Intelligent train sensing method and system

Abstract

The invention discloses an intelligent train sensing method and system. The intelligent train sensing method comprises the following steps: making an electronic map of an intelligent sensing system of each line; formulating a comprehensive decision-making mechanism of the obstacle detection result in the train advancing area; data are collected in real time through a detection unit, obstacles are detected, the obstacle detection result confidence Cd, the obstacle distance measurement result L and the distance measurement confidence Cl are output according to a train advancing area obstacle detection result comprehensive decision-making mechanism, an early warning area LC is dynamically adjusted, and the early warning level is researched and judged; an obstacle ranging result is mapped to an electronic map of the intelligent sensing system, whether the obstacle is within the vehicle limit range of current running of the train or not is judged, if yes, it is confirmed that driving safety is affected, and the final obstacle detection result, early warning and other levels are sent to the train control unit; if not, the detection result is not output. The problems of high sensing cost and low quality in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent sensing technology, and in particular relates to a train intelligent sensing method and system. Background Art

[0002] In the field of rail transit, common intelligent perception systems are composed of sensors, recognition hosts, and high-precision maps. The sensors are responsible for collecting information about the area where the train is traveling. In order to increase the data collection area, the sensors are generally installed on the roof, including using cameras to collect images, using lidar to collect three-dimensional point cloud data, and using millimeter-wave radar to collect speed information; high-precision maps are collected by data collectors riding on trains equipped with lidar and cameras, and then generated after post-processing, recognition, manual verification and other steps. When the system is running, it can be compared and analyzed with the sensor data detection results to assist in recognition; the core of the recognition host is the high-computing power GPU and obstacle detection algorithm.

[0003] The existing solutions have the following problems:

[0004] To build a high-precision map, each route must collect data and build a map, which makes the entire intelligent perception system costly. In real-time applications, the large amount of map data causes a lot of computer performance overhead when calling and identifying.

[0005] Since all types of sensors will show shortcomings under special circumstances, such as poor imaging at night, the recognition effect will be reduced; the recognition effect of lidar will be significantly reduced in special weather conditions such as rain and snow; millimeter-wave radar cannot classify objects due to its low resolution and is easily interfered. Summary of the invention

[0006] The purpose of the present invention is to overcome the problems of the prior art and disclose a train intelligent perception method and system, which solves the problems of high perception cost and low quality in the prior art.

[0007] The object of the present invention is achieved through the following technical solutions:

[0008] A train intelligent perception method, the train intelligent perception method comprising:

[0009] Make electronic maps of intelligent sensing systems for each line;

[0010] Formulate a comprehensive decision-making mechanism for obstacle detection results in the area ahead of the train;

[0011] Obstacles are detected by real-time data collection through the detection unit, and the obstacle detection result confidence Cd, obstacle ranging result L and ranging confidence Cl are output according to the comprehensive decision-making mechanism of the obstacle detection result in the train forward area, and the warning area LC is dynamically adjusted to determine the warning level;

[0012] The obstacle distance measurement results are mapped to the electronic map of the intelligent perception system to determine whether the obstacle is within the vehicle limit range of the train. If it is within the range, it is confirmed that it affects driving safety, and the final obstacle detection result and warning level are sent to the vehicle control unit; if it is not within the range, the detection result is not output.

[0013] In one or more possible embodiments, the process of making an electronic map of an intelligent perception system includes:

[0014] The train runs from the starting point to the end point of the line, and the Beidou coordinates of each position on the line are recorded in real time;

[0015] According to the position information of switches, signals, crossings and tracks in the line section provided by the line construction drawing or signal plan, the corresponding information and the position relationship of the line starting point are calculated to complete the binding of the Beidou positioning system coordinate position of the entire line with the line switches, signals, crossings, tracks and other equipment related to train operation.

[0016] In one or more possible embodiments, the comprehensive decision-making mechanism for obstacle detection results in the train forward area includes:

[0017] When the weather is detected to be normal,

[0018] The laser radar ranging data is used as the output obstacle ranging result L and ranging confidence Cl;

[0019] The confidence of the laser radar and the camera is averaged as the output obstacle detection result confidence Cd.

[0020] In one or more possible embodiments, the comprehensive decision-making mechanism for obstacle detection results in the train forward area includes:

[0021] When bad weather is detected,

[0022] The millimeter wave radar ranging data is used as the output obstacle ranging result L and ranging confidence Cl;

[0023] The confidence level of the millimeter-wave radar detection result is used as the output obstacle detection result confidence level Cd.

[0024] In one or more possible embodiments, weather condition determination is performed by a camera algorithm.

[0025] In one or more possible embodiments, when the warning area LC satisfies: The corresponding area is at level three warning.

[0026] In one or more possible embodiments, when the warning area LC satisfies: The corresponding area is at the second-level warning level.

[0027] In one or more possible embodiments, when the warning area LC satisfies: The corresponding area is at the first-level warning level.

[0028] In one or more possible embodiments, the detection unit includes a camera, a lidar, and a millimeter-wave radar, which is used to collect image data, three-dimensional point cloud data, and train speed data in the direction of the train's forward movement.

[0029] On the other hand, the present invention also discloses:

[0030] A train intelligent sensing system, comprising: a detection unit, a Beidou positioning system, an intelligent sensing host, and a train control unit;

[0031] The detection unit includes: a camera, a laser radar, and a millimeter wave radar, which are used to collect image data, three-dimensional point cloud data, and train speed data in the direction of the train's forward movement;

[0032] The Beidou positioning system is used to collect locomotive position data;

[0033] The intelligent sensing host uses the aforementioned train intelligent sensing method to complete intelligent sensing of obstacles, and sends the corresponding sensing results to the train control unit;

[0034] The empty car unit realizes train control based on the sensing result.

[0035] The aforementioned main scheme of the present invention and its further options can be freely combined to form multiple schemes, all of which are schemes that can be adopted and claimed for protection by the present invention. After understanding the scheme of the present invention, those skilled in the art can understand that there are multiple combinations based on the prior art and common knowledge, all of which are technical schemes to be protected by the present invention, and they are not exhaustively listed here.

[0036] Beneficial effects of the present invention:

[0037] The high-precision maps used in existing intelligent perception systems are collected by data collectors riding on trains equipped with laser radars and cameras, and then generated after post-processing, identification, manual verification and other steps. There are problems such as large map data volume, long production cycle, and the need to update the map when the line environment changes slightly. This application proposes a method for establishing an intelligent perception electronic map of trains based on the Beidou positioning system, which refines the factors that affect the operation of trains, namely the location information of switches, signal machines, crossings, and tracks between the starting point and the end point of the line, and binds the location information of these factors with the positioning data of the Beidou positioning system to form a lightweight electronic map of the intelligent perception system, which reduces the amount of map data while shortening the production cycle and cost. When the line environment changes, only the corresponding location mapping data needs to be modified, which is simple and convenient.

[0038] The existing perception system directly outputs the results or statically allocates the confidence and then outputs the results. The mechanism cannot adapt to the impact of meteorological conditions. With a fixed decision-making mechanism, there is still the possibility of outputting "false" safety information and "false" safety distance in uncertain scenarios. The present invention proposes a comprehensive decision-making mechanism for obstacle detection results in the train forward area, which outputs the obstacle detection results after comprehensive analysis based on the dynamic confidence and dynamic warning area, avoiding the output of unsafe results and causing accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of the train intelligent perception method of the present invention. DETAILED DESCRIPTION

[0040] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0041] refer to Figure 1 As shown, the figure shows a train intelligent perception method, which includes the following steps.

[0042] Step S1: Create an electronic map of each route's intelligent sensing system.

[0043] The process of making the electronic map of the intelligent perception system includes: running from the starting point of the train line to the end point, during which the Beidou coordinate position of each position on the line is recorded in real time; according to the position information of the switches, signals, crossings, and tracks in the line section provided by the line construction drawing or signal plan, the corresponding information and the position relationship of the line starting point are calculated, and the Beidou positioning system coordinate position of the entire line is bound to the position of the line switches, signals, crossings, tracks and other equipment related to train travel.

[0044] If there are multiple routes, repeat the above steps until the production of the intelligent perception system electronic map of all routes is completed.

[0045] Step S2: Formulate a comprehensive decision-making mechanism for obstacle detection results in the train forward area. As shown in Table 1, the main factor affecting the performance of the intelligent perception system sensor is the weather condition.

[0046] Table 1

[0047]

[0048] Therefore, when the camera algorithm detects that the weather is normal (clear), the lidar ranging and positioning capability is optimal, and the lidar ranging data is used as the obstacle ranging result L and ranging confidence Cl output by the entire system; the lidar and camera obstacle detection capabilities are optimal, and the confidence of the two is averaged as the obstacle detection result confidence Cd output by the entire system.

[0049] When the camera algorithm detects that the weather is bad (snow, rain, fog, etc.), the millimeter-wave radar ranging and positioning capability is better, and the millimeter-wave radar ranging data is used as the obstacle ranging result L and ranging confidence Cl output by the entire system; the millimeter-wave radar detection capability is better, and the millimeter-wave radar detection result confidence is used as the obstacle detection result confidence Cd output by the entire system.

[0050] Step S3: Obstacles are detected by real-time data collection through the detection unit, and the obstacle detection result confidence Cd, obstacle ranging result L and ranging confidence Cl are output according to the comprehensive decision-making mechanism of the obstacle detection result in the train forward area, and the warning area LC is dynamically adjusted to determine the warning level.

[0051] In one or more possible embodiments, the detection unit includes a camera, a lidar, and a millimeter-wave radar, which is used to collect image data, three-dimensional point cloud data, and train speed data in the direction of the train's forward movement.

[0052] In one or more possible embodiments, when the warning area LC satisfies: The corresponding area is at level three warning.

[0053] In one or more possible embodiments, when the warning area LC satisfies: The corresponding area is at the second-level warning level.

[0054] In one or more possible embodiments, when the warning area LC satisfies: The corresponding area is at the first-level warning level.

[0055] Step S4: Map the obstacle distance measurement result to the electronic map of the intelligent perception system to determine whether the obstacle is within the vehicle limit range of the train. If it is within the range, it is confirmed that it affects driving safety, and the final obstacle detection result and warning level are sent to the vehicle control unit; if it is not within the range, the detection result is not output.

[0056] On the other hand, the present invention also discloses: a train intelligent sensing system, the train intelligent sensing system comprising: a detection unit, a Beidou positioning system, an intelligent sensing host, and a train control unit;

[0057] The detection unit includes: a camera, a laser radar, and a millimeter wave radar, which are used to collect image data, three-dimensional point cloud data, and train speed data in the direction of the train's forward movement;

[0058] The Beidou positioning system is used to collect locomotive position data;

[0059] The intelligent sensing host uses the aforementioned train intelligent sensing method to complete intelligent sensing of obstacles, and sends the corresponding sensing results to the train control unit;

[0060] The empty car unit realizes train control based on the sensing result.

[0061] Compared with other perception systems that use laser radar and camera data to build high-precision maps, the present invention proposes a method for building train intelligent perception electronic maps based on the Beidou positioning system. The Beidou positioning system is installed on the train, and then the train runs from the starting point of the line to the end point, during which the Beidou coordinate positions of various positions on the line are recorded in real time; according to the position information of the switches, signal machines, crossings, tracks, etc. of the line section provided by the line construction drawing or signal plan, the above information and the position relationship of the line starting point are calculated, and the Beidou positioning system coordinate position of the entire line is bound to the line switches, signal machines, crossings, tracks, and other equipment related to train travel, forming a lightweight intelligent perception system electronic map suitable for the field of rail transit.

[0062] Compared with other perception systems that directly output results or statically allocate confidence and then output results, the present invention proposes a comprehensive decision-making mechanism for obstacle detection results in the train forward area. In order to solve the problem that the existing decision-making mechanism does not consider the problem of sensor detection performance degradation caused by changes in weather conditions, a comprehensive decision-making mechanism for obstacle detection results in the train forward area is proposed. First, use the camera to collect high-resolution environmental images for weather condition recognition, and then dynamically allocate the confidence of the detection results of three types of sensor data: camera, lidar, and millimeter-wave radar; second, based on the electronic map established above, dynamically adjust the horizontal and vertical warning areas of the train forward area; third, output the obstacle detection results after comprehensive analysis of the first and second dynamic confidence and dynamic warning areas. That is, the existing perception system uses a fixed decision-making mechanism, and there is still the output of "pseudo" safety information and "pseudo" safety distance in uncertain scenarios. The present invention proposes a comprehensive decision-making mechanism for obstacle detection results in the train forward area, which outputs obstacle detection results after comprehensive analysis of dynamic confidence and dynamic warning areas, avoiding the output of unsafe results and causing accidents.

[0063] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A train intelligent perception method, characterized in that: The train intelligent perception method comprises: Make electronic maps of intelligent sensing systems for each line; Formulate a comprehensive decision-making mechanism for obstacle detection results in the area ahead of the train; Obstacles are detected by real-time data collection through the detection unit, and the obstacle detection result confidence Cd, obstacle ranging result L and ranging confidence Cl are output according to the comprehensive decision-making mechanism of the obstacle detection result in the train forward area, and the warning area LC is dynamically adjusted to determine the warning level; The obstacle distance measurement results are mapped to the electronic map of the intelligent perception system to determine whether the obstacle is within the vehicle limit range of the train. If it is within the range, it is confirmed that it affects driving safety, and the final obstacle detection result and warning level are sent to the vehicle control unit; if it is not within the range, the detection result is not output.

2. The train intelligent perception method according to claim 1, characterized in that: The process of making electronic maps of intelligent perception system includes: The train runs from the starting point to the end point of the line, and the Beidou coordinates of each position on the line are recorded in real time; According to the position information of switches, signals, crossings and tracks in the line section provided by the line construction drawing or signal plan, the corresponding information and the position relationship of the line starting point are calculated to complete the binding of the Beidou positioning system coordinate position of the entire line with the line switches, signals, crossings, tracks and other equipment related to train operation.

3. The train intelligent sensing method according to claim 1, characterized in that: The comprehensive decision-making mechanism for obstacle detection results in the train forward area includes: When the weather is detected to be normal, The laser radar ranging data is used as the output obstacle ranging result L and ranging confidence Cl; The confidence of the laser radar and the camera is averaged as the output obstacle detection result confidence Cd.

4. The train intelligent sensing method according to claim 1, characterized in that: The comprehensive decision-making mechanism for obstacle detection results in the train forward area includes: When bad weather is detected, The millimeter wave radar ranging data is used as the output obstacle ranging result L and ranging confidence Cl; The confidence level of the millimeter-wave radar detection result is used as the output obstacle detection result confidence level Cd.

5. The train intelligent sensing method according to claim 3 or 4, characterized in that: The camera algorithm determines the weather conditions.

6. The train intelligent perception method according to claim 1, characterized in that: When the warning area LC meets: The corresponding area is at level three warning.

7. The train intelligent sensing method according to claim 1, characterized in that: When the warning area LC meets: The corresponding area is at the second-level warning level.

8. The train intelligent sensing method according to claim 1, characterized in that: When the warning area LC meets: The corresponding area is at the first-level warning level.

9. The train intelligent sensing method according to claim 1, characterized in that: The detection unit includes a camera, a laser radar, and a millimeter wave radar, which are used to collect image data, three-dimensional point cloud data, and train speed data in the direction of the train's forward movement.

10. A train intelligent sensing system, characterized in that: The train intelligent sensing system includes: a detection unit, a Beidou positioning system, an intelligent sensing host, and a train control unit; The detection unit includes: a camera, a laser radar, and a millimeter wave radar, which are used to collect image data, three-dimensional point cloud data, and train speed data in the direction of the train's forward movement; The Beidou positioning system is used to collect locomotive position data; The intelligent sensing host uses the train intelligent sensing method according to any one of claims 1 to 8 to complete the intelligent sensing of obstacles, and sends the corresponding sensing results to the train control unit; The empty car unit realizes train control based on the sensing result.

Images