Automatic driving control method for rail transit train

Abstract

The invention relates to the technical field of rail transit automatic driving, and particularly discloses a rail transit train automatic driving control method, which is characterized in that a vehicle-mounted control center cooperates with a signal system, a traction unit and a braking system, and centimeter-level real-time positioning is realized by adopting extended Kalman filtering and fusing satellite positioning, an odometer and trackside beacon data; the parking precision of the platform is controlled based on a graded braking strategy (coasting-common-precise braking) and pulse type terminal point fine adjustment; the method dynamically adapts to line condition changes in combination with feedforward-feedback composite speed control, and establishes a multi-level safety monitoring network (obstacle detection / overspeed protection / vehicle door linkage) to realize full-process closed-loop control, and finally, high-precision, high-safety and high-coordination operation under unmanned driving is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automatic driving of rail transit, and more particularly, to an automatic driving control method for rail transit trains. BACKGROUND

[0002] With the acceleration of urbanization, the rail transit system is under pressure to operate at high density and high punctuality rate, and automatic driving technology has become a key development direction to improve transportation efficiency and safety. The traditional mode based on manual driving by drivers is limited by reaction speed and operation consistency, and it is difficult to meet the running demand of minute-level tracking interval. The existing ATO (Automatic Train Operation) system still has defects such as large positioning deviation, rough braking control, and insufficient system linkage in complex scenarios, which restricts the improvement of full automation level. The existing technology has the following deficiencies:

[0003] 1. Defects in parking precision control

[0004] The existing positioning scheme relies on a single signal source (such as track circuit or satellite positioning), and in complex environments such as tunnels and dense building areas, the positioning error is often more than ±0.5 meters, which leads to misalignment of the train door and the platform door exceeding the safety tolerance (standard requirement ≤0.3 meters). More seriously, the cumulative error of the odometer lacks an effective correction mechanism, and the parking deviation after continuous inter-station operation is exponentially amplified, forcing the system to frequently degrade to manual intervention mode.

[0005] 2. Imbalance between braking process comfort and safety

[0006] The traditional braking strategy uses a fixed deceleration curve control, without considering factors such as actual load and changes in track adhesion conditions. When empty, excessive braking force causes passenger falling risk (measured deceleration >1.3 m / s 2 ), and when heavily loaded, it may cause over-station due to insufficient braking force (deceleration <0.6 m / s 2 ). The emergency braking response delay is as high as 1.5 seconds or more, and lacks a hierarchical warning mechanism, causing more than 80% of unnecessary emergency braking triggers.

[0007] 3. Subsystem coordination barriers

[0008] Key systems such as signals, traction, braking, and doors operate independently, and information interaction relies on low-speed buses (such as CAN bus transmission cycle >200ms). In typical scenarios such as entering a station, the braking system does not send a "zero speed lock" signal to the door system in time, causing the doors to open incorrectly when the train is not completely stationary. In the past three years, the accident rate caused by such failures has increased by 37%.

[0009] 4. Passive fault response

[0010] The prior art adopts a post-response mechanism for sudden abnormalities (such as track intrusion and equipment failure) and relies on manual dispatching decisions. Data analysis shows that, from the occurrence of a fault to the execution of safety measures by the system, an average of 8.6 seconds is consumed, far exceeding the safety time window (such as the critical time for obstacle collision is usually < 3 seconds). Moreover, fault diagnosis only covers 15% of the core components and cannot achieve full-life-cycle predictive protection.

[0011] Therefore, an automatic driving control method for rail transit trains is proposed to solve the above problems. SUMMARY

[0012] In order to overcome the above-mentioned defects of the prior art, embodiments of the present application provide an automatic driving control method for rail transit trains to solve the problems raised in the above background.

[0013] To achieve the above-mentioned purpose, the present application provides the following technical scheme: an automatic driving control method for rail transit trains, comprising: a vehicle-mounted control center receiving a dispatching instruction after confirming that a departure condition is met when the train is at a station platform, the departure condition needing to include signal system authorization, closed and locked state of the train door, relieved state of the braking system and ready state of the traction system; then a traction control unit driving the train to accelerate to a target cruising speed according to a preset acceleration curve; in the interval running stage, the speed is dynamically adjusted to track the planned running curve by real-time matching of train positioning data and electronic map; when the train approaches the target platform to a preset distance, a staged braking program is started, and in turn, coasting, normal braking and precise parking braking are performed; the braking system is controlled by using the fusion data of the vehicle-mounted positioning device and the platform positioning beacon, so that the train stopping position error is not more than ±0.3 meters; after the train is stopped stably, the alignment state of the platform door and the train door position is verified and a train door enabling instruction is generated; the track environment and running speed are monitored in real time through multiple sensors throughout the journey, and emergency braking is triggered immediately when an obstacle intrusion safety distance or overspeed is detected.

[0014] Preferably, the departure condition confirmation specifically includes: obtaining the mechanical locking signal of all train door state sensors through the train network system to confirm that the train door is completely closed; synchronously collecting the brake cylinder pressure sensor data, determining that the braking system is completely relieved when the pressure value is continuously less than 5 kPa for 3 seconds, and receiving the self-check completion signal of the traction system.

[0015] Preferably, the interval speed adjustment adopts a feedforward-feedback compound control mechanism: the feedforward controller pre-calculates the traction compensation value based on the real-time obtained line slope data and the front curve radius; the feedback controller dynamically corrects the traction or braking force output at a frequency of 10 times per second by comparing the deviation value of the actual speed and the planned speed curve.

[0016] Preferably, the staged braking program contains distance-triggered logic: cut off traction at 500 meters from the platform and enter coasting mode; apply 50% of the rated braking force when the remaining distance is 200 meters; activate the precise stopping algorithm when entering the 50-meter range, dynamically adjust the braking force according to the ratio of real-time speed and remaining distance, and switch to pulse-type point braking within the last 5 meters until complete stop.

[0017] Preferably, the positioning data fusion employs extended Kalman filtering technology to process multi-source signals: fuse the latitude and longitude data of the global satellite positioning system, the pulse count information of the vehicle-mounted odometer, and the position reference of the RFID tags buried between the tracks, to generate the real-time offset between the train center line and the platform reference line.

[0018] Preferably, the precise stopping control contains a three-level correction mechanism: activate the first-level position calibration when the train speed drops to 5 kilometers per hour, correct the odometer cumulative error based on the platform end beacon; start the second-level braking curve optimization when the speed drops to 2 kilometers per hour, adjust the braking force gradient according to the position deviation; and use the third-level intermittent braking within the last 0.5 meters, apply braking force with a 0.3-second on-off period until stopping.

[0019] Preferably, the full-process safety protection contains a parallel monitoring strategy: detect obstacles within 200 meters in front of the train through the fusion of millimeter wave radar and visual sensors installed on the train head, and start emergency braking when the distance to the obstacle is less than the dynamically calculated safe braking distance; monitor the train's rear slip displacement at the same time, and immediately apply full braking force if the rear slip exceeds 0.5 meters after stopping; continuously detect foreign objects in the gap between the train and the platform during the opening of the train doors, and prevent the train doors from moving and issue an alarm when foreign objects are found.

[0020] Technical effects and advantages of the present application:

[0021] Compared with the prior art, the present application constructs a collaborative control system for full automation of rail transit trains, realizes centimeter-level positioning through multi-source data fusion (satellite / odometer / beacon), ensures ±0.3-meter precise stopping by combining three-level braking strategies and end-point fine-tuning control, dynamically adjusts traction / braking force using a feedforward-feedback composite control algorithm, adapts to slope changes and load fluctuations, establishes a multi-level parallel safety monitoring network to achieve millisecond-level response to risks such as obstacle identification and overspeed protection, forms a full-process closed-loop control chain from departure condition verification to platform door linkage, and is equipped with a dual protection mechanism of hard-wire backup and network redundancy to automatically degrade operation and accurately handle faults when faults occur. This technical system realizes the full-cycle unmanned operation of train autonomous perception-decision-execution, and reaches the industry-leading level in safety, punctuality and comfort. BRIEF DESCRIPTION OF DRAWINGS

[0022] Fig. 1 The figure is a system framework diagram of the present application.

[0023] Fig. 2 Workflow diagram for the present invention. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0025] Embodiment one:

[0026] As shown in the accompanying drawings, Figs. 1-2 (1) a rail transit train automatic driving control method, comprising: a vehicle-mounted control center receiving a dispatching instruction after confirming that a departure condition is met when the train stops at a platform, the departure condition including signal system authorization, closed and locked state of the train door, released state of the braking system, and ready state of the traction system; then a traction control unit outputs traction force to drive the train to accelerate to a target cruising speed according to a preset acceleration curve; in the interval running stage, the speed is dynamically adjusted to track the planned running curve by real-time matching of train positioning data and electronic map; when the train approaches the target platform to a preset distance, a staged braking program is started, and in turn, coasting, normal braking, and precise parking braking are performed; the braking system is controlled by using the fusion data of the vehicle-mounted positioning device and the platform positioning beacon, so that the train stopping position error is not more than ±0.3 meters; after the train is stopped stably, the alignment state of the platform door and the train door position is verified, and a train door enabling instruction is generated; the track environment and running speed are monitored in real time through multiple sensors throughout the journey, and emergency braking is triggered immediately when an obstacle intrudes into a safety distance or overspeed is detected, wherein the vehicle-mounted control center (a SIL-4 level safety computer with redundant design) continuously monitors four departure conditions during the train stopping at the platform: receiving a moving authorization instruction sent by the wayside signal system through wireless communication; obtaining all door controller locking feedback signals (mechanical lock tongue insertion depth ≥8mm is determined as locked) through MVB train bus; collecting brake air cylinder pressure sensor data (pressure value <5kPa for 3 seconds is determined as released); receiving a state code of traction converter self-check completion (0xAA indicates readiness). When the four conditions are met at the same time, the vehicle-mounted center executes the dispatching instruction, and the traction control unit drives the train to accelerate to the target cruising speed according to the stored acceleration curve (0.5m / s 2 linearly increases to 1.2m / s 2) output PWM control signal to drive traction motor. When running in section, by fusing GNSS positioning module (update frequency 10 Hz) and route coordinates of electronic map, speed regulation module compares actual speed with planned curve deviation (allowance ± 2 km / h) at a frequency of 20 times per second, dynamically adjusts traction / braking instruction. Trigger braking program when approaching station 500 meters: first coasting slip to reduce kinetic energy, apply electric brake when 200 meters (braking torque is 50% of rated value), start hydraulic braking system within 50 meters and close loop control based on real-time position feedback (error sampling period 50 ms), finally make train stop position deviation ≤ ± 0.3 meters relative to station center reference line. After parking, verify that train door and station door center line are aligned (deviation < 10 mm) by laser alignment sensor, generate car door enable signal. Scan the 200-meter track area in front by multi-source perception system (including millimeter wave radar and infrared camera), if an obstacle is detected and the distance is less than the dynamically calculated safe braking distance (formula V 2 / 2a+ safety margin 10m, a takes the maximum deceleration 1.5m / s 2 ), immediately trigger emergency braking circuit.

[0027] (2) The departure condition confirmation specifically includes: obtaining the mechanical locking signal of all door state sensors through the train network system, confirming that the doors are completely closed; synchronously collecting brake cylinder pressure sensor data, determining that the brake system is completely relieved when the pressure value is continuously lower than 5 kPa for 3 seconds, and receiving the traction system self-check completion signal. In the departure condition verification stage, the vehicle-mounted center polls each car door control unit through the train network (MVB bus protocol), and when all door state bits bit0 (closing signal) and bit1 (locking signal) are high at the same time, it is determined that the doors are completely locked. Brake cylinder pressure data is collected through a pressure transmitter (range 0-10 kPa, accuracy ± 0.1 kPa), and if the pressure value is continuously lower than 5 kPa for 3 sampling periods (1 second), it is determined that the brake is completely relieved. Traction system readiness status detection includes: traction inverter DC link voltage is stable at 1500V±5%, IGBT temperature <85℃, no fault code is reported. The above states are transmitted to the vehicle-mounted center through hard-wire signal (110V level) and network message (process data 0x3A8) double channels, and any abnormal channel is prohibited to depart.

[0028] (3) The interval speed regulation adopts a feedforward-feedback composite control mechanism: the feedforward controller calculates the traction force compensation value based on the real-time acquired line slope data and the front curve radius; the feedback controller dynamically corrects the traction or braking force output at a frequency of 10 times per second by comparing the deviation value of the actual speed and the planned speed curve, wherein the interval speed control adopts a feedforward-feedback double-loop structure: the feedforward controller calculates the traction force compensation value (ΔF = train mass × slope × gravitational coefficient) in advance to adjust the traction instruction based on the electronic map pre-stored line data (slope resolution 1‰, curve radius accuracy 1m) 500 meters before entering the slope change section; the feedback controller adopts an incremental PID algorithm (proportional coefficient Kp = 0.8, integral time Ti = 15s) to collect the actual speed (Doppler radar speed measurement accuracy ± 0.1 km / h) at a frequency of 10 Hz, and outputs a dynamic correction after comparing with the planned curve. When the speed deviation exceeds +3 km / h, the electric brake is started, and when the deviation is lower than -2 km / h, the traction level is increased, and the traction / braking force change rate is limited to ±5 kN / s within the control period to avoid impact.

[0029] (4) The staged braking program contains distance trigger logic: cut off the traction force 500 meters from the platform to enter the coasting mode; apply 50% of the rated braking force when the remaining distance reaches 200 meters; start the precise parking algorithm when entering the 50-meter range, dynamically adjust the braking force according to the real-time speed and remaining distance ratio, and switch to pulse type point braking in the last 5 meters until complete parking, wherein the staged braking program triggers by distance: 500 meters from the platform, the traction control unit outputs zero traction instruction and activates the energy recovery mode; when the remaining distance is 200 meters, the brake control unit outputs an electric brake instruction of 50% of the rated braking force (corresponding torque 3200 Nm); after entering the 50-meter range, the precise parking algorithm starts, dynamically adjusts the braking force (output range 50%-100% of the rated value) by calculating the braking deceleration a = (current speed2-target speed2) / (2×remaining distance) in real time; switch to pulse braking mode in the last 5 meters, alternately apply 100% braking force (for 0.3 seconds) and release braking force (for 0.2 seconds) at a period of 0.5 seconds until the speed drops below 0.2 km / h to trigger the holding brake.

[0030] (5) The positioning data fusion adopts extended Kalman filter technology to process multi-source signals: fusion of global satellite positioning system latitude and longitude data, vehicle odometer pulse count information, and RFID tag position reference buried between tracks, to generate real-time offset of train center line and platform reference line, wherein the positioning fusion adopts an extended Kalman filter, and the input sources include: WGS-84 coordinates (latitude / longitude error ±1.5m) output by the GNSS module, which are mapped to the line coordinate system through a coordinate conversion matrix; the running distance measured by the pulse encoder (2000 pulses per revolution) with a sampling period of 10ms; passive RFID tags (storing absolute mileage value) buried every 50 meters between tracks. The filter prediction stage is based on the motion model (acceleration integration), and the update stage resets the position when an RFID tag is detected (read / write distance ±0.3m). The final output is the lateral deviation (accuracy ±2cm) and longitudinal deviation (accuracy ±5cm) of the train center line relative to the platform reference line.

[0031] (6) Precise parking control contains three-level correction mechanism: when the train speed drops to 5km / h, the first level position calibration is activated, based on the platform end beacon to correct the cumulative error of the odometer; when the speed drops to 2km / h, the second level brake curve optimization is started, and the brake force gradient is adjusted according to the position deviation; the third level intermittent braking is used in the last 0.5m distance, and the brake force is applied with a 0.3s on-off period until parking, wherein the precise parking three-level correction process: when the speed drops to 5km / h (first level), the vehicle center reads the absolute position of the platform starting end beacon (installation accuracy ±1cm), and resets the cumulative error of the odometer; when the speed drops to 2km / h (second level), the brake force is adjusted according to the position deviation ΔS (ΔS = actual position-target position) by gradient, if ΔS>0.2m, increase the brake force by 10%, if ΔS<-0.2m, reduce the brake force by 15%; within the last 0.5m (third level), intermittent braking is started, and the brake command is turned on and off with a 0.3s period (0.2s on to apply 80% brake force, 0.1s off to release), until the speed sensor detects a value of zero for 10 consecutive sampling points (sampling rate 100Hz) to lock the brake cylinder.

[0032] (7) The full range of safety protection includes parallel monitoring strategy: through the installation of millimeter wave radar and visual sensor fusion detection of obstacles within 200 meters in front of the vehicle head, when the obstacle distance is less than the dynamically calculated safe braking distance, start emergency braking; At the same time, monitor the displacement of the train after slipping, if the rear sliding of more than 0.5 meters occurs after stopping, immediately apply full braking force; During the opening of the door, continuously detect the foreign matter in the platform gap, and prevent the door from moving and issue an alarm when foreign matter is found, wherein the safety protection system executes three-channel monitoring in parallel: forward obstacle detection uses 77GHz millimeter wave radar (detection angle 120°, resolution 0.5m) and visible light camera (2 million pixels, 30fps) fusion, identifies track intrusions through a deep learning model (YOLOv5 architecture), and when the obstacle distance is < (current speed 2 / (2*maximum deceleration) + 10m safety margin), triggers emergency braking; After stopping, real-time monitor the displacement of the train body, if the laser range finder detects that the rear sliding distance is > 0.5cm, immediately apply 100% parking brake; During the opening of the door, scan the platform gap through the TOF sensor (range 0-1m, accuracy ±1mm) installed at the door, and when detecting ≥30mm*30mm foreign matter, block the door opening signal and trigger the audible alarm (105dB buzzer).

[0033] Embodiment two: multi-source data joint modeling scene

[0034] Stage 1: Safety verification before departure

[0035] The on-board control center (dual-core safety computer) continuously checks four key states during train station stop

[0036] Signal authorization: receive the green signal instruction code (code "0x5A" represents permission to start) sent by the ground signal device through wireless communication

[0037] Door status: poll the door controller of each car, when all the doors feedback "closed and locked" signal (mechanical lock tongue insertion depth ≥8mm)

[0038] Brake release: collect brake air cylinder pressure sensor data (pressure unit: kilopascal), if the pressure value is detected to be <5 kilopascal (equivalent to complete release of braking force) for 3 seconds in a row

[0039] Traction readiness: the traction system self-check completion flag is "0xAA" (indicating that the inverter voltage is stable at 1500±75 volts and the device temperature is normal)

[0040] Execution logic: when the four conditions are met at the same time, the on-board center lights the "departure ready" indicator light on the driver's console, waits for the driver to press the start button or automatically receives the dispatch departure instruction.

[0041] Stage 2: Smooth start and acceleration

[0042] The traction control unit controls the train start according to the pre-stored acceleration curve (such as accelerating to 30 km / h in 0-10 seconds, acceleration 0.8 m / s 2 ) : first send the PWM modulation instruction of initial 30% duty cycle to the traction inverter, then dynamically adjust the duty cycle every 0.1 second (step ± 2%), and monitor the deviation of motor speed from the target value in real time through the encoder. If 90% of the target speed is not reached within 5 seconds, trigger fault diagnosis and automatically reduce the traction.

[0043] Stage 3: Intelligent tracking of section speed

[0044] The traction control adopts a "prediction + correction" dual-mode strategy: prediction control is based on electronic map to obtain line slope (such as +3 ‰) 500 meters in advance, and dynamically calculates the compensation traction (compensation value = total weight of train × slope × 9.8); real-time correction is achieved by collecting Doppler radar speed data 10 times per second, comparing with the planned curve (allowing ± 2 km / h deviation), and dynamically adjusting according to "1 km / h overspeed reduces 3% traction / 1 km / h underspeed increases 4% traction". Automatic speed reduction when approaching a curve, target speed calculated as √(track radius × 0.8) (such as 45 km / h speed limit for 300-meter radius curve), achieving closed-loop control throughout the process.

[0045] Stage 4: Station entry staged braking

[0046] The train adopts a three-stage braking strategy triggered by distance: cut off traction and enable regenerative braking (efficiency > 85%) at 500 meters from the target; apply 50% rated braking force (such as 160 kN) at 200 meters, maintain -0.5 m / s 2 deceleration; start precise braking at 50 meters, dynamically calculate the required deceleration according to the formula (current speed 2 / 2 × remaining distance) (such as 1.2 m / s 2 at 20 meters from the platform when the speed is 15 km / h), and accurately execute by the hydraulic system. The whole process realizes staged deceleration control from coasting to precise parking.

[0047] Stage 5: Millimeter-level precise parking

[0048] The train realizes ± 0.3-meter precise parking through multi-source fusion positioning: through satellite positioning (± 2 meters), wheel encoder (2.5 cm per pulse) and track RFID tag (50-meter interval, error < 1 cm) three data sources, using prediction-calibration algorithm - usually calculate position by speed integration, correct cumulative error when passing through RFID tag. Finally, switch to pulse braking mode (0.3 seconds braking / 0.2 seconds release alternately) within 5 meters, until the speed drops below 0.5 km / h, to ensure that the train head is accurately aligned with the center line of the platform.

[0049] Stage 6: Safe linkage of vehicle doors

[0050] Three verifications are performed after parking:

[0051] Zero speed confirmation: the speed sensor is continuously sampled for 10 times (once every 0.01 seconds) and all the values are zero Position alignment: the laser alignment instrument measures the gap between the vehicle door and the platform door (standard value 110 ± 5 mm)

[0052] Platform readiness: receiving the "door locked" signal (voltage signal 24V) sent by the platform door controller

[0053] Perform action: when the three conditions are met, the vehicle-mounted center sends a 12V high-level opening instruction to the vehicle door controller, and after a delay of 0.5 seconds, the door opening motor is started.

[0054] Stage 7: Full-time active protection

[0055] Three major safety monitoring are simultaneously performed during train operation: millimeter wave radar (200 meters / 120°) and visual recognition (minimum 30 cm object) are used for front obstacle protection, dynamic braking distance is calculated (formula: speed 2 / 3 + 10 meters buffer), and emergency braking is immediately performed in case of danger; the laser range finder monitors the vehicle body after parking, and the vehicle is automatically locked if the rear sliding exceeds 5 cm; infrared detection (850 nm / 5-15 cm height) is used for door gap detection, and the door is immediately terminated if ≥10 cm foreign matter is found. Three protections form a three-dimensional protection in the whole process.

[0056] Stage 8: Fault safety processing

[0057] The train adopts a three-level intelligent fault response mechanism: for mild faults (such as single sensor abnormalities), the train automatically reduces the speed to 15 km / h and maintains operation to the next station; for moderate faults (such as positioning loss), the train immediately stops and attempts to restart the system; for severe faults (such as brake failure), the train triggers mechanical emergency braking, automatically plays a 90 decibel evacuation broadcast, and sends fault positioning information to the dispatch center through GSM-R, achieving fault classification and precise disposal.

[0058] Finally, it should be noted that in the description of the present application, it should be noted that unless otherwise specified and limited, the terms "mounting", "connection", "connection" should be broadly understood, which can be mechanical connection or electrical connection, or the internal connection of two elements, or direct connection, "up", "down", "left", "right" and the like are only used to indicate the relative positional relationship, when the absolute position of the described object changes, the relative positional relationship may change;

[0059] Secondly: the present application discloses the structure involved in the embodiment of the present application, and other structures can be referred to the usual design, and in the case of no conflict, the same embodiment and different embodiments of the present application can be combined with each other;

[0060] Finally: the above only for the preferred embodiments of the present application, and not for limiting the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application, should be included in the scope of protection of the present application.

Claims

1. A rail transit train automatic driving control method, characterized in that include: When the train stops at the platform, the onboard control center confirms that the departure conditions are met and receives the dispatch instruction. The departure conditions must also include signal system authorization, door closed and locked status, brake system release status, and traction system ready status. The traction control unit then outputs traction force according to a preset acceleration curve to drive the train to accelerate to the target cruising speed; During the section operation phase, the train positioning data is matched with the electronic map in real time, and the speed is dynamically adjusted to track the planned operation curve; when the train approaches the target platform to the preset distance, the graded braking procedure is initiated, and the coasting slip, normal braking and precision parking braking are performed in sequence; the braking system is controlled by the fusion data of the on-board positioning equipment and the platform positioning beacon, so that the train stop position error does not exceed ±0.3 meters; after the train stops, the alignment status of the platform door and the train door is verified and a door enable instruction is generated; the track environment and operating speed are monitored in real time through multiple sensors throughout the entire process, and emergency braking is triggered immediately when an obstacle is detected to intrude into the safe distance or the speed exceeds the limit.

2. The automatic driving control method for a rail transit train according to claim 1, characterized in that The departure condition confirmation specifically includes: obtaining the mechanical locking signals of all door status sensors through the train network system to confirm that the doors are completely closed; synchronously collecting brake cylinder pressure sensor data, and determining that the braking system is completely relieved when the pressure value is lower than 5 kPa for 3 seconds, and at the same time receiving the traction system self-test completion signal.

3. The automatic driving control method for a rail transit train according to claim 1, characterized in that The interval speed regulation adopts a feedforward-feedback composite control mechanism: the feedforward controller pre-calculates the traction force compensation value based on the real-time acquired line gradient data and the radius of the curve ahead; the feedback controller dynamically corrects the traction or braking force output at a frequency of 10 times per second by comparing the deviation between the actual speed and the planned speed curve.

4. The automatic driving control method for a rail transit train according to claim 1, characterized in that The graded braking program includes distance-triggered logic: traction is cut off at 500 meters from the platform and the vehicle enters coasting mode; 50% of the rated braking force is applied when the remaining distance reaches 200 meters; a precise parking algorithm is activated after entering the 50-meter range, dynamically adjusting the braking force based on the ratio of real-time speed to remaining distance, and switching to pulsed point braking in the last 5 meters until the vehicle comes to a complete stop.

5. The automatic driving control method for a rail transit train according to claim 1, characterized in that The positioning data fusion uses extended Kalman filtering technology to process multi-source signals: fusing the latitude and longitude data of the global satellite positioning system, the pulse count information of the on-board odometer, and the position reference of the RFID tag buried between the tracks to generate a real-time offset between the train centerline and the platform baseline.

6. The automatic driving control method for a rail transit train according to claim 1, characterized in that The precise parking control includes a three-level correction mechanism: when the train speed drops to 5 kilometers per hour, the first level position calibration is activated, and the accumulated error of the odometer is corrected based on the beacon at the platform end; when the speed drops to 2 kilometers per hour, the second level braking curve optimization is started, and the braking force gradient is adjusted according to the position deviation; in the last 0.5 meters of distance, the third level intermittent braking is used, and the braking force is applied in a 0.3-second on-off cycle until the train stops.

7. The automatic driving control method for a rail transit train according to claim 1, characterized in that The full-process safety protection includes a parallel monitoring strategy: the millimeter-wave radar installed on the front of the train is integrated with the visual sensor to detect obstacles within 200 meters ahead, and emergency braking is initiated when the obstacle distance is less than the dynamically calculated safe braking distance; the train's backward sliding displacement is also monitored, and if a backward sliding of more than 0.5 meters occurs after stopping, full braking force is immediately applied; foreign objects in the platform gap are continuously detected during the door opening phase. If foreign objects are found, the door movement is blocked and an alarm is issued.

Images