Park logistics robot and passing method

Abstract

A park logistics robot comprises a radar system, a visual system and a control system, and a passing method based on the park logistics robot comprises the following steps: collecting a plurality of obstacle images through the visual system in a delay interval shooting mode when the robot stops passing due to obstacles, and sending obstacle image data to the control system; the control system recognizes the type and state of an obstacle according to the received obstacle image data and calls or generates a passing scheme, and particularly brings logistics vehicles and / or walking bearing machines used in the normalized work of the park into a recognition matching database. The multi-state trafficability pre-judgment information of the specific obstacles is pre-stored through the control system and can be compared with obstacle image data collected in real time, the types of the obstacles are matched, the trafficability pre-judgment information possibly existing in the current state is called, and the trafficability pre-judgment information serves as comparison data of the mode that the robot selects and disposes a traffic event behavior. And the capacity of the logistics robot for automatically handling the passing event and scientifically selecting the passing path is improved.

Description

Technical Field

[0001] The present invention relates to the field of robot control application technology, and in particular to a park logistics robot and a passage method. Background Art

[0002] The existing road planning within the factory cannot be carried out through social maps such as Gaode, because the road conditions within the factory cannot be obtained. At present, if the intelligent production park wants the park logistics robot to perform intelligent road condition analysis, it is necessary to set up cameras on the roads in the factory. The robot builds a factory road map, and then analyzes the images of the cameras running on the road ahead, obtains congestion and other conditions, and recalculates the route. This will greatly increase the investment management and operating costs of basic equipment, as well as the research and development and analysis costs of the robot.

[0003] In fact, production parks are different from social roads. In production parks, pedestrians or passenger vehicles basically do not cause traffic congestion and obstacles to the passability of logistics robots. The obstacles to the passability of logistics robots are basically the factory logistics vehicles that need to unload or stop waiting and some walking load-bearing machinery required for the flow of workstations. These specific obstacles are also the main obstacles that occur frequently and repeatedly in park operations. Only by qualitatively identifying these normalized obstacles can the response plan of the logistics robot be prepared in advance, improving the ability of the logistics robot to automatically handle traffic events and scientifically select traffic paths. Summary of the invention

[0004] To solve the above problems, the present invention proposes a park logistics robot and a corresponding traffic method. In a park without road condition recognition system facilities, through simple visual information collection and program setting, qualitative identification of normalized obstacles in the park is introduced, and response plans of the logistics robot are prepared in advance, thereby improving the logistics robot's ability to automatically handle traffic events and scientifically select traffic paths.

[0005] The objectives of the present invention are achieved through the following technical solutions.

[0006] A park logistics robot, comprising:

[0007] A radar system, which is used for obstacle recognition and navigation. After detecting an obstacle on the robot's running path, the radar system identifies the obstacle size and sends the obstacle size data to the control system;

[0008] A visual system, wherein the visual system is used to collect multiple obstacle images by time-delay interval shooting when the robot stops passing due to an obstacle and send the obstacle image data to the control system;

[0009] A control system, the control system is used to determine the passability of the robot based on the received obstacle size data and identify or generate a passage path, identify the obstacle type and state based on the received obstacle image data and retrieve or generate a passage plan;

[0010] The obstacles include specific obstacles S and newly added obstacles N. The specific obstacles S are logistics vehicles and / or walking load-bearing machinery used for routine work in the park. The data of the specific obstacles S are pre-stored in the control system. The data at least includes the type definition of the specific obstacle S and the multi-state passability prediction information of the specific obstacle S under the type definition. The multi-state passability prediction information at least includes the obstacle removal remaining time prediction information T1 of the specific obstacle S under different cargo loading states. The newly added obstacles N are obstacles collected for the first time by the robot's visual system in addition to the above-mentioned specific obstacles S.

[0011] The control system compares the obstacle image data collected by the visual system when the robot stops passing due to an obstacle with the pre-stored data of the specific obstacle S to determine whether it is a pre-defined specific obstacle S type. In addition, the control system also compares multiple obstacle images collected by time-lapse interval shooting to determine the current state of the obstacle. When the collected obstacle image data matches the pre-defined specific obstacle S type, the control system further compares the current state of the obstacle with the pre-stored multi-state data of the specific obstacle S, and directly reads the passability prediction information of the state according to the matched state, and sends an action instruction to the robot; when the collected obstacle image data does not match the pre-defined specific obstacle S type, the control system sends an action instruction to the robot based on the current state of the obstacle.

[0012] For the above-mentioned park logistics robot, the type definition of the specific obstacle S includes one or more of the model, capacity, license plate number, and cargo type of the logistics vehicle and / or walking load-bearing machinery.

[0013] A method for passing through the above-mentioned park logistics robot includes the following steps:

[0014] Step A: The radar system identifies obstacles, and the control system determines whether the obstacle can be avoided by detouring in the current path according to the obstacle size data. If yes, the robot is directly controlled to avoid the obstacle by detouring. If not, it goes to step B.

[0015] Step B: The control system determines whether there is an alternative path based on the map data. If there is no alternative path, the control system controls the robot to select from the following responses: a. Wait for the obstacle to be removed and then continue to complete the current task; b. Return; c. Wait for a preset time; if there is an alternative path, proceed to step C;

[0016] Step C, the control system predicts the mileage L and / or the delivery time T2 of the identified alternative path. If the mileage L and / or the delivery time T2 of the alternative path is within the preset value range, the robot is directly controlled to avoid obstacles by changing the path. If not, the process proceeds to step D.

[0017] Step D, the control system compares the obstacle image data collected by the visual system when the robot stops passing due to the obstacle with the pre-stored data of the specific obstacle S to determine whether it is a pre-defined specific obstacle S type, and the control system also compares the multiple obstacle images collected by time-lapse interval shooting to determine the current state of the obstacle. When the collected obstacle image data matches the pre-defined specific obstacle S type, the control system further compares the current state of the obstacle with the pre-stored multi-state data of the specific obstacle S, and directly reads the passability prediction information of the state according to the matched state, compares the remaining time prediction information T1 for obstacle removal in the state with the transportation time T2 predicted according to the identified alternative path, if T1<T2, executes waiting, if T1≥T2, the control system controls the robot to select from the following responses: a, avoid the obstacle by changing the path; b, return; when the collected obstacle image data does not match the pre-defined specific obstacle S type, the control system sends an action instruction to the robot based on the current state of the obstacle.

[0018] Preferably, in the above-mentioned passage method, the method of sending action instructions to the robot based on the current state of the obstacle described in step D is: when it is judged that the current state of the obstacle is static, the control system controls the robot to select from the following responses: a, avoiding the obstacle by changing the path; b, returning; c, waiting for a preset time; when it is judged that the current state of the obstacle is dynamic, the control system controls the robot to wait for a preset time.

[0019] Preferably, in the above-mentioned passage method, during the waiting period of a preset time described in step B, the control system continues to judge the passage conditions, and if the passage conditions are still not met after the preset time, it returns.

[0020] Preferably, in the above-mentioned passage method, as described in step D, when it is judged that the current state of the obstacle is static, the control system continues to judge the passage conditions while waiting for a preset time. If the passage conditions are still not met at the end of the preset time or the state of the obstacle has not changed (has not become dynamic), the obstacle is avoided or the return is made by changing the path.

[0021] Preferably, in the above-mentioned passage method, as described in step D, when it is determined that the current state of the obstacle is dynamic, the control system controls the robot to wait for a preset time, and the control system continues to judge the passage conditions. If the passage conditions are still not met at the end of the preset time or the state of the obstacle changes (becomes static), the obstacle is avoided or the return is made by changing the path.

[0022] Preferably, in the above-mentioned passage method, T1 described in step D is less than T2, and the waiting is performed until the obstacle is removed or a preset time is waited. During the waiting period of the preset time, the control system continues to judge the passage conditions. If the passage conditions are still not met at the end of the preset time or the state of the obstacle changes (becomes static), the obstacle is avoided or the return is made by changing the path.

[0023] The beneficial effects of the present invention are:

[0024] The park logistics robot and the corresponding passage method provided by the present invention use a visual system to collect multiple obstacle images by time-delay interval shooting when the robot stops passing due to an obstacle, and send the obstacle image data to the control system. The control system identifies the obstacle type and state according to the received obstacle image data and retrieves or generates a passage plan, especially incorporating the logistics vehicles and / or walking load-bearing machinery used for routine work in the park into the identification and matching database, and pre-stores the multi-state passability prediction information of these specific obstacles through the control system, which can be compared with the obstacle image data collected in real time, match the obstacle type, retrieve the passability prediction information that may exist in the current state, and use it as comparison data for the robot to select the behavior mode for handling passage events, thereby improving the ability of the logistics robot to automatically handle passage events and scientifically select passage paths. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a schematic diagram of different loading states of a logistics vehicle (specific obstacle) as exemplified in Example 1.

[0026] Figure 2 This is a flowchart of the steps of Example 2. DETAILED DESCRIPTION

[0027] Example 1

[0028] A campus logistics robot includes at least a radar system, a visual system, and a control system required to achieve the purpose of the present invention. The radar system is used for obstacle identification and navigation. After detecting an obstacle on the robot's running path, it identifies the obstacle size and sends the obstacle size data to the control system. The visual system is used to capture multiple obstacle images by time-lapse interval shooting when the robot stops passing due to an obstacle and send the obstacle image data to the control system. The control system is used to determine the robot's passability based on the received obstacle size data and identify or generate a passage path, identify the obstacle type and status based on the received obstacle image data, and retrieve or generate a passage plan;

[0029] The obstacles include specific obstacles S and newly added obstacles N. The specific obstacles S are logistics vehicles and / or walking load-bearing machinery used in the normal work of the park, such as Figure 1 A logistics vehicle is shown, in which data of a specific obstacle S is pre-stored in the control system, and the data at least includes a type definition of the specific obstacle S (which may be a custom parameter, or may include model, capacity, license plate number, cargo type, etc. as identification parameters) and multi-state passability prediction information of the specific obstacle S under the type definition.

[0030] The multi-state passability prediction information includes at least the remaining time prediction information T1 of the obstacle removal of a specific obstacle S under different cargo loading states, for example, Figure 1 The figure shows a schematic diagram of different loading states of a logistics vehicle (a "specific obstacle" in the park defined in the present invention). The specific logistics vehicles that appear regularly in the park usually have repetitive cargo types and loading and unloading cycles. For example, the park is a manufacturing factory, and a parts manufacturing department in the factory transports parts to an assembly department, or a storage department in the factory transports raw materials to a parts manufacturing department or an assembly department. The robot can manually define or repeatedly machine learn the following. Figure 1 The remaining time for removing obstacles under different loading conditions shown in the figure is predicted, so that the prediction information T1 of the remaining time for removing obstacles under different loading conditions can be used as comparison data for the robot to select the behavior mode for handling the passage event, for example, Figure 1 (a)-(d) in the figure represent four different states in the dynamic process of cargo loading and unloading, representing different remaining travel times. It should be noted that although the attached figure shows different loading states of logistics vehicles (specific obstacles) in the form of a side view, according to the road conditions in the park, the system can pre-store multiple views of different directions or angles, which can further improve the accuracy of subsequent matching.

[0031] The newly added obstacle N described in the present invention is an obstacle that is first collected by the robot's visual system in addition to the above-mentioned specific obstacle S. For a park with specific production or life activities, although the frequency of occurrence of such temporary obstacles is not high, it is also a frequent event. The definition of the newly added obstacle N can change according to the number of times it is recognized by the robot. For example, when the newly added obstacle N is recognized multiple times, the newly added obstacle N can be regarded as a new specific obstacle based on the recognition frequency or directly based on external information, and then its definition and pre-stored data can be changed through machine learning or manual modification of the database.

[0032] The control system of the present invention compares the obstacle image data collected by the visual system when the robot stops passing due to an obstacle with the pre-stored data of the specific obstacle S to determine whether it is a pre-defined specific obstacle S type. In addition, the control system also compares multiple obstacle images collected by time-lapse interval shooting to determine the current state of the obstacle. When the collected obstacle image data matches the pre-defined specific obstacle S type, the control system further compares the current state of the obstacle with the pre-stored multi-state data of the specific obstacle S, and directly reads the passability prediction information of the state according to the matched state, and sends an action instruction to the robot; when the collected obstacle image data does not match the pre-defined specific obstacle S type, the control system sends an action instruction to the robot based on the current state of the obstacle.

[0033] In the following Example 2, the detailed implementation principle of the present invention will be further introduced in combination with the common method.

[0034] Example 2

[0035] Combination Figure 2 This embodiment provides a method for the park logistics robot to pass based on the embodiment 1, comprising the following steps:

[0036] In step A, the radar system identifies obstacles, and the control system determines whether the obstacle can be avoided by detouring in the current path based on the obstacle size data. If so, the robot is directly controlled to avoid the obstacle by detouring. If not, it goes to step B.

[0037] In step B, the control system determines whether there is an alternative path based on the map data. When there is no alternative path, the control system controls the robot to choose from the following responses: a, wait for the obstacle to be removed and then continue to complete the current task; b, return; c, wait for a preset time; when there is an alternative path, go to step C.

[0038] In step C, the control system predicts the mileage L and / or delivery time T2 of the identified alternative path. If the mileage L and / or delivery time T2 of the alternative path are within the preset value range, the robot is directly controlled to avoid obstacles by changing the path. If not, it goes to step D.

[0039] Step D, the control system compares the obstacle image data collected by the visual system when the robot stops passing due to the obstacle with the pre-stored data of the specific obstacle S to determine whether it is a pre-defined specific obstacle S type, and the control system also compares the multiple obstacle images collected by time-lapse interval shooting to determine the current state of the obstacle. When the collected obstacle image data matches the pre-defined specific obstacle S type, the control system further compares the current state of the obstacle with the pre-stored multi-state data of the specific obstacle S, and directly reads the passability prediction information of the state according to the matched state, compares the remaining time prediction information T1 for obstacle removal in the state with the transportation time T2 predicted according to the identified alternative path, if T1<T2, executes waiting, if T1≥T2, the control system controls the robot to select from the following responses: a, avoid the obstacle by changing the path; b, return; when the collected obstacle image data does not match the pre-defined specific obstacle S type, the control system sends an action instruction to the robot based on the current state of the obstacle.

[0040] Preferably, in the above-mentioned passage method, the method of sending action instructions to the robot based on the current state of the obstacle described in step D is: when it is judged that the current state of the obstacle is static, the control system controls the robot to select from the following responses: a, avoiding the obstacle by changing the path; b, returning; c, waiting for a preset time; when it is judged that the current state of the obstacle is dynamic, the control system controls the robot to wait for a preset time.

[0041] Preferably, in the above-mentioned passage method, during the waiting period of a preset time described in step B, the control system continues to judge the passage conditions, and if the passage conditions are still not met after the preset time, it returns.

[0042] Preferably, in the above-mentioned passage method, as described in step D, when it is judged that the current state of the obstacle is static, the control system continues to judge the passage conditions while waiting for a preset time. If the passage conditions are still not met at the end of the preset time or the state of the obstacle has not changed (has not become dynamic), the obstacle is avoided or the return is made by changing the path.

[0043] Preferably, in the above-mentioned passage method, as described in step D, when it is determined that the current state of the obstacle is dynamic, the control system controls the robot to wait for a preset time, and the control system continues to judge the passage conditions. If the passage conditions are still not met at the end of the preset time or the state of the obstacle changes (becomes static), the obstacle is avoided or the return is made by changing the path.

[0044] Preferably, in the above-mentioned passage method, T1 described in step D is less than T2, and the waiting is performed until the obstacle is removed or a preset time is waited. During the waiting period of the preset time, the control system continues to judge the passage conditions. If the passage conditions are still not met at the end of the preset time or the state of the obstacle changes (becomes static), the obstacle is avoided or the return is made by changing the path.

[0045] It can be seen from the above embodiments that the present invention uses a visual system to collect multiple obstacle images by time-delay interval shooting when the robot stops passing due to an obstacle, and sends the obstacle image data to the control system. The control system identifies the obstacle type and state according to the received obstacle image data and retrieves or generates a passage plan, especially incorporating the logistics vehicles and / or walking load-bearing machinery used for routine work in the park into the identification and matching database, and pre-stores the multi-state passability prediction information of these specific obstacles through the control system, which can be compared with the obstacle image data collected in real time, match the obstacle type, retrieve the possible passability prediction information in the current state, and use it as comparison data for the robot to select the behavior mode of handling passage events, thereby improving the ability of the logistics robot to automatically handle passage events and scientifically select passage paths.

[0046] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A park logistics robot, characterized in that: include: A radar system, which is used for obstacle recognition and navigation. After detecting an obstacle on the robot's running path, the radar system identifies the obstacle size and sends the obstacle size data to the control system; A visual system, wherein the visual system is used to collect multiple obstacle images by time-delay interval shooting when the robot stops passing due to an obstacle and send the obstacle image data to the control system; A control system, the control system is used to determine the passability of the robot based on the received obstacle size data and identify or generate a passage path, identify the obstacle type and state based on the received obstacle image data and retrieve or generate a passage plan; The obstacles include specific obstacles S and newly added obstacles N. The specific obstacles S are logistics vehicles and / or walking load-bearing machinery used for routine work in the park. The data of the specific obstacles S are pre-stored in the control system. The data at least includes the type definition of the specific obstacle S and the multi-state passability prediction information of the specific obstacle S under the type definition. The multi-state passability prediction information at least includes the obstacle removal remaining time prediction information T1 of the specific obstacle S under different cargo loading states. The newly added obstacles N are obstacles collected for the first time by the robot's visual system in addition to the above-mentioned specific obstacles S. The control system compares the obstacle image data collected by the visual system when the robot stops passing due to an obstacle with the pre-stored data of the specific obstacle S to determine whether it is a pre-defined specific obstacle S type. In addition, the control system also compares multiple obstacle images collected by time-lapse interval shooting to determine the current state of the obstacle. When the collected obstacle image data matches the pre-defined specific obstacle S type, the control system further compares the current state of the obstacle with the pre-stored multi-state data of the specific obstacle S, and directly reads the passability prediction information of the state according to the matched state, and sends an action instruction to the robot; when the collected obstacle image data does not match the pre-defined specific obstacle S type, the control system sends an action instruction to the robot based on the current state of the obstacle.

2. A park logistics robot according to claim 1, characterized in that: The type definition of the specific obstacle S includes one or more of the model, capacity, license plate number, and cargo type of the logistics vehicle and / or the mobile load-bearing machinery.

3. A method for passing through a park logistics robot based on claim 1 or 2, characterized in that: The following steps are involved: Step A: The radar system identifies obstacles, and the control system determines whether the obstacle can be avoided by detouring in the current path according to the obstacle size data. If yes, the robot is directly controlled to avoid the obstacle by detouring. If not, it goes to step B. Step B: The control system determines whether there is an alternative path based on the map data. If there is no alternative path, the control system controls the robot to select from the following responses: a. Wait for the obstacle to be removed and then continue to complete the current task; b. Return; c. Wait for a preset time; if there is an alternative path, proceed to step C; Step C, the control system predicts the mileage L and / or the delivery time T2 of the identified alternative path. If the mileage L and / or the delivery time T2 of the alternative path is within the preset value range, the robot is directly controlled to avoid obstacles by changing the path. If not, the process proceeds to step D. Step D, the control system compares the obstacle image data collected by the visual system when the robot stops passing due to the obstacle with the pre-stored data of the specific obstacle S to determine whether it is a pre-defined specific obstacle S type, and the control system also compares the multiple obstacle images collected by time-lapse interval shooting to determine the current state of the obstacle. When the collected obstacle image data matches the pre-defined specific obstacle S type, the control system further compares the current state of the obstacle with the pre-stored multi-state data of the specific obstacle S, and directly reads the passability prediction information of the state according to the matched state, compares the remaining time prediction information T1 for obstacle removal in the state with the transportation time T2 predicted according to the identified alternative path, if T1<T2, executes waiting, if T1≥T2, the control system controls the robot to select from the following responses: a, avoid the obstacle by changing the path; b, return; when the collected obstacle image data does not match the pre-defined specific obstacle S type, the control system sends an action instruction to the robot based on the current state of the obstacle.

4. The passage method according to claim 3, characterized in that: The method of sending action instructions to the robot based on the current state of the obstacle described in step D is: when the current state of the obstacle is judged to be static, the control system controls the robot to select from the following responses: a, avoiding the obstacle by changing the path; b, returning; c, waiting for a preset time; when the current state of the obstacle is judged to be dynamic, the control system controls the robot to wait for a preset time.

5. The passage method according to claim 3 or 4, characterized in that: During the waiting period of a preset time described in step B, the control system continues to judge the traffic conditions. If the traffic conditions are still not met after the preset time, the system returns.

6. The passage method according to claim 4, characterized in that: When it is determined in step D that the current state of the obstacle is static, the control system continues to determine the passage conditions during a preset period of time. If the passage conditions are still not met at the end of the preset time or the state of the obstacle has not changed (not become dynamic), the obstacle is avoided or the return is made by changing the path.

7. The passage method according to claim 4, characterized in that: As described in step D, when it is determined that the current state of the obstacle is dynamic, the control system controls the robot to wait for a preset time, and the control system continues to judge the passage conditions. If the passage conditions are still not met at the end of the preset time or the state of the obstacle changes (becomes static), the obstacle is avoided or returned by changing the path.

8. The passage method according to claim 3, characterized in that: T1<T2 described in step D, the waiting execution is to wait until the obstacle is removed or wait for a preset time. During the waiting period of the preset time, the control system continues to judge the traffic conditions. If the traffic conditions are still not met at the end of the preset time or the state of the obstacle changes (becomes static), the obstacle is avoided or the return is made by changing the path.

Images