Locomotive automatic driving risk detection and emergency disposal system based on environmental perception

Abstract

The invention provides a locomotive automatic driving risk detection and emergency disposal system based on environmental perception, and the system comprises a personnel perception system which is used for carrying out the perception and detection of the behaviors of safety personnel in a locomotive cab, and outputting a first detection result; the operation environment sensing system is used for sensing and detecting the operation track and / or the operation environment of the locomotive and outputting a second detection result; the automatic driving system is used for determining the current operation risk level of the locomotive according to the first detection result and / or the second detection result and outputting a control instruction according to the operation risk level; and the locomotive control system is used for executing the control instruction to complete locomotive operation and / or locomotive braking action. According to the invention, through a cooperation mechanism among the systems, timely detection and rapid response of potential risks are realized, the safety and reliability of locomotive automatic driving are greatly improved, and a powerful guarantee is provided for stable operation of rail transit.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of automatic driving of freight locomotives, in particular to a locomotive automatic driving risk detection and emergency disposal system based on environment perception and a freight train. BACKGROUND

[0002] With the rapid development of the rail transportation industry, automatic driving technology of locomotives has become an important direction to improve transportation efficiency and safety. However, the current automatic driving system of locomotives still faces many technical challenges in practical application: on the one hand, the existing system lacks sufficient perception of the state of the safety officer in the cab, and it is difficult to detect behaviors such as fatigue, distraction or violation of the safety officer in real time, which may lead to safety hazards when the safety officer cannot respond to unexpected situations in time; on the other hand, the perception accuracy of the running track and the surrounding environment of the locomotive is limited, especially in complex weather (such as heavy rain, heavy fog) or special road conditions (such as track foreign objects, line abnormalities), it is difficult to accurately identify obstacles and track features, resulting in delayed risk judgment.

[0003] At the same time, the risk level division of the existing system has weak correlation with the emergency disposal mechanism, and often can only trigger alarm or braking action singly, lacking the ability to dynamically adjust the disposal strategy according to the risk level. In addition, when the automatic driving system itself fails, the system lacks sufficient redundancy design, and cannot realize seamless switching of the safe state, which may lead to serious consequences such as out-of-control of the locomotive. These problems make the safety, reliability and intelligence level of the existing automatic driving system of locomotives difficult to meet the needs of complex operation scenarios, and an integrated system that can integrate personnel state perception, environment perception, dynamic risk assessment and multi-level emergency disposal is urgently needed. SUMMARY

[0004] The purpose of the embodiments of the present disclosure is to provide a locomotive automatic driving risk detection and emergency disposal system based on environment perception and a freight train to solve the problems existing in the prior art.

[0005] The embodiments of the present disclosure adopt the following technical solutions: a locomotive automatic driving risk detection and emergency disposal system based on environment perception, comprising: a personnel perception system for perceiving and detecting the behavior of a safety officer in the cab of a locomotive and outputting a first detection result; a running environment perception system for perceiving and detecting the running track and / or the running environment of the locomotive and outputting a second detection result; an automatic driving system for determining the current running risk level of the locomotive according to the first detection result and / or the second detection result, and outputting a control instruction according to the running risk level; and a locomotive control system for executing the control instruction to complete the locomotive running and / or locomotive braking action.

[0006] The embodiment of the present disclosure further provides a freight locomotive, comprising at least the locomotive automatic driving risk detection and emergency disposal system as described above.

[0007] The embodiment of the present disclosure has the beneficial effects that: through the cooperation of the personnel perception system and the running environment perception system, the behaviors of the safety personnel in the cockpit and the running track and environment of the locomotive are detected respectively, the overall capture of personnel and environmental risks is realized, and the limitations of single perception of the traditional system are overcome; the automatic driving system determines the running risk level according to the two types of detection results and outputs the control instruction, so that the risk assessment is more in line with the actual running condition, the blindness of emergency disposal is avoided, and through the cooperation mechanism between the systems, the potential risks are detected and responded quickly, the safety and reliability of the automatic driving of the locomotive are greatly improved, and a strong guarantee is provided for the stable operation of the rail transit. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the one or more embodiments of the present disclosure or the prior art, the drawings needed to be used in the embodiment or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments described in the present disclosure, and for those skilled in the art, other drawings can also be obtained without creative labor under the premise of the drawings.

[0009] Figure 1 The figure is a schematic diagram of the architecture of the locomotive automatic driving risk detection and emergency disposal system in the first embodiment of the present disclosure.

[0010] Figure 2 The figure is a safety personnel behavior monitoring flowchart in the first embodiment of the present disclosure.

[0011] Figure 3 The figure is a running environment perception detection flowchart in the first embodiment of the present disclosure.

[0012] Figure 4 The figure is a system health monitoring flowchart in the first embodiment of the present disclosure.

[0013] Figure 5 The figure is an emergency control response schematic diagram in the first embodiment of the present disclosure. DETAILED DESCRIPTION

[0014] In order for those skilled in the art to better understand the technical solutions in the one or more embodiments of the present specification, the technical solutions in the one or more embodiments of the present specification will be clearly and completely described in the following with reference to the drawings in the one or more embodiments of the present specification. Obviously, the described embodiments are only a part of the embodiments of the present specification, rather than all the embodiments. Based on the one or more embodiments of the present specification, all other embodiments obtained by those of ordinary skill in the art without creative labor should belong to the protection scope of the present document.

[0015] In order to solve the problems existing in the prior art, the first embodiment of the present disclosure provides an environment perception based automatic driving risk detection and emergency disposal system for locomotive, which is configured on a freight locomotive with a running track to realize risk detection and emergency disposal in the freight transportation process.

[0016] Figure 1 The architecture schematic diagram of the automatic driving risk detection and emergency disposal system for locomotive of the present embodiment is shown, which mainly includes the following parts: a personnel perception system 10, mainly used for perceiving and detecting the behavior of the safety officer in the locomotive cab and outputting a first detection result; a running environment perception system 20, used for perceiving and detecting the running track and / or the running environment of the locomotive and outputting a second detection result; an automatic driving system 30, used for determining the current running risk level of the locomotive according to the first detection result and / or the second detection result, and outputting a control instruction according to the running risk level; and a locomotive control system 40, used for executing the control instruction to complete the locomotive running and / or locomotive braking action. The functions of the systems in the present embodiment will be described in detail below. Figures 2 to 5 The functions of the systems in the present embodiment will be described in detail below.

[0017] At present, automatic driving of locomotive is still based on automatic driving and running of freight locomotive under the supervision of a person, so whether the safety officer in the cab normally and compliantly performs the duties is very important for the reliability and safety of automatic driving of the locomotive. Therefore, the present embodiment deploys a personnel perception system 10 to monitor the behavior of the safety officer in the cab, including but not limited to playing a mobile phone, dozing off, long-time burying head, long-time turning head, not on duty and other behaviors affecting safety.

[0018] Specifically, the personnel perception system 10 relies on the cooperation of image sensors and computer vision algorithms to monitor the behavior of the safety officer based on the closed-loop logic of "hardware acquisition-software analysis-decision warning". The personnel perception system 10 at least includes a first infrared camera and a personnel perception processing unit. The first infrared camera can be selected from a 200-wire pixel or more infrared high-definition camera to adapt to the complex light in the cockpit. The shooting range should cover the face and upper body operation area of the safety officer to avoid missing detection caused by too narrow viewing angle. In actual installation, the first infrared camera is usually fixed above the driver's seat or inside the front windshield, with a 45° downward angle to the safety officer's face, which can avoid blocking the safety officer's line of sight and clearly capture facial features (eyes, mouth) and body movements (hands, safety belt). In addition, a vibration compensation module can be configured to offset the vibration during train operation to avoid image blur, or a fill light can be configured to improve image shooting effect according to actual needs.

[0019] The personnel perception processing unit is used to receive the safety officer image collected by the first infrared camera in real time, extract the facial features and body features in the safety officer image, identify whether the safety officer has any of the behaviors of fatigue, distraction and illegal operation, and form a first detection result according to the behavior recognition result and the behavior duration. Specifically, Figure 2 The personnel perception processing unit can perform noise reduction (remove sensor noise), enhancement (contrast adjustment, highlight facial contours), distortion correction (correct lens perspective error) and other preprocessing operations on the safety officer image in priority, to provide clear input for subsequent analysis. Then, through lightweight target detection algorithms such as YOLOv5, the key targets such as "safety officer's head", "hand", "safety belt" in the image are located in real time, and the interference of irrelevant objects (such as cups, files) in the cockpit is excluded. Then, feature extraction is performed. The facial key points are extracted through Dlib or MTCNN algorithm, the facial features such as eye opening degree, head pitch angle and mouth opening degree are calculated, and the hand coordinates are extracted through the skeletal key point detection algorithm to judge whether the hand is in the operation area and whether the hand holds a mobile phone. After extracting the above features, the personnel perception processing unit can classify the feature sequence through a deep learning model. For example, according to the "closed eye duration" and "yawn frequency", the classification result of "fatigue" or "normal" is output, or through the time sequence change of the head turning angle and the hand position, combined with the rule base, it is judged as "distraction". The illegal operation of holding a mobile phone by hand can be directly judged through static body features.

[0020] In some embodiments, the railway industry safety behavior standards (such as TB / T3563-2020 "Railway Locomotive Driver Occupational Health Examination Standard") can be built into the personnel perception processing unit, and the judgment thresholds of various behaviors are set according to the above-mentioned standards, for example: closing eyes for more than 2 seconds or yawning more than 3 times per minute is judged as fatigue behavior; looking down for more than 5 seconds or turning head angle more than 45° for more than 3 seconds is judged as distraction behavior; not wearing safety belt for more than 10 seconds, both hands off the operation table for more than 15 seconds, holding a mobile phone with hands, etc. are judged as illegal operation. The personnel perception processing unit forms a first detection result according to the actual behavior recognition result and the behavior duration, and sends it to the automatic driving system 30.

[0021] The running environment perception system 20 mainly performs perception detection on the external environment in which the locomotive is located, mainly including the running track and the running environment. This system is based on "multi-sensor fusion + environment perception algorithm", and realizes precise monitoring of the line through the complementation of visual information of the camera and three-dimensional space information of the laser radar. Specifically, the running environment perception system 20 mainly includes a laser radar, a second infrared camera, a synchronization and calibration unit, and an edge computing unit; wherein the laser radar is arranged at the top of the front end of the locomotive, emits laser beams to the direction of locomotive travel, calculates the three-dimensional coordinates and distance of the target in front through the reflected light, and generates point cloud data; the second infrared camera is arranged below the laser radar, the shooting range overlaps the field of view of the laser radar, and is used to collect the environment image of the direction of locomotive travel, so as to identify the target category and texture features in the locomotive running environment; the synchronization and calibration unit is used to synchronize the point cloud data and the environment image through the time stamp, and complete the spatial calibration between the point cloud data and the environment image through the calibration board, to ensure the consistency of multi-source data; the edge computing unit carries a GPU to process point cloud and image data in real time, and is used to determine the track features and obstacle recognition results of the direction of locomotive travel according to the point cloud data and the environment image, and in the case of existing obstacles, compare whether the obstacles invade the preset line limit according to the preset line limit, calculate the invasion depth, area and predicted collision time, and form a second detection result according to the track features, invasion depth, area and predicted collision time.

[0022] In some embodiments, the edge computing unit specifically implements the running environment perception detection process as shown in Figure 3 , and specifically includes the following operations:

[0023] (1) Data preprocessing: for laser radar point cloud, ground points and noise points can be removed, and effective points within the line range (0-5m above the track surface, 3m on the left and right) can be retained; for environment image, distortion correction and exposure compensation can be performed, and image enhancement algorithm can be used to improve the detail clarity;

[0024] (2) Target detection and fusion: For the laser radar point cloud, based on the point cloud clustering algorithm, the points with close distances in space are aggregated into targets, and the envelope (length, width, height) of the target, the centroid coordinates and the relative speed (e.g. calculated by the position change of the continuous frame point cloud) are calculated; for the environment image, the obstacles (pedestrians, vehicles, falling rocks, etc.) and track features (rails, fasteners, switches) in the image are detected by a deep learning model (such as YOLOv8), and the target category and two-dimensional bounding box are output; then the target categories recognized by the camera are associated with the three-dimensional information detected by the laser radar using the calibration parameters, and the advantages of both are combined to improve the accuracy of target detection;

[0025] (3) Track and clearance analysis: The track line is fitted by the rail points in the point cloud whose reflection intensity exceeds a certain threshold, and the track gauge (the distance between the parallel lines of the left and right tracks) and the track surface continuity are calculated. If the track line in the continuous frame is interrupted, it is determined that the rail is broken; the small size point cloud feature of the fastener regularly arranged on the sleeper is identified, and the number of missing fasteners is counted. If more than 3 consecutive fasteners are missing, an early warning is triggered; based on the pre-set three-dimensional model of the line clearance (such as a cuboid region with a width of 4.88m and a height of 5.5m), it is compared in real time whether the target in the point cloud invades the region, and the invasion depth and area are calculated;

[0026] (4) Risk assessment: In the presence of obstacles, based on the distance of the obstacle target and the real-time speed of the locomotive, the collision time TTC = distance / relative speed (if the obstacle target is stationary, the relative speed = locomotive speed) is calculated.

[0027] In some embodiments, whether it is a laser radar or an infrared camera, the data collected by it may decrease in accuracy due to changes in the external environment, especially for weather such as rain, snow and heavy fog, which seriously affects the laser reflection effect and image collection accuracy. Therefore, the edge computing unit can also determine the track extension distance using the track features and judge the clarity of the environment image based on the data collected by the current laser radar and the second infrared camera, and output the detectable distance and / or the weather influence degree to reflect the influence of the external environment on the collection effect. Finally, the confidence of the invasion depth, the area and the predicted collision time and other parameters is determined using the detectable distance and / or the weather influence degree, and the confidence is also sent to the automatic driving system 30 as part of the second detection result when the second detection result is formed, so that the automatic driving system 30 can generate an automatic driving strategy in combination with the confidence during the automatic driving process.

[0028] The automatic driving system 30 of the embodiment is in communication connection with the personnel perception system 10, the operating environment perception system 20 and the locomotive control system 40, and mainly performs the operation risk judgment of the current locomotive in the driving process by receiving the first detection result and the second detection result, and performs the corresponding disposal by indicating the locomotive control system to avoid the risk. In addition, it should be noted that the automatic driving system 30 of the embodiment should also be able to generate the instructions of the locomotive control system 40 based on the conventional automatic driving algorithm, and the locomotive control system 40 can operate the brake, acceleration, traction and other systems of the locomotive according to the above instructions to complete the normal operation of the locomotive. The embodiment does not specifically limit and detail the above automatic driving algorithm and the way how the locomotive control system drives the locomotive to operate normally.

[0029] In the embodiment, the automatic driving system 30 at least includes an alarm device and an automatic driving control unit, wherein the alarm device is arranged in the cab and is used for sound and light alarm and / or voice alarm to the safety officer, and the automatic driving control unit is the control core of risk detection and emergency disposal, which combines the detection results of other perception systems and the preset protection operation strategy to evaluate the risk level and generate and execute the protection strategy. Specifically, the automatic driving control unit is used for: according to the behavior recognition result and the behavior duration contained in the first detection result, issuing a warning to the safety officer through the sound and light alarm device, and uploading the first detection result to the ground control center. In actual operation, the automatic driving control unit uses the alarm device installed in the cab to alarm the safety officer to improve his attention according to the behavior recognition result and the behavior duration contained in the first detection result, and uploads the behavior of the safety officer to the ground control center as the basis for the examination of the safety officer. At the same time, the automatic driving control unit can also judge whether the current safety officer has the ability to take over the locomotive according to the behavior recognition result and the behavior duration. If the safety officer is in a fatigue behavior, it is possible that the safety officer cannot take over the locomotive at present, if the safety officer is in a distraction behavior, it is determined that the safety officer has the ability to take over the locomotive, and if the safety officer has a rule violation at present, the locomotive can be taken over after the rule violation of the safety officer is eliminated.

[0030] The automatic driving control unit is also configured to determine a current operation risk type and an operation risk level of the locomotive according to the current operation speed of the locomotive, the track feature, and the predicted collision time, and determine an automatic protection strategy and a remaining response time according to the operation risk type, the operation risk level, and a preset protection operation; control the alarm device to deliver operation risk information to the safety officer, and in the case that no manual control action of the safety officer is received before the remaining response time arrives, autonomously execute the automatic protection strategy and generate a control instruction according to the automatic protection strategy. It should be noted that the current operation risk type and the operation risk level of the locomotive determined by the automatic driving control unit and the corresponding automatic protection strategy can be autonomously set according to different operation environments, different freight transport demands, and different types of locomotives.

[0031] For example, the current operation speed of the locomotive is 30 km / h, the track feature is a straight section, the predicted collision time is 12 seconds, and the obstacle is a small metal foreign matter invading the track limit (with an invasion depth of 0.3 meters). The automatic driving control unit determines that the current operation risk type is track foreign matter invasion, the operation risk level is level two (medium risk), and the corresponding generated automatic protection strategy is to reduce the speed to 15 km / h, with a remaining response time of 10 seconds. At this time, the automatic driving control unit can drive the alarm unit to trigger the sound and light alarm in the cab (yellow warning light flashing + intermittent buzzer prompt at intervals of 3 seconds), the on-board display screen pops up a diagram showing the position of the obstacle and a text prompt “suggest reducing the speed to 15 km / h”, and the voice broadcast “detected track foreign matter 60 meters ahead, please confirm whether to reduce the speed”. If no manual control action (such as brake pedal operation, traction system control operation, etc.) of the safety officer is received within 10 seconds, the automatic driving control unit automatically executes the protection strategy of reducing the speed to 15 km / h and generates a corresponding control instruction.

[0032] Alternatively, the current operation speed of the locomotive is 25 km / h, the track feature is a curve section (with a radius of 300 meters), the predicted collision time is zero (no obstacle), but the personnel perception system detects that the safety officer has been continuously looking down at the mobile phone for 15 seconds. The operation risk type is personnel operation disability, the operation risk level is level three (high risk), the corresponding generated automatic protection strategy is to temporarily limit the acceleration function (lock the maximum speed at the current speed), and the remaining response time is 7 seconds. At the same time, the alarm device is driven to trigger the high-frequency sound and light alarm (red warning light always on + continuous buzzer), and the voice is urgently broadcast “safety officer, please pay attention! Please pay attention to the driving state immediately!”. If no action of the safety officer is detected within 7 seconds, the automatic driving control unit automatically executes an emergency speed reduction to 10 km / h and continues to alarm.

[0033] Or, the current running speed of the locomotive is 50 km / h (the maximum limit speed), the track feature is a straight section in the tunnel, the predicted collision time is 10 seconds, the obstacle is a collapsed construction machinery (completely blocking the track), the automatic driving control unit determines that the running risk type is a fatal collision risk (extremely high risk), the running risk level is level four (the highest risk), the automatic protection strategy is to immediately start the maximum force air brake and simultaneously send an emergency avoidance signal to the rear train, and the remaining response time is 0 seconds, that is, to directly trigger the system autonomous emergency handling without waiting for manual response, and at the same time trigger the highest level of sound and light alarm (red flashing light + continuous high decibel alarm), and the voice is broadcasted synchronously "emergency! Large obstacle found in front, emergency braking soon!", prompting the safety officer to prepare for the collision accident handling.

[0034] At the same time, the automatic driving system 30 should also make real-time judgments and treatments for internal system function failures while sensing external environmental risks, for example, in the case of failures of the personnel perception system 10 and the running environment perception system 20, it may not be able to feedback the safety officer state or the external environment situation, or the detection result is wrong, which will affect the execution of the automatic driving algorithm and the judgment of the risk, at this time the automatic driving function becomes unreliable; if the locomotive control system 40 fails, it cannot accurately execute the instructions of the automatic protection strategy, which is easy to cause safety accidents. Therefore, the automatic driving system 30 of the embodiment also has the function of monitoring the health of other system functions, and real-time acquisition of safety risks and hidden dangers affecting the safe driving of automatic driving.

[0035] Specifically, the automatic driving control unit is also used to send health monitoring instructions to the personnel perception system 10, the running environment perception system 20 and the locomotive control system 40, and determine whether the system functions of the locomotive control system, the personnel perception system and the running environment perception system are normal according to the response of the health monitoring instructions; if the above systems can respond to the health monitoring instructions and send correct instruction confirmation feedback and instruction response feedback to the automatic driving control unit, it means that the system functions are normal; if the instruction response feedback can be received but the instruction confirmation feedback cannot be received, it means that the system communication is abnormal, if only the instruction confirmation feedback of the corresponding system can be received and the instruction response feedback cannot be received, or both feedbacks cannot be received, it means that the system function is abnormal, such as Figure 4 indicated.

[0036] In the case of abnormal system functions of the personnel perception system 10 and / or the running environment perception system 20, the automatic driving control unit cannot accurately know the outside situation and cannot give accurate automatic driving strategy, at this time the driving alarm device transmits the automatic driving function abnormal information to the safety officer and prompts the safety officer to take over the locomotive running control. If no manual control action of the safety officer is perceived within a certain time (for example, 15s), the automatic driving control unit ensures the safety of the locomotive through emergency braking.

[0037] In the event of a malfunction in the locomotive control system 40, the fault type of the locomotive control system is determined. Based on the fault type and preset protective operations, fault handling measures are determined and implemented. Simultaneously, the alarm device is controlled to transmit the fault handling measures to the safety officer. For example, if the fault type of the locomotive control system 40 is determined to be a braking control function fault, traction control function fault, or acceleration / deceleration function fault that affects driving safety, the corresponding fault handling measures include, but are not limited to, decelerating and stopping the locomotive by means of mechanical braking. If the fault type of the locomotive control system 40 is a communication function fault that does not affect driving safety, the safety officer may be prompted to carry out maintenance at an appropriate time simply through audible and visual warnings.

[0038] It should be noted that the preset protection operations in this embodiment refer to the collection of historical protection operations performed by the safety operator on the locomotive in the event of various emergencies under manual driving conditions. The specific content of the preset protection operations may vary depending on the vehicle type, the formation environment, and the freight transport requirements; this embodiment does not impose specific limitations. When generating automatic protection strategies or fault handling measures, the automatic driving control unit should, in addition to the preset protection operations, also adhere to the "principle of minimum intervention" and the "safety fallback principle." The former avoids over-response leading to efficiency losses, while the latter prioritizes stopping the vehicle in extreme situations.

[0039] In some embodiments, the automatic driving control unit is also configured to: display any one of the following on the vehicle's current operational risk type, operational risk level, fault type, and abnormal automatic driving function information on the onboard human-machine interface for confirmation by the safety operator; and receive the confirmation result from the safety operator on the onboard human-machine interface. In some cases, for warnings that the safety operator has determined do not require processing, the automatic driving control unit will not perform protective measures when it detects the corresponding warning again.

[0040] like Figure 1 As shown, the locomotive automatic driving risk detection and emergency response system of this embodiment also includes an emergency control processing system 50. This emergency control processing system 50 includes: safety relays connected to the locomotive's braking system and automatic driving control unit, respectively, and an emergency brake button configured in the driver's cab. Under normal circumstances, the automatic driving control unit outputs a redundant asynchronous dynamic signal to the safety relay. When the safety relay receives the redundant asynchronous dynamic signal output by the automatic driving control unit, it will not trigger the braking system. If the safety relay does not receive the redundant asynchronous dynamic signal or the redundant asynchronous dynamic signal does not meet the specifications, the safety relay triggers the braking system to achieve emergency braking protection for the locomotive. Alternatively, when the emergency brake button is pressed, the safety relay will also trigger the braking system to achieve emergency braking protection for the locomotive.

[0041] Specifically, as shown in Figure 5 In fact, there are two control loops between the automatic driving system and the braking system of the locomotive, one of which is implemented through the locomotive control system, and the other of which is implemented through the emergency control processing system. When the locomotive control system functions normally, the automatic driving control unit can drive the locomotive control system through a control instruction to operate the braking system of the locomotive to complete the corresponding braking operation; if the locomotive control system cannot complete the braking instruction due to abnormal function, the automatic driving control unit can suspend the output of the redundant asynchronous dynamic signal to trigger the safety relay to actuate the braking system; if the automatic driving system also has a fault or the safety officer actively intervenes, the safety relay can be directly driven through the emergency braking button to trigger the braking system to actuate, thereby ensuring the safety of the locomotive in a non-controlled state.

[0042] It should be noted that, in order to confirm that manual intervention safety operation can be achieved at any time of the system, the safety relay configured in the emergency control processing system 50 can achieve system isolation, forced emergency stop, etc., and the manual emergency control function still takes effect when the system is powered off.

[0043] In some embodiments, the locomotive automatic driving risk detection and emergency disposal system further includes a storage system (not shown in the figure), which is mainly used for storing full data of the personnel perception system, the operating environment perception system, the locomotive control system and the automatic driving system, including but not limited to collected data of various sensors, protection strategies, execution results, alarm conditions, etc., and can also include the states of various actuators, communication logs, etc., to provide a basis for subsequent fault analysis and system optimization.

[0044] The embodiment integrates personnel perception and environment perception, realizes omnidirectional detection of personnel behavior risk and external environment risk, solves the single perception problem of traditional systems, dynamically matches emergency strategies according to risk levels, avoids “one-size-fits-all” disposal, improves response accuracy, builds a man-machine cooperation mode, reserves manual response time after early warning, forms double safety protection, enhances fault tolerance through system health monitoring and emergency control redundancy design, ensures safe switching, and stores full data to provide support for accident tracing and system optimization. In summary, the system greatly improves the safety, reliability and intelligent level of automatic driving of the locomotive, and ensures efficient operation of rail transit.

[0045] Based on the same inventive concept, the second embodiment of the present disclosure provides a freight locomotive, which at least includes the locomotive automatic driving risk detection and emergency disposal system provided in the first embodiment of the present disclosure, to realize timely detection and rapid response of potential risks, greatly improve the safety and reliability of automatic driving of the locomotive, and provide strong guarantee for stable operation of rail transit.

[0046] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present disclosure, and are not intended to limit the present disclosure; although the present disclosure has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions recorded in the above embodiments can still be modified, or some or all of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present disclosure.

Claims

1. An environment-sensing-based locomotive automatic driving risk detection and emergency disposal system, characterized in that, The personnel perception system is used for perceiving and detecting the behaviors of the safety officer in the cab of the locomotive and outputting a first detection result. The running environment perception system is used for perceiving and detecting the running track and / or running environment of the locomotive and outputting a second detection result. The automatic driving system is used for determining the current running risk level of the locomotive according to the first detection result and / or the second detection result, and outputting a control instruction according to the running risk level. The locomotive control system is used for executing the control instruction to complete the locomotive running and / or locomotive braking action. The personnel perception system at least comprises:

2. The automatic driving risk detection and emergency disposal system for locomotive according to claim 1, characterized in that, The first infrared camera is arranged in the cab, and the shooting range covers the face and upper body operation area of the safety officer. The personnel perception processing unit is used for receiving the safety officer image collected by the first infrared camera in real time, extracting the face features and limb features in the safety officer image, identifying whether the safety officer has any one of the behaviors of fatigue, distraction and rule violation, and forming the first detection result according to the behavior identification result and the behavior duration. The running environment perception system at least comprises:

3. The automatic driving risk detection and emergency disposal system for locomotive according to claim 1, characterized in that, The laser radar is arranged at the top of the front end of the locomotive, emits laser beams to the direction of travel of the locomotive, calculates the three-dimensional coordinates and distance of the front target through reflected light, and generates point cloud data. The second infrared camera is arranged below the laser radar, and the shooting range overlaps the field of view of the laser radar, which is used for collecting the environment image of the direction of travel of the locomotive. The synchronization and calibration unit is used for synchronizing the point cloud data and the environment image through time stamp, and completing the spatial calibration between the point cloud data and the environment image through the calibration board. The edge computing unit is used for determining the track features and obstacle identification result of the direction of travel of the locomotive according to the point cloud data and the environment image, and in the case that there is an obstacle, comparing whether the obstacle invades the preset line limit according to the preset line limit in real time, calculating the invasion depth, area and predicted collision time, and forming the second detection result according to the track features, the invasion depth, the area and the predicted collision time. The edge computing unit is further used for:

4. The automatic driving risk detection and emergency disposal system for locomotive according to claim 3, characterized in that, Determining the track extension distance according to the track features; Outputting the detectable distance and / or weather influence degree according to the track extension distance and the definition of the environment image; Determining the confidence of the invasion depth, the area and the predicted collision time according to the detectable distance and / or the weather influence degree. The automatic driving system at least comprises:

5. The automatic driving risk detection and emergency disposal system for locomotive according to claim 1, characterized in that, The alarm device is arranged in the cab, and is used for acousto-optic alarm and / or voice alarm to the safety officer. The automatic driving control unit is used for: According to the behavior identification result and the behavior duration contained in the first detection result, issuing a warning to the safety officer through the acousto-optic alarm device, and uploading the first detection result to the ground control center. ​ determine a current operation risk type and an operation risk level of the locomotive according to the current operation speed of the locomotive, the track feature, and the predicted collision time, and determine an automatic protection strategy and a remaining response time according to the operation risk type, the operation risk level, and a preset protection operation; control the alarm device to deliver operation risk information to a safety officer, and autonomously execute the automatic protection strategy and generate a control instruction according to the automatic protection strategy in a case where no manual control action of the safety officer is received before the remaining response time arrives.

6. The automatic driving risk detection and emergency disposal system for locomotive according to claim 5, characterized in that, The automatic driving control unit is further configured to: send a health monitoring instruction to the locomotive control system, the personnel perception system, and the operation environment perception system, and determine whether system functions of the locomotive control system, the personnel perception system, and the operation environment perception system are normal according to responses to the health monitoring instruction; in a case where system functions of the locomotive control system are abnormal, determine a fault type of the locomotive control system, determine and execute a fault handling measure according to the fault type and a preset protection operation, and control the alarm device to deliver the fault handling measure to a safety officer; in a case where system functions of the personnel perception system and / or the operation environment perception system are abnormal, drive the alarm device to deliver automatic driving function abnormality information to a safety officer, and prompt the safety officer to take over locomotive operation control.

7. The automatic driving risk detection and emergency disposal system of a locomotive according to claim 6, characterized in that, The automatic driving control unit is further configured to: display any one of the current operation risk type, the operation risk level, the fault type, and the automatic driving function abnormality information of the locomotive on a vehicle-mounted human-machine interface for confirmation by a safety officer; receive a confirmation result of the safety officer on the vehicle-mounted human-machine interface.

8. The automatic driving risk detection and emergency disposal system of a locomotive according to claim 6, characterized in that, Further comprising: an emergency control processing system including a safety relay connected with the braking system of the locomotive and the automatic driving control unit, and an emergency brake button; wherein the safety relay does not trigger the braking system to act when receiving a redundant asynchronous dynamic signal output by the automatic driving control unit, and triggers the braking system to implement locomotive emergency stop protection when not receiving the redundant asynchronous dynamic signal or the redundant asynchronous dynamic signal does not conform to a specification; in a case where the emergency brake button is pressed, the safety relay triggers the braking system to implement locomotive emergency stop protection.

9. The automatic driving risk detection and emergency disposal system for locomotive according to any one of claims 1 to 8, characterized in that, Further comprising: a storage system configured to store full-amount data of the personnel perception system, the operation environment perception system, the locomotive control system, and the automatic driving system.

10. A freight locomotive characterized by, at least comprising the locomotive automatic driving risk detection and emergency disposal system according to any one of claims 1 to 9.

Images