Dynamic path planning method and system for train inspection robot and train inspection robot

Abstract

The invention provides a dynamic path planning method for a train inspection robot. The method comprises the following steps: S1, presetting all point location and route attributes on a tunnel, a train and a map; s2, identifying a train number, and calculating an optimal path of a current train parking direction; s3, train deviation is detected, a deviation value is obtained, and each preset point position value under the current train stopping position is adjusted; s4, according to different building modes inside and outside the tunnel, a robot driving path plan is dynamically established; and S5, the robot finely adjusts the associated driving tracks inside and outside the tunnel according to the dynamic planning path and the actual situation, reaches the target detection point, completes the detection task and returns to the charging point. The invention further discloses a dynamic path planning system of the train inspection robot and the train inspection robot. Configuration of attributes such as the train attribute, the tunnel attribute and the lifting platform is added, the path can be dynamically planned, the train states under different conditions are met, the train inspection flexibility is improved, and the field application requirement is met.

Description

Technical Field

[0001] The present invention relates to the technical field of train inspection robots, and in particular to a dynamic path planning method and system for a train inspection robot and the train inspection robot. Background Art

[0002] The train inspection robot is an intelligent inspection device designed for special industries such as subways and high-speed railways. It aims to replace manual labor to complete various complex inspection tasks. The train inspection robot uses advanced technologies such as autonomous navigation technology, multi-eye 3D detection technology, and AI deep learning to achieve efficient and accurate inspection operations, providing important guarantees for the safety and reliability of subway operations. Path planning is an indispensable key module for robots in the inspection process.

[0003] Current inspection robots all implement path planning according to preset routes, but this fixed path planning is not suitable for train inspection robots. The difference between train inspection robots and inspection robots is that the equipment that inspection robots need to inspect is fixed, but the trains inspected by train inspection robots are not fixed, and variable factors include: deviations in direction and stop position, etc. Summary of the invention

[0004] To solve the above problems, the present invention aims to propose a dynamic path planning method, system and train inspection robot for a train inspection robot, which can dynamically change the preset path planning and provide support and guarantee for the inspection work of the train inspection robot.

[0005] To achieve the above object, the technical solution of the present invention is achieved as follows:

[0006] A dynamic path planning method for a train inspection robot, the method comprising the following steps:

[0007] S1. Preset the tunnel, train and all points and route attributes on the map;

[0008] S2, identify the train number and calculate the optimal path for the current train stop direction;

[0009] S3, detect train deviation, obtain deviation value, and adjust each preset point value under the current train stop position;

[0010] S4. Dynamically establish the robot's driving path planning according to different construction methods inside and outside the tunnel;

[0011] S5. The robot follows the dynamically planned path, fine-tunes the driving trajectory associated with the inside and outside of the tunnel according to the actual situation, reaches the target detection point, completes the detection task, and returns to the charging point;

[0012] Furthermore, in step S1, the following specific process is included:

[0013] S11: generating a map outside the tunnel and multiple maps of the bottom of the train inside the tunnel, wherein the generated map includes a map number and an area number, wherein the area number is an upper area or a lower area, representing a map outside the tunnel or a map of the bottom of the train inside the tunnel, each map coordinate system is independent, and the map number of the same area number is unique;

[0014] S12, presetting the coordinates and attributes of each path point on the map and the attributes of the path line;

[0015] S13. Currently, there are two ways to transfer from outside the tunnel to inside the tunnel. The first way is through a ramp, and the second way is through a lifting platform.

[0016] Furthermore, in step S2, the following specific process is included:

[0017] S21: The train inspection robot receives the train number identification task issued by the backend control center server, and calculates the point closest to the robot's current coordinates and the preset path according to the preset path point information as the starting point;

[0018] Use the vector method to find the distance between the current coordinate point and each preset inspection route, obtain the route with the shortest distance that is less than a certain threshold, and then calculate the distance between the current coordinate point and the two points on this route respectively. The point with the shortest distance is the starting point.

[0019] The preset line where the current coordinate point is on the extension line of the preset inspection route is unavailable; the current coordinate point and the preset inspection route are not in the same area and are unavailable; the current coordinate point and the preset inspection route are in the same area but have different numbers and are unavailable; the current preset route is disabled and is unavailable;

[0020] S22: Use Dijkstra algorithm to calculate the shortest path from the starting point to the target point. In this shortest path set, delete the first path point to solve the problem of robot turning and causing collision due to the narrow road and the path point behind the robot.

[0021] S23: Pass each path point in turn, arrive at the target point, send control instructions to the robotic arm, and control the robotic arm to reach the preset position; control the 3D high-definition camera to take photos, and send them back to the background control center server. After algorithm recognition, the preset train attributes are used as the standard to obtain the train number that currently needs to be inspected, and store it in the database of the background control center server.

[0022] Furthermore, in step S3, the following specific process is included:

[0023] S31: The robot receives the train deviation detection task sent by the background control center server, arrives at the target point, sends control instructions to the robotic arm, controls the robotic arm to reach the preset position; controls the 3D high-definition camera to take several photos and sends them back to the background control center server.

[0024] S32: After algorithm recognition, the deviation value between the current train stop position and the preset train stop position is obtained and sent to the robot.

[0025] S33: The robot receives the deviation value and dynamically adjusts the coordinate position of each preset point on the map outside the tunnel according to the deviation value.

[0026] Furthermore, in step S4, the following specific process is included:

[0027] S41: Loading configuration file parameters preset in the robot system;

[0028] S42: The robot receives a task list sent to it by the backend control center server, including the train car number and the tunnel number, and saves the task list in the database in sequence;

[0029] S43: According to preset parameters, different construction methods outside the tunnel and inside the tunnel are switched, and different dynamic path planning is implemented.

[0030] Furthermore, the route and walking direction connecting the outside and inside of the tunnel are set; if it is a one-way tunnel, one virtual connection line needs to be configured, and if it is a two-way tunnel, two virtual connections need to be configured;

[0031] Among them, a one-way tunnel means that there is a downhill slope at one end of the train, and a two-way tunnel means that there are downhill slopes at both ends of the train;

[0032] When parking in the forward direction, the upper lift point and the lower tunnel head lift point are used as map switching points, a virtual line is configured between the two points, and the line speed is set to 0; the upper lift point is used as the starting point, and the lower tunnel head lift point is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking;

[0033] When parking in reverse, the upper lift point and the lower tunnel end lift point are used as map switching points, a virtual line is configured between these two points, and the line speed is set to 0; the upper lift point is used as the starting point, and the lower tunnel end lift point is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking.

[0034] Furthermore, it also includes:

[0035] When it is a one-way tunnel and forward parking, the inner point of the tunnel head and the front path point are used as map switching points, a virtual line is configured between these two points, and the line speed is set to 0; the inner point of the tunnel head is used as the starting point, and the front path point is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking;

[0036] When it is a one-way tunnel or reverse parking, the driving direction of the uphill and downhill routes needs to be changed. The reverse direction is changed to forward direction from the inner point of the tunnel head to the outer point of the tunnel head, and the reverse direction is changed to forward direction from the outer point of the tunnel head to the inner point of the tunnel head. The inner point of the tunnel head and the path point at the rear of the vehicle are used as map switching points, and a virtual line is configured between the two points, and the line speed is set to 0; the inner point of the tunnel head is used as the starting point, and the path point at the rear of the vehicle is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking;

[0037] When it is a two-way tunnel and forward parking, it is necessary to add a pair of map switching points on the basis of the one-way tunnel, namely the internal point at the end of the tunnel and the path point at the rear of the vehicle, configure a virtual line between these two points, and set the line speed to 0; the internal point at the end of the tunnel is used as the starting point, and the path point at the rear of the vehicle is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking;

[0038] When it is a two-way tunnel and reverse parking, on the basis of the one-way tunnel, the driving direction of the uphill and downhill road route at the other end needs to be changed, and the direction from the external point at the end of the tunnel to the internal point at the end of the tunnel is changed from reverse walking to forward walking, and the direction from the internal point at the end of the tunnel to the external point at the end of the tunnel is changed from forward walking to reverse walking; on the basis of the one-way tunnel, it is necessary to add a pair of map switching points, namely the internal point at the end of the tunnel and the path point at the front of the vehicle, configure a virtual line between these two points, and set the line speed to 0; the internal point at the end of the tunnel is used as the starting point, and the path point at the front of the vehicle is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking;

[0039] When the lift platform method is used, the train car number and tunnel number are obtained from the task, and the information in the preset parameter file is used as the standard to determine the train's stop direction and obtain the lift platform number bound to the tunnel number.

[0040] Furthermore, in step S5, the following specific process is included:

[0041] S51: Read a task from the database, the robot drives along the dynamically planned path, and calculates the arrival angle according to the actual position parameters returned by the navigation system;

[0042] Determine whether the next point is the target point. If it is the target point, the in-position angle is based on the preset value. If it is not the target point but a path point, then calculate the angle relative to the map based on the current robot coordinates returned by the navigation and the preset coordinates of the next path point. If the direction of the current path line is reverse, then determine whether the in-position angle is greater than 0. If it is greater than 0, subtract 180° from the in-position angle. Otherwise, add 180° to the in-position angle. If the current path line and the next path line are in opposite directions, then determine whether the in-position angle is greater than 0. If it is greater than 0, subtract 180° from the in-position angle. Otherwise, add 180° to the in-position angle.

[0043] S52: each time a path point is reached, its attributes are obtained, and relevant services are implemented according to the path point attributes;

[0044] Among them, when the ramp method is adopted, if the current attribute is a map-cutting point, a map-cutting instruction is sent to the navigation program, and after waiting for the navigation map to switch, the path point information returned by the navigation is obtained, and the sending of the next path point is adjusted according to the situation; if the current path point and the next path point are both map-cutting points, and they are driving from the lower area to the upper area, the next path point is deleted, and the next path point information is directly sent to the navigation program; thereby solving the problem of inaccurate navigation position after cutting the map from inside the tunnel to outside the tunnel, and the robot turning in a narrow tunnel;

[0045] Among them, when the lifting platform method is adopted, the attribute of the next point of the current position point is obtained. If it is a lifting platform point, it is determined whether the current parking area of ​​the lifting platform is consistent with the current area of ​​the robot. If the areas are inconsistent, the lifting platform is controlled to the area where the robot is located; after the robot arrives at the lifting platform point, the arrival results of the navigation system and the lifting platform control system are obtained. If one of the returns is incorrect, an abnormal warning is given. When both return to the position, the lifting platform is controlled to reach the lifting platform point of the specified area, and a map switching instruction is issued to the navigation program at the same time. Wait for the navigation map switching to be completed, obtain the arrival information of the path point returned by the navigation, and if it is in place, send the next path point; Among them, if the first lifting platform point is in the upper area and the second lifting platform point is in the lower area, it is necessary to record the coordinates of the lifting platform point in the lower area and save them in the robot database as the arrival coordinates of the lifting platform point in the lower layer when the robot moves from the lower area to the upper area;

[0046] S53, when the robot leaves the previous target point and moves to the next target point, the robot arm is controlled to move to the preset position of the next target point to improve the train inspection efficiency;

[0047] S54. When all tasks in the database are completed, the robot calculates the optimal path to return to the charging point according to the preset path, arrives at the charging point, and waits for the next task.

[0048] In order to achieve the above object, the present invention also provides a dynamic path planning system for a train inspection robot, comprising the following modules:

[0049] Input module: used to preset the attributes of all points and routes in tunnels, trains and maps;

[0050] Train path calculation module: used to identify the train number and calculate the optimal path for the current train stop direction;

[0051] Deviation correction module: used to detect train deviation, obtain deviation value, and adjust each preset point value under the current train stop position;

[0052] Robot path planning module: used to dynamically establish robot driving path planning according to different construction methods inside and outside the tunnel;

[0053] Navigation module: It is used for the robot to follow the dynamically planned path, fine-tune the driving trajectory related to the inside and outside of the tunnel according to the actual situation, reach the target detection point, complete the detection task, and return to the charging point.

[0054] In order to achieve the above-mentioned purpose, the present invention also provides a train inspection robot, which adopts the dynamic path planning method of the train inspection robot as mentioned above.

[0055] Beneficial effect: Since the direction of the train and the stop position each time are not fixed, the present invention adds configurations of train attributes, tunnel attributes, lifting platforms and other attributes, which can dynamically plan the path to meet the train status in different situations, improve the flexibility of train inspection, and meet on-site application needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the accompanying drawings:

[0057] Figure 1 This is a flow chart of a dynamic path planning method for a train inspection robot according to an embodiment of the present invention;

[0058] Figure 2 A schematic diagram of robot path points and path lines in a dynamic path planning method for a train inspection robot according to an embodiment of the present invention;

[0059] Figure 3 It is a schematic diagram of the walking trajectory of the train forward parking robot in the single-pit ramp construction mode in the dynamic path planning method of the train inspection robot according to the embodiment of the present invention;

[0060] Figure 4It is a schematic diagram of the walking trajectory of the train reverse parking robot in the single tunnel ramp construction mode in the dynamic path planning method of the train inspection robot according to the embodiment of the present invention;

[0061] Figure 5 It is a schematic diagram of the walking trajectory of the train forward parking robot in the double-pit ramp construction mode in the dynamic path planning method of the train inspection robot according to the embodiment of the present invention;

[0062] Figure 6 It is a schematic diagram of the walking trajectory of the train reverse parking robot in the double-pit ramp construction mode in the dynamic path planning method of the train inspection robot according to the embodiment of the present invention;

[0063] Figure 7 It is a schematic diagram of the walking trajectory of the train forward parking robot in the lifting platform construction mode in the dynamic path planning method of the train inspection robot according to the embodiment of the present invention;

[0064] Figure 8 It is a schematic diagram of the walking trajectory of the train reverse parking robot in the lifting platform construction mode in the dynamic path planning method of the train inspection robot according to the embodiment of the present invention;

[0065] Fig. 9 It is a structural schematic diagram of the dynamic path planning system of the train inspection robot according to an embodiment of the present invention. DETAILED DESCRIPTION

[0066] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0067] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0068] Example 1

[0069] See also Figure 1-8 : A dynamic path planning method for a train inspection robot, the method comprising the following steps:

[0070] S1. Preset the tunnel, train and all points and route attributes on the map;

[0071] S2, identify the train number and calculate the optimal path for the current train stop direction;

[0072] S3, detect train deviation, obtain deviation value, and adjust each preset point value under the current train stop position;

[0073] S4. Dynamically establish the robot's driving path planning according to different construction methods inside and outside the tunnel;

[0074] S5. The robot follows the dynamically planned path, fine-tunes the driving trajectory associated with the inside and outside of the tunnel according to the actual situation, reaches the target detection point, completes the detection task, and returns to the charging point;

[0075] It should be noted that since the direction of the train and the stop position each time are not fixed, this embodiment adds the configuration of train attributes, tunnel attributes, lift attributes and other attributes, which can dynamically plan the route to meet the train status in different situations, improve the flexibility of train inspection, and meet on-site application needs.

[0076] In a specific example, in step S1, the following specific process is included:

[0077] S11: generating a map outside the tunnel and multiple maps of the bottom of the train inside the tunnel, wherein the generated map includes a map number and an area number, wherein the area number is an upper area or a lower area, representing a map outside the tunnel or a map of the bottom of the train inside the tunnel, each map coordinate system is independent, and the map number of the same area number is unique;

[0078] It should be noted that two maps are generated using the configuration tool, denoted as Figure 1 peacefully Figure 2 ,land Figure 1 In the upper area, the Figure 2 In the lower area, the Figure 1 It is a map outside the tunnel. Figure 2 It is the bottom map of the train in the tunnel, and each map has an independent coordinate system;

[0079] S12, presetting the coordinates and attributes of each path point on the map and the attributes of the path line;

[0080] like Figure 2 The coordinates and attributes of waypoint 1 to waypoint N on the preset map and the attributes of configured routes 1 to route N are shown;

[0081] S13. Currently, there are two ways to transfer from outside the tunnel to inside the tunnel. The first way is through a ramp, and the second way is through a lifting platform.

[0082] It should be noted that if Figure 5 : When using method 1 (ramp), you need to configure the train attributes, tunnel attributes and route attributes.

[0083] Train attributes include train number, front car number, rear car number, front waypoint ID, rear waypoint ID, and map number.

[0084] Tunnel attributes include tunnel number, tunnel type (one-way, two-way), tunnel head external point ID, tunnel head internal point ID, tunnel tail external point ID, and tunnel tail internal point ID.

[0085] Except for the tunnel type attribute, all the above attributes are unique.

[0086] The front path point ID and the rear path point ID are the point IDs on the bottom map of the train in the tunnel; the tunnel head external point ID, tunnel head internal point ID, tunnel tail external point ID, and tunnel tail internal point ID are the point IDs on the map outside the tunnel, for example Figure 3 Point 1 is the external point ID of the tunnel head, point 2 is the internal point ID of the tunnel head, point 3 is the internal point ID of the tunnel tail, and point 4 is the external point ID of the tunnel tail.

[0087] In terms of actual physical position, the path point at the front of the vehicle (point 5) is approximately similar to the internal point at the head of the tunnel (point 2), the path point at the rear of the vehicle (point 8) is approximately similar to the internal point at the tail of the tunnel (point 3), and the distance between the path point at the front of the vehicle (point 2) and the front of the vehicle is approximately equal to the distance between the path point at the rear of the vehicle (point 3) and the rear of the vehicle.

[0088] When using mode 2 (lifting platform), if Figure 7 : Train properties and lift properties need to be configured.

[0089] Train attributes include the train number, the front car number, and the rear car number, all of which are unique.

[0090] The attributes of the lift platform are unique, including the lift platform number, the upper lift platform point ID, the lower tunnel head lift platform point ID, and the lower tunnel tail lift platform point ID. Among them, the upper lift platform point ID is the point ID on the map outside the tunnel, and the lower tunnel head lift platform point ID and the tunnel tail lift platform point ID are the point IDs on the bottom map of the train inside the tunnel, such as Figure 4 : Specifically, point 2 is the upper lift point ID, point 3 is the lower tunnel head lift point ID, and point 6 is the lower tunnel tail lift point ID. Among them, point 2 is the upper ground Figure 2 Point 3 and point 4 are the lower ground Figure 2 The point on.

[0091] In actual physical position, the upper lift point (point 2) is approximately similar to the lower tunnel head lift point (point 3), and the distance between the lower tunnel head lift point (point 3) and the front of the vehicle is approximately equal to the distance between the lower tunnel tail lift point (point 6) and the rear of the vehicle.

[0092] This information is stored in the robot system in the form of a file (xml), and the robot needs to load this information when it runs.

[0093] In order to meet the construction methods inside and outside the tunnels of different sites, namely the ramp and the lifting platform, this embodiment uses two path planning schemes, thereby improving the practicality of the robot.

[0094] In a specific example, in step S2, the following specific process is included:

[0095] S21: The train inspection robot receives the train number recognition task issued by the backend control center server, and calculates the point closest to the robot's current coordinates (horizontal x and vertical y) and the preset path according to the preset path point information as the starting point;

[0096] Use the vector method to find the distance between the current coordinate point and each preset inspection route, obtain the route with the shortest distance that is less than a certain threshold, and then calculate the distance between the current coordinate point and the two points on this route respectively. The point with the shortest distance is the starting point.

[0097] The preset line where the current coordinate point is on the extension line of the preset inspection route is unavailable; the current coordinate point and the preset inspection route are not in the same area and are unavailable; the current coordinate point and the preset inspection route are in the same area but have different numbers and are unavailable; the current preset route is disabled and is unavailable;

[0098] In the specific implementation, Figure 2 As shown, the robot needs to reach the target point 4 from the current position. The vector method is used to calculate the distance between the robot's current coordinate point and route 1, route 2, and finally route 5, and the route with the smallest distance that is less than a certain threshold is obtained. After calculation, route 2 should be selected, and the distance from the robot to point 2 and point 3 is further calculated respectively. The point with the shortest distance is the starting point. As shown in the figure, point 2 is the starting point.

[0099] S22: Use Dijkstra algorithm to calculate the shortest path from the starting point to the target point. In this shortest path set, delete the first path point to solve the problem of robot turning and causing collision due to the narrow road and the path point behind the robot.

[0100] In the specific implementation, Figure 2 As shown, the Dijkstra algorithm is used to calculate that the shortest path from the current position to the target point 4 is point 2 to point 3 to point 4. The robot is currently parked in the forward direction. If it wants to reach point 2, it needs to turn the direction. However, since the road is narrow, the robot's turning will cause a collision, so the first path point is deleted.

[0101] S23: Pass each path point in turn, arrive at the target point, send control instructions to the robotic arm, and control the robotic arm to reach the preset position; control the 3D high-definition camera to take photos, and send them back to the background control center server. After algorithm recognition, the preset train attributes are used as the standard to obtain the train number that currently needs to be inspected, and store it in the database of the background control center server.

[0102] From this, it can be seen that: this embodiment determines the attributes of the path point in advance. If it is a lift point, the lift is controlled to be in place in advance to reduce the lifting time of the lift; when the robot leaves the current target point to go to the next target point, the robotic arm is controlled to move to the target position at the same time. When the robot arrives at the next target point, it can immediately execute the task, thereby improving the inspection efficiency and shortening the inspection time.

[0103] In a specific example, in step S3, the following specific process is included:

[0104] S31: The robot receives the train deviation detection task sent by the background control center server, arrives at the target point, sends control instructions to the robotic arm, controls the robotic arm to reach the preset position; controls the 3D high-definition camera to take several photos and sends them back to the background control center server.

[0105] S32: After algorithm recognition, the deviation value between the current train stop position and the preset train stop position is obtained and sent to the robot.

[0106] S33: The robot receives the deviation value and dynamically adjusts the coordinate position of each preset point on the map outside the tunnel according to the deviation value.

[0107] In a specific example, in step S4, the following specific process is included:

[0108] S41: Loading configuration file parameters preset in the robot system;

[0109] S42: The robot receives a task list sent to it by the backend control center server, including the train car number and the tunnel number, and saves the task list in the database in sequence;

[0110] S43: According to preset parameters, different construction methods outside the tunnel and inside the tunnel are switched, and different dynamic path planning is implemented.

[0111] It should be noted that when using the ramp method, the train car number and tunnel number are obtained from the task, and the information in the preset parameter file is used as the standard to determine the train's stop direction and tunnel type; if the train car number is not in the preset parameter file, the return car number does not exist; if the channel number is not in the preset parameter file, the return channel number does not exist.

[0112] In a specific example, the route and walking direction connecting the outside and inside of the tunnel are set; if it is a one-way tunnel, one virtual connection line needs to be configured, and if it is a two-way tunnel, two virtual connections need to be configured;

[0113] Among them, a one-way tunnel means that there is a downhill slope at one end of the train, and a two-way tunnel means that there are downhill slopes at both ends of the train; Figure 3As shown: if there is only ramp 1 or only ramp 2, it is a one-way tunnel; if there are both ramp 1 and ramp 2, it is a two-way tunnel;

[0114] When parking in the forward direction, Figure 7 : The upper lift point 3 and the lower tunnel head lift point 4 are used as map switching points, a virtual line is configured between these two points, and the line speed is set to 0; the upper lift point 3 is used as the starting point, and the lower tunnel head lift point 4 is used as the end point. The route from the starting point to the end point is set to forward walking, and the route from the end point to the starting point is set to reverse walking;

[0115] When reverse parking, such as Figure 8 : The upper lift point 3 and the lower tunnel end lift point 7 are used as map switching points, a virtual line is configured between these two points, and the line speed is set to 0; the upper lift point 3 is used as the starting point, and the lower tunnel end lift point 7 is used as the end point. The route from the starting point to the end point is set to forward walking, and the route from the end point to the starting point is set to reverse walking.

[0116] In a specific example, it also includes:

[0117] When it is a one-way tunnel and forward parking, Figure 3 : The internal point 2 of the tunnel head and the front path point 3 are used as the map switching points. A virtual line is configured between these two points, and the line speed is set to 0; the internal point 2 of the tunnel head is used as the starting point, and the front path point 3 is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking;

[0118] When it is a one-way tunnel or reverse parking, Figure 4 : It is necessary to change the driving direction of the uphill and downhill routes, change the reverse direction from the internal point 2 of the tunnel head to the external point 1 of the tunnel head to the forward direction, and change the reverse direction from the external point 1 of the tunnel head to the internal point 2 of the tunnel head to the forward direction. Set the internal point 2 of the tunnel head and the rear path point 6 as the map switching points, configure a virtual line between these two points, and set the line speed to 0; the internal point 2 of the tunnel head is used as the starting point, and the rear path point 6 is used as the end point. The route from the starting point to the end point is set to forward walking, and the route from the end point to the starting point is set to reverse walking;

[0119] When it is a two-way tunnel and forward parking, Figure 5 : It is necessary to add a pair of map switching points based on the one-way tunnel, namely the internal point 3 at the end of the tunnel and the path point 8 at the rear of the vehicle, configure a virtual line between these two points, and set the line speed to 0; the internal point 3 at the end of the tunnel is used as the starting point, and the path point 8 at the rear of the vehicle is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking;

[0120] When it is a two-way tunnel and reverse parking, Figure 6 : On the basis of the one-way tunnel, it is necessary to change the driving direction of the uphill and downhill routes at the other end, and change the reverse walking direction from the external point 4 at the end of the tunnel to the internal point 3 at the end of the tunnel to the external point 4 at the end of the tunnel from the forward walking direction to the reverse walking direction; it is necessary to add a pair of map switching points on the basis of the one-way tunnel, namely the internal point 3 at the end of the tunnel and the front path point 5, configure a virtual line between these two points, and set the line speed to 0; the internal point 3 at the end of the tunnel is used as the starting point, and the front path point 5 is used as the end point. The route from the starting point to the end point is set to forward walking, and the route from the end point to the starting point is set to reverse walking;

[0121] When the lift platform method is used, the train car number and tunnel number are obtained from the task, and the information in the preset parameter file is used as the standard to determine the train's stop direction and obtain the lift platform number bound to the tunnel number.

[0122] In a specific example, in step S5, the following specific process is included:

[0123] S51: Read a task from the database, the robot drives along the dynamically planned path, and calculates the arrival angle according to the actual position parameters returned by the navigation system;

[0124] Determine whether the next point is the target point. If it is the target point, the in-position angle is based on the preset value. If it is not the target point but a path point, then calculate the angle relative to the map based on the current robot coordinates returned by the navigation and the preset coordinates of the next path point. If the direction of the current path line is reverse, then determine whether the in-position angle is greater than 0. If it is greater than 0, subtract 180° from the in-position angle. Otherwise, add 180° to the in-position angle. If the current path line and the next path line are in opposite directions, then determine whether the in-position angle is greater than 0. If it is greater than 0, subtract 180° from the in-position angle. Otherwise, add 180° to the in-position angle.

[0125] S52: each time a path point is reached, its attributes are obtained, and relevant services are implemented according to the path point attributes;

[0126] Among them, when using the ramp method, if the current attribute is a cut map point, such as Figure 3 As shown: When the robot reaches waypoint 2, it needs to go to waypoint 3. At this time, the attribute of waypoint 2 is a map cut point. The map cut command is sent to the navigation program, and the navigation map switch is completed. After obtaining the waypoint arrival information returned by the navigation, the next waypoint is adjusted according to the situation. If the current waypoint and the next waypoint are both map cut points, and they are driving from the lower area to the upper area, the next waypoint is deleted, and the next waypoint information is directly sent to the navigation program. Figure 3As shown: that is, the areas where waypoints 2 and 3 are located are determined. If waypoint 2 is in the upper area and waypoint 3 is in the lower area, waypoint 3 is sent to the navigation program. When the robot reaches waypoint 3, it needs to go to waypoint 2, which is from the lower area to the upper area. Then the next waypoint 2 is deleted and waypoint 1 is directly sent to the navigation program. This solves the problem of inaccurate navigation position after cutting the map from inside the tunnel to outside the tunnel, and the robot turning in a narrow tunnel.

[0127] When the lifting platform method is used, the properties of the next point of the current point are obtained. If it is a lifting platform point, such as Figure 7 As shown: that is, when the robot reaches path point 1, the current area is the upper area, and it needs to pass path point 2, path point 3, and path point 4. At this time, the attribute of path point 3 is a lift, and it is determined whether the current parking area of ​​the lift is consistent with the current area of ​​the robot, that is, the upper area. If the areas are inconsistent, the lift is controlled to the area where the robot is located; after the robot reaches lift point 3, the arrival results of the navigation system and the lift control system are obtained. If one of the returns is incorrect, an abnormal warning is given. When both return to the position, the lift is controlled to reach the lift point of the specified area, and a map switching instruction is issued to the navigation program at the same time, and the navigation map switching is completed. The path point arrival information returned by the navigation is obtained. If it is in place, the next path point 4 is issued; among them, if the first lift point 3 is in the upper area and the second lift point 4 is in the lower area, the coordinates of the lift point 4 in the lower area need to be recorded and saved in the robot database as the arrival coordinates of the lift point in the lower layer when the robot moves from the lower area to the upper area;

[0128] S53, when the robot leaves the previous target point and moves to the next target point, the robot arm is controlled to move to the preset position of the next target point to improve the train inspection efficiency;

[0129] S54. When all tasks in the database are completed, the robot calculates the optimal path to return to the charging point according to the preset path, arrives at the charging point, and waits for the next task.

[0130] In summary, this embodiment has the following technical effects:

[0131] 1. The direction of the train and the stop position each time are not fixed, so this embodiment adds the configuration of train attributes, tunnel attributes, and lift attributes, which can dynamically plan the path to meet the train status in different situations, improve the flexibility of train inspection, and meet the needs of on-site applications.

[0132] 2. In order to meet the construction methods inside and outside the tunnels of different venues, namely ramps and lifting platforms, this embodiment uses two path planning schemes, thereby improving the practicality of the robot.

[0133] 3. To improve inspection efficiency and shorten inspection time, this embodiment makes the following improvements:

[0134] (1) Determine the attributes of the path point in advance. If it is a lift point, control the lift to be in place in advance to reduce the lift time.

[0135] (2) When the robot leaves the current target point and moves to the next target point, the robot arm is controlled to move to the target position at the same time. When the robot reaches the next target point, the task can be executed immediately.

[0136] 4. Since the path for the robot to travel is narrow, this embodiment has made the following improvements:

[0137] (1) Under certain conditions, find the nearest path point and remove the first path point, and go directly to the next point to prevent the robot from turning around.

[0138] (2) When using the ramp method, the navigation of the lower map will not turn, but the navigation of the upper map will turn. When going from the lower map to the upper map, the cut-off point of the upper map is removed to prevent navigation collision due to turning.

[0139] (3) When using the lift platform method, the lower map is based on the bottom of the train, and the upper map is based on the tunnel. During driving, the actual coordinates of the lower map are recorded when the upper map is transferred to the lower map to prevent the lift platform from failing to reach its destination due to the inconsistency between the switching point of the lower lift platform and the actual switching point when the train stops offset.

[0140] Example 2

[0141] To achieve the above purpose, see Fig. 9 : This embodiment also provides a dynamic path planning system for a train inspection robot, comprising the following modules:

[0142] Input module: used to preset the attributes of all points and routes in tunnels, trains and maps;

[0143] Train path calculation module: used to identify the train number and calculate the optimal path for the current train stop direction;

[0144] Deviation correction module: used to detect train deviation, obtain deviation value, and adjust each preset point value under the current train stop position;

[0145] Robot path planning module: used to dynamically establish robot driving path planning according to different construction methods inside and outside the tunnel;

[0146] Navigation module: It is used for the robot to follow the dynamically planned path, fine-tune the driving trajectory related to the inside and outside of the tunnel according to the actual situation, reach the target detection point, complete the detection task, and return to the charging point.

[0147] The dynamic path planning system of the train inspection robot of this embodiment has the same advantages as the dynamic path planning method of the train inspection robot described above over the prior art, which will not be described in detail here.

[0148] Example 3

[0149] In order to achieve the above-mentioned purpose, the present embodiment further provides a train inspection robot, which adopts the dynamic path planning method of the train inspection robot as described above.

[0150] 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, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A dynamic path planning method for a train inspection robot, characterized in that: The method comprises the following steps: S1. Preset the tunnel, train and all points and route attributes on the map; S2, identify the train number and calculate the optimal path for the current train stop direction; S3, detect train deviation, obtain deviation value, and adjust each preset point value under the current train stop position; S4. Dynamically establish the robot's driving path planning according to different construction methods inside and outside the tunnel; S5. The robot follows the dynamically planned path, fine-tunes the driving trajectory associated with the inside and outside of the tunnel according to the actual situation, reaches the target detection point, completes the detection task, and returns to the charging point.

2. The dynamic path planning method of the train inspection robot according to claim 1, characterized in that: In the step S1, the following specific processes are included: S11: generating a map outside the tunnel and multiple maps of the bottom of the train inside the tunnel, wherein the generated map includes a map number and an area number, wherein the area number is an upper area or a lower area, representing a map outside the tunnel or a map of the bottom of the train inside the tunnel, each map coordinate system is independent, and the map number of the same area number is unique; S12, presetting the coordinates and attributes of each path point on the map and the attributes of the path line; S13. Currently, there are two ways to transfer from outside the tunnel to inside the tunnel. The first way is through a ramp, and the second way is through a lifting platform.

3. The dynamic path planning method of the train inspection robot according to claim 1, characterized in that: In the step S2, the following specific processes are included: S21: The train inspection robot receives the train number identification task issued by the backend control center server, and calculates the point closest to the robot's current coordinates and the preset path according to the preset path point information as the starting point; Use the vector method to find the distance between the current coordinate point and each preset inspection route, obtain the route with the shortest distance that is less than a certain threshold, and then calculate the distance between the current coordinate point and the two points on this route respectively. The point with the shortest distance is the starting point. The preset line where the current coordinate point is on the extension line of the preset inspection route is unavailable; The current coordinate point is not in the same area as the preset inspection route and is unavailable; The current coordinate point and the preset inspection route are in the same area with different numbers and cannot be used; the current preset route is not enabled and cannot be used; S22: Use Dijkstra algorithm to calculate the shortest path from the starting point to the target point. In this shortest path set, delete the first path point to solve the problem of robot turning and causing collision due to the narrow road and the path point behind the robot. S23: Pass each path point in turn, arrive at the target point, send control instructions to the robotic arm, and control the robotic arm to reach the preset position; control the 3D high-definition camera to take photos, and send them back to the background control center server. After algorithm recognition, the preset train attributes are used as the standard to obtain the train number that currently needs to be inspected, and store it in the database of the background control center server.

4. The dynamic path planning method of the train inspection robot according to claim 1, characterized in that: In the step S3, the following specific processes are included: S31: The robot receives the train deviation detection task sent by the background control center server, arrives at the target point, sends control instructions to the robotic arm, controls the robotic arm to reach the preset position; controls the 3D high-definition camera to take several photos and sends them back to the background control center server. S32: After algorithm recognition, the deviation value between the current train stop position and the preset train stop position is obtained and sent to the robot. S33: The robot receives the deviation value and dynamically adjusts the coordinate position of each preset point on the map outside the tunnel according to the deviation value.

5. The dynamic path planning method of the train inspection robot according to claim 1, characterized in that: In the step S4, the following specific processes are included: S41: Loading configuration file parameters preset in the robot system; S42: The robot receives a task list sent to it by the backend control center server, including the train car number and the tunnel number, and saves the task list in the database in sequence; S43: According to preset parameters, different construction methods outside the tunnel and inside the tunnel are switched, and different dynamic path planning is implemented.

6. The dynamic path planning method of the train inspection robot according to claim 5, characterized in that: Set the route and walking direction connecting the outside and inside of the tunnel; if it is a one-way tunnel, you need to configure one virtual line, if it is a two-way tunnel, you need to configure two virtual lines; Among them, a one-way tunnel means that there is a downhill slope at one end of the train, and a two-way tunnel means that there are downhill slopes at both ends of the train; When parking in the forward direction, the upper lift point and the lower tunnel head lift point are used as map switching points, a virtual line is configured between the two points, and the line speed is set to 0; the upper lift point is used as the starting point, and the lower tunnel head lift point is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking; When parking in reverse, the upper lift point and the lower tunnel end lift point are used as map switching points, a virtual line is configured between these two points, and the line speed is set to 0; the upper lift point is used as the starting point, and the lower tunnel end lift point is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking.

7. The dynamic path planning method of the train inspection robot according to claim 6, characterized in that: Also includes: When it is a one-way tunnel and forward parking, the inner point of the tunnel head and the front path point are used as map switching points, a virtual line is configured between these two points, and the line speed is set to 0; the inner point of the tunnel head is used as the starting point, and the front path point is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking; When it is a one-way tunnel or reverse parking, the driving direction of the uphill and downhill routes needs to be changed. The reverse direction is changed to forward direction from the inner point of the tunnel head to the outer point of the tunnel head, and the reverse direction is changed to forward direction from the outer point of the tunnel head to the inner point of the tunnel head. The inner point of the tunnel head and the path point at the rear of the vehicle are used as map switching points, and a virtual line is configured between the two points, and the line speed is set to 0; the inner point of the tunnel head is used as the starting point, and the path point at the rear of the vehicle is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking; When it is a two-way tunnel and forward parking, it is necessary to add a pair of map switching points on the basis of the one-way tunnel, namely the internal point at the end of the tunnel and the path point at the rear of the vehicle, configure a virtual line between these two points, and set the line speed to 0; the internal point at the end of the tunnel is used as the starting point, and the path point at the rear of the vehicle is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking; When it is a two-way tunnel and reverse parking, on the basis of the one-way tunnel, the driving direction of the uphill and downhill road route at the other end needs to be changed, and the direction from the external point at the end of the tunnel to the internal point at the end of the tunnel is changed from reverse walking to forward walking, and the direction from the internal point at the end of the tunnel to the external point at the end of the tunnel is changed from forward walking to reverse walking; on the basis of the one-way tunnel, it is necessary to add a pair of map switching points, namely the internal point at the end of the tunnel and the path point at the front of the vehicle, configure a virtual line between these two points, and set the line speed to 0; the internal point at the end of the tunnel is used as the starting point, and the path point at the front of the vehicle is used as the ending point. The route from the starting point to the ending point is set to forward walking, and the route from the ending point to the starting point is set to reverse walking; When the lift platform method is used, the train car number and tunnel number are obtained from the task, and the information in the preset parameter file is used as the standard to determine the stop direction of the train and obtain the lift platform number bound to the tunnel number.

8. The dynamic path planning method of the train inspection robot according to claim 1, characterized in that: In the step S5, the following specific processes are included: S51: Read a task from the database, the robot drives along the dynamically planned path, and calculates the arrival angle according to the actual position parameters returned by the navigation system; Determine whether the next point is the target point. If it is the target point, the arrival angle is based on the preset value. If the target point is not a path point, then the angle relative to the map is calculated based on the current robot coordinates returned by the navigation and the preset coordinates of the next path point. If the direction of the current path line is reverse, then determine whether the in-position angle is greater than 0. If it is greater than 0, subtract 180° from the in-position angle, otherwise, add 180° to the in-position angle. If the current path line and the next path line are in opposite directions, then determine whether the in-position angle is greater than 0. If it is greater than 0, subtract 180° from the in-position angle, otherwise, add 180° to the in-position angle. S52: each time a path point is reached, its attributes are obtained, and relevant services are implemented according to the path point attributes; Among them, when the ramp method is adopted, if the current attribute is a map-cutting point, a map-cutting instruction is sent to the navigation program, and after waiting for the navigation map to switch, the path point information returned by the navigation is obtained, and the sending of the next path point is adjusted according to the situation; if the current path point and the next path point are both map-cutting points, and they are driving from the lower area to the upper area, the next path point is deleted, and the next path point information is directly sent to the navigation program; thereby solving the problem of inaccurate navigation position after cutting the map from inside the tunnel to outside the tunnel, and the robot turning in a narrow tunnel; Among them, when the lifting platform method is adopted, the attribute of the next point of the current position point is obtained. If it is a lifting platform point, it is determined whether the current parking area of ​​the lifting platform is consistent with the current area of ​​the robot. If the areas are inconsistent, the lifting platform is controlled to the area where the robot is located; after the robot arrives at the lifting platform point, the arrival results of the navigation system and the lifting platform control system are obtained. If one of the returns is incorrect, an abnormal warning is given. When both return to the position, the lifting platform is controlled to reach the lifting platform point of the specified area, and a map switching instruction is issued to the navigation program at the same time. Wait for the navigation map switching to be completed, obtain the arrival information of the path point returned by the navigation, and if it is in place, send the next path point; Among them, if the first lifting platform point is in the upper area and the second lifting platform point is in the lower area, it is necessary to record the coordinates of the lifting platform point in the lower area and save them in the robot database as the arrival coordinates of the lifting platform point in the lower layer when the robot moves from the lower area to the upper area; S53, when the robot leaves the previous target point and moves to the next target point, the robot arm is controlled to move to the preset position of the next target point to improve the train inspection efficiency; S54. When all tasks in the database are completed, the robot calculates the optimal path to return to the charging point according to the preset path, arrives at the charging point, and waits for the next task.

9. A dynamic path planning system for a train inspection robot, characterized in that: Includes the following modules: Input module: used to preset the attributes of all points and routes in tunnels, trains and maps; Train path calculation module: used to identify the train number and calculate the optimal path for the current train stop direction; Deviation correction module: used to detect train deviation, obtain deviation value, and adjust each preset point value under the current train stop position; Robot path planning module: used to dynamically establish robot driving path planning according to different construction methods inside and outside the tunnel; Navigation module: It is used for the robot to follow the dynamically planned path, fine-tune the driving trajectory related to the inside and outside of the tunnel according to the actual situation, reach the target detection point, complete the detection task, and return to the charging point.

10. A train inspection robot, characterized in that: A dynamic path planning method for a train inspection robot as described in any one of claims 1 to 8 is adopted.

Images