A safety alarm method and system for electric locomotive automatic driving

Abstract

The present invention relates to the field of locomotive driving alarm, and specifically discloses a method and system for automatic driving safety alarm of an electric locomotive. The method first determines a plurality of preset time nodes, and marks the interval duration between two adjacent preset time nodes as the node interval duration. Whenever the time reaches any preset time node, the driver is prompted to step on the pedal within the node interval duration between the preset time node and the next preset time node. If no stepping signal is received, it means that the driver has not stepped on the pedal within each node interval duration, and the driver may be in a drowsy state. A controller determines whether the driver is indeed in a drowsy state based on a real-time face video. If the driver is indeed in a drowsy state, the controller controls an audible and visual alarm to sound an alarm to remind the driver to stay awake and focused.

Description

Technical Field

[0001] The present invention relates to the technical field of locomotive travel alarms, and in particular to a method and system for automatically driving an electric locomotive to provide a safety alarm. Background Art

[0002] Unmanned driving (autonomous driving) is the integration of multiple high-tech disciplines and technologies in vehicles. Regarding the classification of autonomous driving, the SAE (Society of Automotive Engineers) classification standard is currently generally recognized internationally, which is divided into six levels from L0 to L5.

[0003] Since electric locomotives run on tracks, the road environment is relatively simple and there are relatively few external interference factors. Compared with complex road conditions outside, it is easier to achieve automatic driving. Existing electric locomotives (such as subway trains) are generally equipped with driving assistance functions, which can start, drive or brake by themselves under normal circumstances. However, the driver, as the controller of vehicle safety, needs to be able to intervene in the driving of the vehicle in real time, rather than relying entirely on automatic driving. Therefore, it is necessary to ensure that the driver stays awake at all times during driving, especially to prevent the driver from dozing off or falling asleep. However, existing automatic driving solutions cannot automatically identify the driver's driving status in real time, thereby preventing the driver from dozing off or falling asleep. Summary of the invention

[0004] The main purpose of the present invention is to provide an electric locomotive automatic driving safety alarm method and system, which aims to solve the problem that the existing automatic driving solutions cannot automatically and real-time identify the driver's driving status, thereby preventing the driver from dozing off and falling asleep.

[0005] The technical solution proposed by the present invention is:

[0006] A method for automatic driving safety alarm of an electric locomotive is applied to an automatic driving safety alarm system of an electric locomotive; the system comprises a controller, a camera module, a loudspeaker, an audible and visual alarm and a pedal arranged on the locomotive; the camera module, the loudspeaker and the audible and visual alarm are all communicatively connected to the controller; the pedal is provided with a contact sensor which is communicatively connected to the controller and is used to detect a stepping signal of the pedal; the method comprises:

[0007] The controller acquires a driving mode of the vehicle;

[0008] When the driving mode of the vehicle is the automatic driving mode, the controller determines a plurality of preset time nodes after the current moment, and marks the interval between two adjacent preset time nodes as the node interval duration, wherein the time interval between the first preset time node and the current moment is the first preset duration, and any two adjacent node interval durations are different;

[0009] At any preset time node, the controller generates and plays a prompt sound message through the speaker, wherein the prompt sound message is used to prompt the driver to step on the pedal within the node interval length between the preset time node and the next preset time node;

[0010] The controller determines whether a stepping signal from the contact sensor is received before the next preset time node arrives;

[0011] If no stepping signal is received from the contact sensor, the controller obtains the real-time facial video of the driver captured by the camera module, and determines whether the driver is in a drowsy state based on the real-time facial video;

[0012] If the user is in a dozing state, the controller controls the sound and light alarm to sound an alarm.

[0013] Preferably, the system further comprises a cloud server in communication with the controller; the cloud server stores standard facial images corresponding to each driver; the controller obtains the real-time facial video of the driver captured by the camera module, and determines whether the driver is in a drowsy state based on the real-time facial video, including:

[0014] The controller intercepts the real-time face video to obtain a video with the current moment as the end moment and the moment of the second preset time length before the current moment as the start moment, and marks it as a video to be analyzed;

[0015] The controller performs video frame decomposition on the video to be analyzed to obtain a plurality of target frame images;

[0016] The controller acquires a first actual distance and a second actual distance corresponding to the target frame image, wherein the first actual distance is the distance between a line connecting the left corner of the mouth and the right corner of the mouth of the face in the target frame image and the tip of the nose, and the second actual distance is the distance between a line connecting the left pupil center point and the right pupil center point of the face in the target frame image and the tip of the nose;

[0017] The controller obtains a standard face image of the driver corresponding to the target frame image from the cloud server, and obtains a first standard distance and a second standard distance corresponding to the standard face image, wherein the first standard distance is the distance between a line connecting a left corner of the mouth and a right corner of the mouth of the face in the standard face image and a nose tip point, and the second standard distance is the distance between a line connecting a left pupil center point and a right pupil center point of the face in the standard face image and a nose tip point;

[0018] The controller determines the actual state of the driver at the time corresponding to the target frame image based on the first actual distance, the second actual distance, the first standard distance, and the second standard distance, wherein the actual state includes a head-down state and a non-head-down state;

[0019] The controller determines whether the following condition is satisfied: the driver's actual state at the moment corresponding to the target frame image within a third consecutive preset time length is all in the head-down state;

[0020] If so, the controller determines that the driver is in a drowsy state.

[0021] Preferably, the controller determines the actual state of the driver at the moment corresponding to the target frame image based on the first actual distance, the second actual distance, the first standard distance and the second standard distance, including:

[0022] The controller calculates a first ratio and a second ratio corresponding to the target frame image based on the first actual distance, the second actual distance, the first standard distance, and the second standard distance:

[0023] ,

[0024] ,

[0025] In the formula, is the first ratio; is the first actual distance, is the first standard distance; is the second ratio; is the second actual distance, is the second standard distance;

[0026] The controller determines whether the following conditions are met: the first ratio and the second ratio are both less than a first preset value, and no nostrils are detected in the target frame image corresponding to the first ratio;

[0027] If so, the controller determines that the actual state of the driver at the moment corresponding to the target frame image is a head-down state;

[0028] If not, the controller determines that the actual state of the driver at the moment corresponding to the target frame image is a non-head-down state.

[0029] Preferably, the controller obtains a standard face image of the driver corresponding to the target frame image from the cloud server, and obtains a first standard distance and a second standard distance corresponding to the standard face image, and then further includes:

[0030] The controller calculates a first ratio of each target frame image based on a first standard distance and a first actual distance of each target frame image to obtain a plurality of first ratios, wherein the value of the first ratio is a ratio of the first actual distance of the target frame image to the first standard distance:

[0031] The controller establishes a corresponding relationship between the first ratio of each target frame image and the shooting time of the target frame image;

[0032] The controller determines whether a nostril is detected in the target frame image corresponding to the smallest first ratio through image analysis;

[0033] If not, the controller marks the smallest first ratio as the first target value, and determines whether there is a second target value in each of the first ratios that satisfies the following conditions: the two first ratios adjacent to the second target value are both greater than the second target value, the second target value is different from the first target value, and no nostrils are detected in the target frame image corresponding to the second target value, wherein the first moment is the shooting moment corresponding to the first target value, and the second moment is the shooting moment corresponding to the second target value;

[0034] If so, the controller marks the first ratio value that takes a value of 1 after the first moment as the third target value, and marks the shooting moment corresponding to the third target value as the third moment, marks the first ratio value that takes a value of 1 before the first moment as the fourth target value, and marks the shooting moment corresponding to the fourth target value as the fourth moment;

[0035] The controller marks the first ratio value that takes a value of 1 after the second moment as a fifth target value, and marks the shooting moment corresponding to the fifth target value as a fifth moment, marks the first ratio value that takes a value of 1 before the second moment as a sixth target value, and marks the shooting moment corresponding to the sixth target value as a sixth moment;

[0036] The controller determines whether the driver is in a drowsy state based on the first target value, the first moment, the second target value, the second moment, the third target value, the third moment, the fourth target value, the fourth moment, the fifth target value, the fifth moment, the sixth target value and the sixth moment.

[0037] Preferably, the controller determines whether the driver is in a drowsy state based on the first target value, the first moment, the second target value, the second moment, the third target value, the third moment, the fourth target value, the fourth moment, the fifth target value, the fifth moment, the sixth target value and the sixth moment, including:

[0038] The controller obtains the interval duration between the first moment and the third moment, and marks it as the first interval duration, and calculates the first slope value based on the first target value, the third target value and the first interval duration:

[0039] ,

[0040] In the formula, is the first slope value, is the third target value, is the first target value, is the first interval duration;

[0041] The controller obtains the interval duration between the second moment and the fifth moment, marks it as the second interval duration, and calculates the second slope value based on the second target value, the fifth target value and the second interval duration:

[0042] ,

[0043] In the formula, is the second slope value, is the fifth target value, is the second target value, is the duration of the second interval;

[0044] The controller obtains the interval duration between the first moment and the fourth moment, and marks it as the third interval duration, and calculates the third slope value based on the first target value, the fourth target value and the third interval duration:

[0045] ,

[0046] In the formula, is the third slope value, is the fourth target value, is the duration of the third interval;

[0047] The controller obtains the interval duration between the second moment and the sixth moment, and marks it as the fourth interval duration, and calculates the fourth slope value based on the second target value, the sixth target value and the fourth interval duration:

[0048] ,

[0049] In the formula, is the fourth slope value, is the sixth target value, is the fourth interval duration;

[0050] The controller determines whether the driver is in a drowsy state based on the first slope value, the second slope value, the third slope value, and the fourth slope value.

[0051] Preferably, the controller determines whether the driver is in a drowsy state based on the first slope value, the second slope value, the third slope value and the fourth slope value, including:

[0052] The controller determines whether the following conditions are met: the first slope value and the second slope value are both greater than a second preset value, and the third slope value and the fourth slope value are both less than a third preset value, wherein the second preset value is greater than the third preset value, and the second preset value and the third preset value are both positive values;

[0053] If so, the controller determines that the driver is in a drowsy state.

[0054] Preferably, the system further comprises a cloud server in communication connection with the controller; the method further comprises:

[0055] The controller obtains weather forecast information through the cloud server, and determines whether there is an estimated rainfall area covering the track between the current position of the locomotive and the target site based on the weather forecast information;

[0056] If so, the controller obtains the estimated rainfall start time and the estimated rainfall end time of the estimated rainfall area based on the weather forecast information;

[0057] The controller marks the track segment covered by the estimated rainfall area as an estimated rainfall segment;

[0058] The controller marks the track segment between one end of the estimated rainfall segment close to the locomotive and the locomotive as a first target segment, and marks the track segment between one end of the estimated rainfall segment away from the locomotive and the target station as a second target segment;

[0059] The controller obtains the required time limit, and obtains the interval time between the estimated rainfall end time and the current time, and marks it as the fifth interval time, wherein the required time limit is the longest allowed time for the locomotive to travel to the target station from the current time;

[0060] The controller determines whether the fifth interval duration is greater than or equal to the required time limit;

[0061] If so, the controller determines a first driving speed corresponding to the first target segment, a second driving speed corresponding to the second target segment, and a third driving speed corresponding to the estimated rainfall segment.

[0062] Preferably, the controller determines a first driving speed corresponding to the first target segment, a second driving speed corresponding to the second target segment, and a third driving speed corresponding to the estimated rainfall segment, respectively, including:

[0063] The controller sets the first travel speed and the second travel speed to the maximum speed limit corresponding to the locomotive;

[0064] The controller calculates a third travel speed based on the required time limit, the maximum speed limit corresponding to the locomotive, the distance of the first target segment, the distance of the second target segment, and the distance of the estimated rainfall segment:

[0065] ,

[0066] In the formula, is the third driving speed; is the distance of the first target segment; is the distance of the second target segment; The distance of the estimated rainfall section; To request a time limit; The maximum speed limit corresponding to the locomotive.

[0067] Preferably, the system further comprises a rainfall detection module disposed on the locomotive, the rainfall detection module being in communication with the controller; the controller respectively determines a first driving speed corresponding to the first target segment, a second driving speed corresponding to the second target segment, and a third driving speed corresponding to the estimated rainfall segment, and then further comprises:

[0068] When the locomotive is traveling in the estimated rainfall section, the controller obtains the real-time rainfall at the current position of the train through the rainfall detection module, and obtains the maximum speed limit corresponding to the real-time rainfall;

[0069] When the third running speed is higher than the maximum speed limit corresponding to the real-time rainfall, the controller controls the locomotive to decelerate, and controls the locomotive to run at the maximum speed limit corresponding to the real-time rainfall.

[0070] The present invention also proposes an electric locomotive automatic driving safety alarm system, which applies the electric locomotive automatic driving safety alarm method; the system includes a controller, a camera module, a loudspeaker, an audible and visual alarm and a pedal arranged on the locomotive; the camera module, the loudspeaker and the audible and visual alarm are all communicatively connected to the controller; the pedal is provided with a contact sensor which is communicatively connected to the controller and is used to detect a pedal stepping signal.

[0071] Through the above technical solution, the following beneficial effects can be achieved:

[0072] The electric locomotive automatic driving safety alarm method proposed in the present invention can automatically identify the driver's driving status in real time, and alarm when the driver is tired and relaxed, so as to remind the driver to stay focused; first, multiple preset time nodes are determined, and the interval duration between two adjacent preset time nodes is marked as the node interval duration. During the driving process of the locomotive, whenever the time comes to any preset time node, the controller generates and plays a prompt sound information through a speaker to prompt the driver to step on the pedal within the node interval duration between the current preset time node and the next preset time node; if no stepping signal is received, it means that the driver has not stepped on the pedal within each node interval duration, and the driver may be in a drowsy state. For the sake of accuracy, it is necessary to further confirm whether the driver is indeed in a drowsy state. The controller determines whether the driver is indeed in a drowsy state based on the real-time face video. If the driver is indeed in a drowsy state, the controller controls the sound and light alarm to alarm to remind the driver to stay awake and focused. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0074] Figure 1 This is a flow chart of the first embodiment of the electric locomotive automatic driving safety alarm method proposed by the present invention. DETAILED DESCRIPTION

[0075] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0076] The present invention provides a safety alarm method and system for automatic driving of an electric locomotive.

[0077] As attached Figure 1 As shown, in a first embodiment of a method for automatic driving safety alarm of an electric locomotive proposed by the present invention, the method is applied to an automatic driving safety alarm system of an electric locomotive; the system includes a controller, a camera module, a loudspeaker, an audible and visual alarm and a pedal of a locomotive (the camera module, the loudspeaker, the audible and visual alarm and the pedal are all arranged in the cab of the locomotive, and the camera module is used to take a real-time facial image of the driver, and the camera module is facing the driver's face); the camera module, the loudspeaker and the audible and visual alarm are all communicatively connected to the controller; the pedal is provided with a contact sensor which is communicatively connected to the controller and is used to detect a pedal stepping signal; this embodiment includes the following steps:

[0078] Step S110: The controller obtains a driving mode of the vehicle, wherein the driving mode includes an automatic driving mode and a manual driving mode.

[0079] Step S120: When the vehicle's driving mode is the automatic driving mode, the controller determines multiple preset time nodes after the current moment, and marks the interval between two adjacent preset time nodes as the node interval duration, wherein the time interval between the first preset time node and the current moment is the first preset duration (preferably 3 seconds, where the first preset duration is a preparatory duration, so that the subsequent controller can generate a prompt sound message), and the interval between any two adjacent nodes is not the same.

[0080] Specifically, the total duration of the multiple preset time nodes here is determined as the time required to travel from the starting station to the target station. Setting the interval between any two adjacent nodes to be different can prevent the driver from mechanically and regularly stepping on the pedals to avoid drowsiness fatigue detection.

[0081] Step S130: At any preset time node, the controller generates and plays a prompt sound message through the speaker, wherein the prompt sound message is used to prompt the driver to step on the pedal within the node interval duration between the preset time node and the next preset time node.

[0082] Specifically, a sound prompt is used to remind the driver to step on the pedal within the node interval between the current preset time node and the next preset time node. If the driver steps on the pedal as required, it proves that the driver's attention is focused and he is not drowsy.

[0083] Step S140: The controller determines whether a stepping signal from the contact sensor is received before the next preset time node arrives.

[0084] Step S150: If no stepping signal is received from the contact sensor, the controller obtains the real-time facial video of the driver captured by the camera module, and determines whether the driver is in a drowsy state based on the real-time facial video.

[0085] Specifically, if no stepping signal is received, it means that the driver may be drowsy, but for the sake of accuracy, it is necessary to further confirm whether the driver is indeed drowsy. In this case, the controller needs to determine whether the driver is drowsy based on real-time face video.

[0086] Step S160: If the user is in a dozing state, the controller controls the sound and light alarm to sound an alarm.

[0087] The electric locomotive automatic driving safety alarm method proposed in the present invention can automatically identify the driver's driving status in real time, and alarm when the driver is tired and relaxed, so as to remind the driver to stay focused; first, multiple preset time nodes are determined, and the interval duration between two adjacent preset time nodes is marked as the node interval duration. During the driving process of the locomotive, whenever the time comes to any preset time node, the controller generates and plays a prompt sound information through a speaker to prompt the driver to step on the pedal within the node interval duration between the current preset time node and the next preset time node; if no stepping signal is received, it means that the driver has not stepped on the pedal within each node interval duration, and the driver may be in a drowsy state. For the sake of accuracy, it is necessary to further confirm whether the driver is indeed in a drowsy state. The controller determines whether the driver is indeed in a drowsy state based on the real-time face video. If the driver is indeed in a drowsy state, the controller controls the sound and light alarm to alarm to remind the driver to stay awake and focused.

[0088] In a second embodiment of a safety alarm method for electric locomotive automatic driving proposed by the present invention, based on the first embodiment, the system further includes a cloud server in communication connection with the controller; the cloud server stores standard face images corresponding to each driver; step S150 includes the following steps:

[0089] Step S210: The controller intercepts the real-time face video to obtain a video with the current moment as the end moment and the moment of a second preset time length (for example, 8 seconds) before the current moment as the start moment, and marks it as the video to be analyzed.

[0090] Specifically, by analyzing the driver's facial video within a second preset time period (the second preset time period here is a relatively short time period) before the current moment, it is determined whether the driver is in a drowsy state.

[0091] Step S220: the controller decomposes the video to be analyzed into video frames to obtain a plurality of target frame images.

[0092] Step S230: The controller obtains a first actual distance and a second actual distance corresponding to the target frame image, wherein the first actual distance is the distance between a line connecting the left corner of the mouth and the right corner of the mouth of the face in the target frame image and the tip of the nose, and the second actual distance is the distance between a line connecting the left pupil center point and the right pupil center point of the face in the target frame image and the tip of the nose.

[0093] Step S240: The controller obtains a standard facial image of the driver from the cloud server and corresponding to the target frame image, and obtains a first standard distance and a second standard distance corresponding to the standard facial image, wherein the first standard distance is the distance between a line connecting the left corner of the mouth and the right corner of the mouth of the face in the standard facial image and the tip of the nose, and the second standard distance is the distance between a line connecting the left pupil center point and the right pupil center point of the face in the standard facial image and the tip of the nose.

[0094] Specifically, the first standard distance here is the distance between the line between the left corner of the mouth and the right corner of the mouth of the driver and the tip of the nose when the driver's face is in an upright position, and the second standard distance here is the distance between the line between the center point of the left pupil and the center point of the right pupil of the driver's face and the tip of the nose when the driver's face is in an upright position.

[0095] Step S250: The controller determines the actual state of the driver at the moment corresponding to the target frame image based on the first actual distance, the second actual distance, the first standard distance and the second standard distance.

[0096] Specifically, when the driver is in a head-down state, the first actual distance will be smaller than the first standard distance, and the second actual distance will be smaller than the second standard distance. In this way, by comparing the first actual distance corresponding to the target frame image with the first standard distance, and the second actual distance of the target frame image with the second standard distance, it can be known that the driver is in a head-down state, and then determine whether the driver is in a sleepy state.

[0097] Step S260: The controller determines whether the following condition is satisfied: the driver's actual state at the moments corresponding to the target frame images within a third consecutive preset time length (eg, 2 seconds) is all in the head-down state.

[0098] If yes, execute step S270: the controller determines that the driver is in a drowsy state.

[0099] Specifically, under normal circumstances, the driver will not keep his head down for too long, but if he keeps his head down for 2 seconds, it means that the driver is drowsy.

[0100] In a third embodiment of a safety alarm method for electric locomotive automatic driving proposed by the present invention, based on the second embodiment, step S250 includes the following steps:

[0101] Step S310: The controller calculates a first ratio and a second ratio corresponding to the target frame image based on the first actual distance, the second actual distance, the first standard distance, and the second standard distance:

[0102] ,

[0103] ,

[0104] In the formula, is the first ratio; is the first actual distance, is the first standard distance; is the second ratio; is the second actual distance, is the second standard distance.

[0105] Step S320: The controller determines whether the following conditions are met: the first ratio and the second ratio are both smaller than a first preset value (eg, 0.6), and no nostrils are detected in the target frame image corresponding to the first ratio.

[0106] Specifically, when the driver is in a state of lowering his head, the first actual distance will be smaller than the first standard distance, and the second actual distance will be smaller than the second standard distance, so the corresponding first ratio will be less than 1, but when the driver is in a state of raising his head, the corresponding first ratio will also be less than 1. In order to exclude this situation, it is necessary to simultaneously satisfy the condition that no nostrils are detected in the target frame image corresponding to the first ratio (when the driver is in a state of raising his head, the corresponding target frame image will detect nostrils, and when the driver is in a state of lowering his head, the corresponding target frame image will not detect nostrils) in order to determine that the driver is in a state of lowering his head; in this embodiment, the first preset value is set to 0.6. If both the first ratio and the second ratio are less than 0.6, it is determined that the driver in the target frame image is in a state of lowering his head.

[0107] If so, executing step S330: the controller determines that the actual state of the driver at the moment corresponding to the target frame image is the head-down state.

[0108] If not, executing step S340: the controller determines that the actual state of the driver at the moment corresponding to the target frame image is a non-head-down state.

[0109] Specifically, on the contrary, it is determined that the driver in the target frame image is in a non-head-down state.

[0110] In a fourth embodiment of a safety alarm method for electric locomotive automatic driving proposed by the present invention, based on the second embodiment, step S240 further includes the following steps:

[0111] Step S410: The controller calculates a first ratio of each target frame image based on a first standard distance and a first actual distance of each target frame image to obtain multiple first ratios, wherein the first ratio is a ratio of the first actual distance of the target frame image to the first standard distance.

[0112] Specifically, a corresponding first ratio can be calculated for each target frame image here, and the first ratio can reflect whether the driver in the target frame image is in a head-down state (please refer to the third embodiment for details).

[0113] Step S420: the controller establishes a correspondence between the first ratio of each target frame image and the shooting time of the target frame image.

[0114] Step S430: The controller determines whether nostrils are detected in the target frame image corresponding to the smallest first ratio through image analysis.

[0115] If not, execute step S440: the controller marks the smallest first ratio as the first target value, and determines whether there is a second target value in each of the first ratios that satisfies the following conditions: the two first ratios before and after the second target value are both greater than the second target value, the second target value is different from the first target value, and no nostrils are detected in the target frame image corresponding to the second target value, wherein the first moment is the shooting moment corresponding to the first target value, and the second moment is the shooting moment corresponding to the second target value.

[0116] Specifically, if not, it means that the driver is in a head-down state in the target frame image corresponding to the smallest first ratio, so the smallest first ratio is marked as the first target value; and because the first ratio of the first target value is the smallest, the moment corresponding to the first target value here is the moment when the driver lowers his head to the lowest position (the smaller the first ratio, the lower the driver's head);

[0117] Similarly, the second target value here is another first ratio at the bottom, and because no nostrils are detected in the target frame image corresponding to the second target value, the second target value corresponds to the lowest position of another head-lowering action of the driver.

[0118] Specifically, since the first moment and the second moment here are both within the second preset time length, it means that the driver lowered his head at least twice in a relatively short period of time, which further proves that the driver was in a sleepy state.

[0119] Step S450: If it exists, the controller marks the first ratio that takes a value of 1 (the maximum value of the first ratio is 1) after the first moment as the third target value, and marks the shooting moment corresponding to the third target value as the third moment, marks the first ratio that takes a value of 1 before the first moment as the fourth target value, and marks the shooting moment corresponding to the fourth target value as the fourth moment.

[0120] Specifically, the third target value here is the first ratio which is immediately adjacent to the first target value and has a value of 1. Therefore, the driver in the target frame image corresponding to the third target value is in a state of facing the camera module (upright state); that is, the driver transitions from the head-down position at the first moment to the upright position at the third moment, that is, the driver is in a head-up action during the time period from the first moment to the third moment.

[0121] Similarly, the fourth target value here is the first ratio that is adjacent to the first target value in front and has a value of 1. Therefore, the driver in the target frame image corresponding to the fourth target value is in a state facing the camera module (upright state); that is, the driver transitions from the upright position at the fourth moment to the head-down position at the first moment, that is, the driver is in a head-down action during the time period from the fourth moment to the first moment.

[0122] Step S460: The controller marks the first ratio that takes a value of 1 after the second moment as the fifth target value, and marks the shooting moment corresponding to the fifth target value as the fifth moment, marks the first ratio that takes a value of 1 before the second moment as the sixth target value, and marks the shooting moment corresponding to the sixth target value as the sixth moment.

[0123] Specifically, the fifth target value here is the first ratio that is immediately adjacent to the second target value and has a value of 1. Therefore, the driver in the target frame image corresponding to the fifth target value is in a state of facing the camera module (upright state); that is, the driver transitions from the head-lowering position at the second moment to the upright position at the fifth moment, that is, the driver is in a head-up action during the time period from the second moment to the fifth moment.

[0124] Similarly, the sixth target value here is the first ratio that is adjacent to the first target value in front and has a value of 1. Therefore, the driver in the target frame image corresponding to the sixth target value is in a state facing the camera module (upright state); that is, the driver transitions from the upright position at the sixth moment to the head-down position at the second moment, that is, the driver is in a head-down action during the time period from the sixth moment to the second moment.

[0125] Step S470: The controller determines whether the driver is in a drowsy state based on the first target value, the first moment, the second target value, the second moment, the third target value, the third moment, the fourth target value, the fourth moment, the fifth target value, the fifth moment, the sixth target value and the sixth moment.

[0126] From this, we can see that the driver lowered his head and raised his head twice; when a person is drowsy, he should lower his head slowly and raise his head quickly; therefore, by analyzing the corresponding rates of these two head-lowering and head-raising actions, we can determine whether the driver meets the conditions of slowly lowering his head and quickly raising his head. If so, it means that the driver is drowsy.

[0127] In a fifth embodiment of a safety alarm method for electric locomotive automatic driving proposed by the present invention, based on the fourth embodiment, step S470 includes the following steps:

[0128] Step S510: The controller obtains the interval duration between the first moment and the third moment, marks it as the first interval duration, and calculates the first slope value based on the first target value, the third target value and the first interval duration:

[0129] ,

[0130] In the formula, is the first slope value, is the third target value, is the first target value, The first interval duration.

[0131] Specifically, since the driver is in the process of raising his head during the period from the first moment to the third moment, the first slope value here can reflect the speed of the driver in this raising his head action. The larger the first slope value is, the faster the driver is in this raising his head action. And since the third target value is definitely greater than the first target value, the first slope value here must be a positive value.

[0132] Step S520: The controller obtains the interval duration between the second moment and the fifth moment, and marks it as the second interval duration, and calculates the second slope value based on the second target value, the fifth target value and the second interval duration:

[0133] ,

[0134] In the formula, is the second slope value, is the fifth target value, is the second target value, The second interval duration.

[0135] Specifically, since the driver is also in the process of raising his head during the period from the second moment to the fifth moment, the second slope value here can reflect the speed at which the driver is performing another raising his head. The larger the first slope value is, the faster the speed of the driver's another raising his head is. And since the fifth target value is definitely greater than the second target value, the second slope value here must be a positive value.

[0136] Step S530: the controller obtains the interval duration between the first moment and the fourth moment, marks it as the third interval duration, and calculates the third slope value based on the first target value, the fourth target value and the third interval duration:

[0137] ,

[0138] In the formula, is the third slope value, is the fourth target value, The third interval duration.

[0139] Specifically, since the driver is in the head-down action during the period from the fourth moment to the first moment, the third slope value here can reflect the speed of the driver in this head-down action. The larger the third slope value, the faster the driver's head-down action. And since the fourth target value is definitely greater than the first target value, the third slope value here must be a positive value.

[0140] Step S540: the controller obtains the interval duration between the second moment and the sixth moment, marks it as the fourth interval duration, and calculates the fourth slope value based on the second target value, the sixth target value and the fourth interval duration:

[0141] ,

[0142] In the formula, is the fourth slope value, is the sixth target value, This is the fourth interval duration.

[0143] Specifically, since the driver is also in the head-down action during the period from the sixth moment to the second moment, the fourth slope value here can reflect the speed at which the driver is performing another head-down action. The larger the fourth slope value, the faster the driver's head-down action is. And since the sixth target value is definitely greater than the second target value, the fourth slope value here must be a positive value.

[0144] Step S550: The controller determines whether the driver is in a drowsy state based on the first slope value, the second slope value, the third slope value, and the fourth slope value.

[0145] Specifically, if the driver is in a drowsy state, the corresponding action is to lower the head slowly and raise the head quickly; the first slope value and the second slope value can reflect the driver's head-raising speed, and the third slope value and the fourth slope value can reflect the driver's head-lowering speed; therefore, by analyzing the first slope value, the second slope value, the third slope value and the fourth slope value, it can be determined whether the driver is in a drowsy state.

[0146] In a sixth embodiment of a safety alarm method for electric locomotive automatic driving proposed by the present invention, based on the fifth embodiment, step S550 includes the following steps:

[0147] Step S610: The controller determines whether the following conditions are met: the first slope value and the second slope value are both greater than a second preset value (preferably 1.2), and the third slope value and the fourth slope value are both less than a third preset value (preferably 0.7), wherein the second preset value is greater than the third preset value, and the second preset value and the third preset value are both positive values.

[0148] Specifically, since the first slope value, the second slope value, the third slope value and the fourth slope value are all positive values, and the larger the slope value, the faster the corresponding head raising or lowering speed, the second preset value is set to be greater than the third preset value in this implementation.

[0149] Therefore, if the first slope value and the second slope value are both greater than the second preset value, and the third slope value and the fourth slope value are both less than the third preset value, it means that the first slope value and the second slope value are greater than the third slope value and the fourth slope value; that is, it means that the driver's two head-raising actions are faster than the two head-lowering actions, that is, the driver's two head-lowering actions are relatively slower; it means that the driver lowered his head slowly and raised his head quickly, that is, the driver meets the situation of slowly lowering his head and quickly raising his head, so it can be judged that the driver is in a sleepy state.

[0150] If yes, execute step S620: the controller determines that the driver is in a drowsy state.

[0151] In a seventh embodiment of a safety alarm method for electric locomotive automatic driving proposed by the present invention, based on the first embodiment, the system further includes a cloud server in communication connection with the controller; this embodiment further includes the following steps:

[0152] Step S710: The controller obtains weather forecast information through the cloud server, and determines whether there is an estimated rainfall area covering the track between the current position of the locomotive and the target site based on the weather forecast information.

[0153] Step S720: If present, the controller obtains the estimated rainfall start time and the estimated rainfall end time of the estimated rainfall area based on the weather forecast information.

[0154] Step S730: The controller marks the track segment covered by the estimated rainfall area as an estimated rainfall segment.

[0155] Step S740: The controller marks the track segment between the end of the predicted rainfall segment close to the locomotive and the locomotive as the first target segment, and marks the track segment between the end of the predicted rainfall segment away from the locomotive and the target station as the second target segment.

[0156] Step S750: The controller obtains the required time limit, and obtains the interval duration between the estimated rainfall end time and the current time, and marks it as the fifth interval duration, wherein the required time limit is the longest allowed time for the locomotive to travel to the target site from the current time.

[0157] Step S760: The controller determines whether the fifth interval duration is greater than or equal to the required time limit.

[0158] Specifically, if the fifth interval duration is greater than or equal to the required time limit, it means that if the locomotive needs to reach the target site on time, it cannot avoid rain by driving slowly; therefore, the driving strategy is to drive as slowly as possible in the rainy area to ensure safety; drive as fast as possible in the rainless area to ensure that it can arrive at the target site on time; and based on this driving strategy, determine the first driving speed corresponding to the first target segment, the second driving speed corresponding to the second target segment, and the third driving speed corresponding to the estimated rainfall segment.

[0159] If so, execute step S770: the controller determines a first driving speed corresponding to the first target segment, a second driving speed corresponding to the second target segment, and a third driving speed corresponding to the estimated rainfall segment.

[0160] In an eighth embodiment of a safety alarm method for electric locomotive automatic driving proposed by the present invention, based on the seventh embodiment, step S770 includes the following steps:

[0161] Step S810: The controller sets the first travel speed and the second travel speed as the maximum speed limit corresponding to the locomotive.

[0162] Specifically, if both the first target section and the second target section are rain-free sections, the maximum speed limit corresponding to the locomotive is used as the first travel speed corresponding to the first target section and the second travel speed corresponding to the second target section.

[0163] Step S820: The controller calculates a third travel speed based on the required time limit, the maximum speed limit corresponding to the locomotive, the distance of the first target segment, the distance of the second target segment, and the distance of the estimated rainfall segment:

[0164] ,

[0165] In the formula, is the third driving speed; is the distance of the first target segment; is the distance of the second target segment; The distance of the estimated rainfall section; To request a time limit; The maximum speed limit corresponding to the locomotive.

[0166] Specifically, the meaning of the above formula is to minimize the driving speed in the estimated rainfall section (ie, the third driving speed) while ensuring that the locomotive can arrive at the target station on time.

[0167] In a ninth embodiment of a safety alarm method for automatic driving of an electric locomotive proposed by the present invention, based on the seventh embodiment, the system further includes a rain detection module provided in the locomotive, and the rain detection module is communicatively connected with the controller; step S770, and then further includes the following steps:

[0168] Step S910: When the locomotive travels in the estimated rainfall section, the controller obtains the real-time rainfall at the current position of the train through the rainfall detection module, and obtains the maximum speed limit corresponding to the real-time rainfall.

[0169] Specifically, in order to ensure safe driving, the greater the real-time rainfall, the lower the corresponding maximum speed limit, and the smaller the real-time rainfall, the higher the corresponding maximum speed limit.

[0170] Step S920: When the third travel speed is higher than the maximum speed limit corresponding to the real-time rainfall, the controller controls the locomotive to decelerate, and controls the locomotive to travel at the maximum speed limit corresponding to the real-time rainfall.

[0171] Step S930: When the third travel speed is lower than the maximum speed limit corresponding to the real-time rainfall, the controller controls the locomotive to accelerate, and controls the locomotive to travel at the maximum speed limit corresponding to the real-time rainfall.

[0172] Specifically, this embodiment enables the locomotive to adjust the real-time running speed based on the real-time rainfall in the rainy area, which is safer.

[0173] The present invention also proposes an electric locomotive automatic driving safety alarm system, which applies the electric locomotive automatic driving safety alarm method; the system includes a controller, a camera module, a loudspeaker, an audible and visual alarm and a pedal arranged on the locomotive; the camera module, the loudspeaker and the audible and visual alarm are all communicatively connected to the controller; the pedal is provided with a contact sensor which is communicatively connected to the controller and is used to detect a pedal stepping signal.

[0174] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0175] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A safety alarm method for automatic driving of an electric locomotive, characterized in that: Applicable to the electric locomotive automatic driving safety alarm system; the system comprises a controller, a camera module, a speaker, an audible and visual alarm and a pedal arranged on the locomotive; the camera module, the speaker and the audible and visual alarm are all communicatively connected to the controller; The pedal is provided with a contact sensor which is in communication with the controller and is used to detect a stepping signal of the pedal; the method comprises: The controller acquires a driving mode of the vehicle; When the driving mode of the vehicle is the automatic driving mode, the controller determines a plurality of preset time nodes after the current moment, and marks the interval between two adjacent preset time nodes as the node interval duration, wherein the time interval between the first preset time node and the current moment is the first preset duration, and any two adjacent node interval durations are different; At any preset time node, the controller generates and plays a prompt sound message through the speaker, wherein the prompt sound message is used to prompt the driver to step on the pedal within the node interval length between the preset time node and the next preset time node; The controller determines whether a stepping signal from the contact sensor is received before the next preset time node arrives; If no stepping signal is received from the contact sensor, the controller obtains the real-time facial video of the driver captured by the camera module, and determines whether the driver is in a drowsy state based on the real-time facial video; If the user is in a dozing state, the controller controls the sound and light alarm to sound an alarm; The system further includes a cloud server in communication with the controller; the cloud server stores standard facial images corresponding to each driver; the controller obtains a real-time facial video of the driver captured by the camera module, and determines whether the driver is in a drowsy state based on the real-time facial video, including: The controller intercepts the real-time face video to obtain a video with the current moment as the end moment and the moment of the second preset time length before the current moment as the start moment, and marks it as a video to be analyzed; The controller performs video frame decomposition on the video to be analyzed to obtain a plurality of target frame images; The controller acquires a first actual distance and a second actual distance corresponding to the target frame image, wherein the first actual distance is the distance between a line connecting the left corner of the mouth and the right corner of the mouth of the face in the target frame image and the tip of the nose, and the second actual distance is the distance between a line connecting the left pupil center point and the right pupil center point of the face in the target frame image and the tip of the nose; The controller obtains a standard face image of the driver corresponding to the target frame image from the cloud server, and obtains a first standard distance and a second standard distance corresponding to the standard face image, wherein the first standard distance is the distance between a line connecting a left corner of the mouth and a right corner of the mouth of the face in the standard face image and a nose tip point, and the second standard distance is the distance between a line connecting a left pupil center point and a right pupil center point of the face in the standard face image and a nose tip point; The controller determines the actual state of the driver at the time corresponding to the target frame image based on the first actual distance, the second actual distance, the first standard distance, and the second standard distance, wherein the actual state includes a head-down state and a non-head-down state; The controller determines whether the following condition is satisfied: the driver's actual state at the moment corresponding to the target frame image within a third consecutive preset time length is all in the head-down state; If so, the controller determines that the driver is in a drowsy state; The controller obtains a standard face image of the driver corresponding to the target frame image from the cloud server, and obtains a first standard distance and a second standard distance corresponding to the standard face image, and then further includes: The controller calculates a first ratio of each target frame image based on a first standard distance and a first actual distance of each target frame image to obtain a plurality of first ratios, wherein the value of the first ratio is a ratio of the first actual distance of the target frame image to the first standard distance: The controller establishes a corresponding relationship between the first ratio of each target frame image and the shooting time of the target frame image; The controller determines whether a nostril is detected in the target frame image corresponding to the smallest first ratio through image analysis; If not, the controller marks the smallest first ratio as the first target value, and determines whether there is a second target value in each of the first ratios that satisfies the following conditions: the two first ratios adjacent to the second target value are both greater than the second target value, the second target value is different from the first target value, and no nostrils are detected in the target frame image corresponding to the second target value, wherein the first moment is the shooting moment corresponding to the first target value, and the second moment is the shooting moment corresponding to the second target value; If so, the controller marks the first ratio value that takes a value of 1 after the first moment as the third target value, and marks the shooting moment corresponding to the third target value as the third moment, marks the first ratio value that takes a value of 1 before the first moment as the fourth target value, and marks the shooting moment corresponding to the fourth target value as the fourth moment; The controller marks the first ratio value that takes a value of 1 after the second moment as a fifth target value, and marks the shooting moment corresponding to the fifth target value as a fifth moment, marks the first ratio value that takes a value of 1 before the second moment as a sixth target value, and marks the shooting moment corresponding to the sixth target value as a sixth moment; The controller determines whether the driver is in a drowsy state based on the first target value, the first moment, the second target value, the second moment, the third target value, the third moment, the fourth target value, the fourth moment, the fifth target value, the fifth moment, the sixth target value and the sixth moment.

2. The electric locomotive automatic driving safety alarm method according to claim 1, characterized in that: The controller determines the actual state of the driver at the time corresponding to the target frame image based on the first actual distance, the second actual distance, the first standard distance, and the second standard distance, including: The controller calculates a first ratio and a second ratio corresponding to the target frame image based on the first actual distance, the second actual distance, the first standard distance, and the second standard distance: ， ， In the formula, is the first ratio; is the first actual distance, is the first standard distance; is the second ratio; is the second actual distance, is the second standard distance; The controller determines whether the following conditions are met: the first ratio and the second ratio are both less than a first preset value, and no nostrils are detected in the target frame image corresponding to the first ratio; If so, the controller determines that the actual state of the driver at the moment corresponding to the target frame image is a head-down state; If not, the controller determines that the actual state of the driver at the moment corresponding to the target frame image is a non-head-down state.

3. The electric locomotive automatic driving safety alarm method according to claim 1, characterized in that: The controller determines whether the driver is in a drowsy state based on the first target value, the first moment, the second target value, the second moment, the third target value, the third moment, the fourth target value, the fourth moment, the fifth target value, the fifth moment, the sixth target value, and the sixth moment, including: The controller obtains the interval duration between the first moment and the third moment, and marks it as the first interval duration, and calculates the first slope value based on the first target value, the third target value and the first interval duration: ， In the formula, is the first slope value, is the third target value, is the first target value, is the first interval duration; The controller obtains the interval duration between the second moment and the fifth moment, marks it as the second interval duration, and calculates the second slope value based on the second target value, the fifth target value and the second interval duration: ， In the formula, is the second slope value, is the fifth target value, is the second target value, is the duration of the second interval; The controller obtains the interval duration between the first moment and the fourth moment, and marks it as the third interval duration, and calculates the third slope value based on the first target value, the fourth target value and the third interval duration: ， In the formula, is the third slope value, is the fourth target value, is the duration of the third interval; The controller obtains the interval duration between the second moment and the sixth moment, and marks it as the fourth interval duration, and calculates the fourth slope value based on the second target value, the sixth target value and the fourth interval duration: ， In the formula, is the fourth slope value, is the sixth target value, is the fourth interval duration; The controller determines whether the driver is in a drowsy state based on the first slope value, the second slope value, the third slope value, and the fourth slope value.

4. The electric locomotive automatic driving safety alarm method according to claim 3, characterized in that: The controller determines whether the driver is in a drowsy state based on the first slope value, the second slope value, the third slope value, and the fourth slope value, including: The controller determines whether the following conditions are met: the first slope value and the second slope value are both greater than a second preset value, and the third slope value and the fourth slope value are both less than a third preset value, wherein the second preset value is greater than the third preset value, and the second preset value and the third preset value are both positive values; If so, the controller determines that the driver is in a drowsy state.

5. The electric locomotive automatic driving safety alarm method according to claim 1, characterized in that: The system further includes a cloud server in communication with the controller; the method further includes: The controller obtains weather forecast information through the cloud server, and determines whether there is an estimated rainfall area covering the track between the current position of the locomotive and the target site based on the weather forecast information; If so, the controller obtains the estimated rainfall start time and the estimated rainfall end time of the estimated rainfall area based on the weather forecast information; The controller marks the track segment covered by the estimated rainfall area as an estimated rainfall segment; The controller marks the track segment between one end of the estimated rainfall segment close to the locomotive and the locomotive as a first target segment, and marks the track segment between one end of the estimated rainfall segment away from the locomotive and the target station as a second target segment; The controller obtains the required time limit, and obtains the interval time between the estimated rainfall end time and the current time, and marks it as the fifth interval time, wherein the required time limit is the longest allowed time for the locomotive to travel to the target station from the current time; The controller determines whether the fifth interval duration is greater than or equal to the required time limit; If so, the controller determines a first driving speed corresponding to the first target segment, a second driving speed corresponding to the second target segment, and a third driving speed corresponding to the estimated rainfall segment.

6. The electric locomotive automatic driving safety alarm method according to claim 5, characterized in that: The controller determines a first driving speed corresponding to the first target segment, a second driving speed corresponding to the second target segment, and a third driving speed corresponding to the estimated rainfall segment, respectively, including: The controller sets the first travel speed and the second travel speed to the maximum speed limit corresponding to the locomotive; The controller calculates a third travel speed based on the required time limit, the maximum speed limit corresponding to the locomotive, the distance of the first target segment, the distance of the second target segment, and the distance of the estimated rainfall segment: ， In the formula, is the third driving speed; is the distance of the first target segment; is the distance of the second target segment; The distance of the estimated rainfall section; To request a time limit; The maximum speed limit corresponding to the locomotive.

7. The electric locomotive automatic driving safety alarm method according to claim 5, characterized in that: The system further includes a rainfall detection module disposed on the locomotive, the rainfall detection module being in communication with the controller; the controller respectively determines a first driving speed corresponding to the first target segment, a second driving speed corresponding to the second target segment, and a third driving speed corresponding to the estimated rainfall segment, and then further includes: When the locomotive is traveling in the estimated rainfall section, the controller obtains the real-time rainfall at the current position of the train through the rainfall detection module, and obtains the maximum speed limit corresponding to the real-time rainfall; When the third running speed is higher than the maximum speed limit corresponding to the real-time rainfall, the controller controls the locomotive to decelerate, and controls the locomotive to run at the maximum speed limit corresponding to the real-time rainfall.

8. An electric locomotive automatic driving safety alarm system, characterized in that: The electric locomotive automatic driving safety alarm method as described in any one of claims 1 to 7 is applied; the system includes a controller, a camera module, a speaker, an audible and visual alarm and a pedal arranged on the locomotive; the camera module, the speaker and the audible and visual alarm are all communicatively connected to the controller; the pedal is provided with a contact sensor which is communicatively connected to the controller and is used to detect a pedal stepping signal.

Images