verification: after receiving the data frame, the communication switch / SWITCH module verifies the check code. If the verification fails, it sends a retransmission request to the terminal device, with a maximum of 3 retransmission attempts. If it fails, an error log is recorded.

[0058] III. Parallel processing and encapsulation phase Data classification and forwarding: the SWITCH module analyzes the device ID of the data frame and forwards it to the corresponding processing module according to the type: Image type data (obstacles, driver behavior) -> AI intelligent operation module; Sensor type data (vibration, temperature, smoke) -> MPU controller module.

[0059] AI module loads SDK: the AI intelligent operation module receives data and calls the corresponding SDK according to the device ID: Obstacle warning: load target detection SDK; Driver behavior: load behavior analysis SDK.

[0060] Image data processing: the AI module performs target detection on video frames, identifies the type of obstacles, and calculates the distance.

[0061] MPU module loads SDK: the MPU controller module receives data and calls the corresponding SDK: Running part monitoring: load vibration analysis SDK; Fire alarm: load smoke / temperature fusion SDK.

[0062] Sensor data processing: the MPU module performs spectral analysis on vibration signals to identify abnormal frequency peaks. Inter-module data interaction: AI module and MPU module interact data through the 10G PCIE bus of the SWITCH module: The AI module sends the obstacle distance to the MPU module, and calculates the collision time combined with the vehicle speed. The MPU module feeds back the running part state to the AI module to correct the obstacle warning strategy.

[0063] Processing result integration: The SWITCH module collects the processing results of AI and MPU, and integrates them in the format of "timestamp + device ID + result type + value + unit", TRDP data frame packaging: The integrated results are packaged into a unified data frame.

[0064] Four, unified reporting and storage stage TRDP link selection: The TRDP communication module detects the link state of the two independent Ethernet controllers, and preferentially sends data frames through the main link. If the main link fails, it automatically switches to the standby link.

[0065] Data transmission to TCMS: The packaged TRDP data frame is transmitted to the train TCMS system through the M12D type coding connector (device end hole type + cable end needle type).

[0066] TCMS confirmation feedback: After the TCMS system receives the data frame, it parses and verifies the checksum. If correct, it returns an acknowledgment frame (including the received timestamp). After the TRDP module receives the acknowledgment, it marks the data frame as "successfully reported".

[0067] Raw data storage trigger: The SWITCH module synchronously sends the raw data (unprocessed) of the terminal device to the HDD high-speed mechanical hard disk. The storage path is classified by "device ID / date / timestamp".

[0068] Processing result storage: The processing results of AI and MPU are stored in the HDD according to the same path.

[0069] Storage integrity verification: The HDD hard disk automatically performs CRC verification after completing a batch of data storage, generates a verification file to ensure that the data has not been tampered with, and at the same time, the NVR super capacitor module provides power protection for the storage process.

[0070] It will be apparent to those skilled in the art that the application is not limited to the details of the above-exemplified embodiments and that the present application can be implemented in other particular forms without departing from the spirit or essential characteristics of the present application. The embodiments should therefore be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the above description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. No reference signs in the claims should be considered as limiting the scope of the claims with respect to the figures of the patent document.

Claims

1. A multi-system integration platform for rail transit, characterized in that, The platform consists of: a converged platform host, a communication switch, application systems, and their terminal devices; The fusion platform host adopts a standard chassis structure and integrates a PSU power module, FSP processor, SWITCH switch module, HDD high-speed mechanical hard disk, NVR supercapacitor module, AP wireless communication module, AI intelligent computing module, MPU controller module, and TRDP communication module. The communication switch connects the terminal devices of each application system to the converged platform host via Ethernet. The application system and its terminal equipment include at least two of the following: train active obstacle warning system, fire alarm system, driver behavior monitoring system, stable and unstable state monitoring system, running gear monitoring system, passenger behavior monitoring system, and battery status monitoring system. Among them, the terminal devices of each application system collect target monitoring data and transmit it to the converged platform host via Ethernet; The SWITCH module enables data exchange between modules within the converged platform host via a 10 Gigabit bandwidth PCIe bus. The AI ​​intelligent computing module and MPU controller module in the integrated platform host work together to process data, and the processing results are uniformly reported to the train TCMS system through the TRDP communication module. The logical functions of each application system are encapsulated in the form of an SDK and communicate with the converged platform host through API interface calls to achieve functional isolation.

2. The rail transit multi-system fusion platform according to claim 1, characterized in that, The PSU power module is a redundant power supply design. It receives 110VDC input from the train, has a withstand voltage range of DC77V to 137.5V, and outputs DC12V and DC53V voltages. The DC12V power supply powers other modules in the main unit, and the DC53V output POE power supply powers various application systems.

3. The rail transit multi-system integration platform according to claim 1, characterized in that, The SWITCH module is a gigabit Ethernet managed switch, with a minimum of two modules. The backplane interface uses PCIe aggregation technology to achieve 10 gigabit bandwidth, and the front panel is equipped with 12 gigabit Ethernet interfaces.

4. The rail transit multi-system fusion platform according to claim 1, characterized in that, The NVR supercapacitor module provides backup power for ≥30 seconds, covering power outage protection for the AI ​​intelligent computing module, MPU controller module, and HDD high-speed mechanical hard drive.

5. The rail transit multi-system integration platform according to claim 1, characterized in that, The FSP processor is built on FPGA technology, which collects vehicle speed signals in real time and sends the speed signals to the terminal devices of various application systems via Ethernet as a synchronization reference for data collection from multiple systems. At the same time, it reserves an interface for signal output to the vehicle.

6. The rail transit multi-system fusion platform according to claim 1, characterized in that, It also includes at least one independent switch dedicated to transmitting sensor data from the walkway monitoring system, which is directly connected to the converged platform host via 10 Gigabit Ethernet.

7. The rail transit multi-system integration platform according to claim 1, characterized in that, The MPU controller module integrates and processes data from the running gear monitoring system, the stable / unstable state monitoring system, and the fire alarm system, and triggers the AP wireless communication module to send an alarm when an anomaly is detected.

8. The rail transit multi-system integration platform according to claim 1, characterized in that, The TRDP communication module uses an M12D type coded connector that conforms to IEC62076-2-101. The device end is a hole type and the cable end is a pin type. The key subsystem nodes are configured with two independent Ethernet controllers, corresponding to two M12 connectors to achieve redundant and reliable communication.

9. A method for collaborative data processing of multiple systems based on a rail transit multi-system fusion platform according to any one of claims 1-8, characterized in that, The method includes the following steps: Time synchronization: The FSP processor of the fusion platform host collects real-time train speed signals based on FPGA technology, generates timestamps, and broadcasts them to the terminal devices of various application systems via Ethernet, serving as the synchronization benchmark for data collected by the terminal devices. Data acquisition and transmission: The terminal devices of each application system acquire target monitoring data according to the synchronization benchmark. The monitoring data includes image data, sensor signals, and bus data, and is transmitted to the fusion platform host through a unified Ethernet interface with timestamps. Parallel processing and encapsulation: The AI ​​intelligent computing module of the fusion platform host calls the corresponding SDK to process image data, which includes image / point cloud data of obstacle warning, passenger behavior, and driver behavior. The MPU controller module calls the corresponding SDK to process sensor data, which includes vibration of the running gear, stability, and battery status signals. The two communicate and perform parallel computing through the 10 Gigabit PCIe bus of the SWITCH module. After integrating the processing results, they are encapsulated into a unified data frame according to the communication protocol of the TRDP communication module. Unified reporting: The encapsulated data frames are transmitted to the train TCMS system through the M12D type encoding connector of the TRDP communication module, while the raw data and processing results are stored in the HDD high-speed mechanical hard disk.

Images