Path obstacle avoidance method based on fuzzy matching camera video stream and probabilistic cloud computing

Abstract

The present invention discloses a path obstacle avoidance method based on fuzzy matching camera video stream and probabilistic cloud computing, and specifically relates to the field of path obstacle avoidance technology. The method comprises: step one, initial path planning, step two, initial path obstacle probability prediction, step three, analyzing the video stream to identify obstacles, and step four, implementing path obstacle avoidance response measures. The present invention divides the risk area on the initialization path according to the obstacle information on the historical initialization path. When the transport robot reaches the risk area, it adopts more sophisticated image recognition to determine whether there are obstacles and their specific types, so as to facilitate timely path obstacle avoidance. The fuzzy recognition method can significantly improve the accuracy of obstacle recognition, tracking and classification in complex scenes, meet the intelligence and efficiency of factory transport robots, and improve factory production efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of path obstacle avoidance, and in particular to a path obstacle avoidance method based on fuzzy matching camera video stream and probabilistic cloud computing. Background Art

[0002] The development of factory transport robots has become an important trend in the manufacturing and logistics industries, especially in promoting the construction of smart factories, improving production efficiency and reducing labor costs. With the continuous advancement of technology, these robots are developing in a more flexible, efficient and intelligent direction, and can adapt to complex working environments and requirements. With the rise of smart manufacturing and smart logistics, factory transport robots are developing in a more multifunctional and flexible direction, so it is necessary to study the path obstacle avoidance method of factory transport robots.

[0003] The existing technology for factory transport robot path obstacle avoidance can meet the basic requirements, but there are also some potential problems, which are specifically manifested in the following aspects: on the one hand, factory transport robots generally use magnetic strips laid on the ground, and magnetic sensors are installed on the bottom of the robot to track and maintain the driving path by detecting magnetic strips, or use color ribbons or color lines on the ground, and the robot uses photoelectric sensors to identify the path and guide its movement. When encountering obstacles, it is impossible to identify the state of the obstacles in time and re-plan the path, resulting in the factory transport robot being unable to continue working, reducing the factory transportation efficiency, and not meeting the actual application needs.

[0004] On the other hand, part of the existing technology uses laser reflective plates set up in the factory environment for the working path of the transport robot. The robot uses a built-in laser scanner to identify the position of these reflectors to correct and maintain its own position. This method provides higher path flexibility and accuracy, but usually requires more complex initial installation and calibration, and has low efficiency in obstacle recognition. It is impossible to clearly obtain the specific information of the obstacles, determine whether it can continue to pass, and make timely adjustments to the path. This does not meet the construction needs of smart factories, resulting in the transport robot being unable to work efficiently. Summary of the invention

[0005] The purpose of the present invention is to provide a path obstacle avoidance method based on fuzzy matching camera video stream and probabilistic cloud computing, which solves the problems existing in the background technology.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solution: a path obstacle avoidance method based on fuzzy matching camera video stream and probabilistic cloud computing, the method comprising:

[0007] Step 1: Initial path planning: Obtain the working data of the factory transport robot and the information data of each candidate path, calculate the matching index of each candidate path, and plan the initial path of the factory transport robot;

[0008] Step 2: Predict the probability of obstacles on the initial path: Obtain the locations of each obstacle that has appeared in history on the initial path, analyze the probability of each obstacle appearing, and divide the obstacle risk area on the initial path;

[0009] Step 3: Analyze the video stream to identify obstacles: Obtain the video stream of the factory transport robot during its operation, identify obstacles on the path through fuzzy matching, and classify the obstacles on the path to obtain static obstacles and dynamic obstacles.

[0010] Step 4. Implement path obstacle avoidance response measures: obtain various information data of static type obstacles on the path, calculate the obstacle avoidance judgment value of the factory transport robot based on the various information data of the factory transport robot, implement static type obstacle avoidance response measures, obtain the probability values ​​of each future path of dynamic type obstacles on the path, and implement dynamic type obstacle avoidance response measures.

[0011] Furthermore, the matching index of each candidate path is calculated, and the specific analysis method is as follows:

[0012] The working data of the factory transport robot, including the width of the factory transport robot , Transportation time required ;

[0013] The information data of each path to be selected includes the maximum width of each path to be selected , minimum width , the number of obstacles in each transportation process in history , Historical transportation duration ,in , Indicates the number of each candidate path, represents the total number of paths to be selected, , Indicates the number of each historical transportation process, Indicates the total number of historical transportation processes;

[0014] The number of obstacles in each transportation process of each candidate path Perform mean processing, and the mean result is used as the historical average number of obstacles on each candidate path. ;

[0015] By formula: , calculate the width of each candidate path to meet the state value , through the formula: , calculate the matching index of each candidate path ,in Indicates the acceptable number of pre-set path obstacles. represents the exponential function identifier, Represents the logarithmic function identifier.

[0016] Furthermore, the specific analysis method for planning the initial path of the factory transport robot is as follows:

[0017] Based on the matching index of each candidate path Sort them and select the candidate path corresponding to the subscript of the maximum matching index, which is recorded as the initial path of the factory transport robot;

[0018] The command information of the initial path of the factory transport robot is obtained from the database and sent to the control center of the factory transport robot, and the factory transport robot performs the transportation work.

[0019] Furthermore, the specific analysis method for dividing the obstacle risk area on the initial path is as follows:

[0020] The starting position of the factory transport robot is set as the origin coordinate, the vertical direction is the longitudinal axis, and the horizontal right direction is the horizontal axis, and a two-dimensional coordinate system of the initial path of the factory transport robot is established;

[0021] The positions of each obstacle in the history of the initial path are extracted from the database, and mapped to the two-dimensional coordinate system after position conversion to obtain the position coordinates of each obstacle in the history of the initial path. ,in , Indicates the number of obstacles that appeared in the history of the initial path. Indicates the total number of historical obstacles that occurred on the initial path;

[0022] The location coordinates of each obstacle occurrence in the history of the initial path , statistics are located in Area where the secondary obstacle appears Total number of obstacles within Similarly, the total number of obstacles in the area where each obstacle appears is obtained , through the formula: , calculate the first The probability of obstacles in the coordinate area of ​​the location where the obstacle appears ,in, A mathematical symbol for a plane representing the area where obstacles appear. Indicates the preset judgment range scale;

[0023] like , then the first The area where the secondary obstacle appears is the risk area, where Indicates the obstacle probability judgment threshold;

[0024] Otherwise, the initial path is divided into The area where the secondary obstacle appears is the general area.

[0025] Furthermore, the obstacles on the path are identified and classified, and the specific analysis method is as follows:

[0026] Get the video stream of the factory transport robot and perform frame processing. For the video stream of the risk area, set the frame processing extraction interval to , for video streams in the general area, set the frame processing extraction interval to , obtain each frame image of the video stream of the factory transport robot, perform target detection on each frame image of the video stream of the factory transport robot through a deep learning model, and segment obstacles in the scene;

[0027] Analyze the obstacle pixels in each frame of the video stream of the factory transport robot, and use the formula: , calculate the difference between consecutive frames ,in The current frame of the video stream representing the factory transport robot Pixel coordinates, The previous frame of the video stream of the factory transport robot Pixel coordinates, , represents the number of each pixel in the frame of the video stream of the factory transport robot, The total number of pixels in the frame of the video stream representing the factory transport robot;

[0028] like , then determine the The pixels are dynamic areas, and the obstacle is a dynamic type obstacle, where: represents the difference threshold of consecutive frames obtained from the database;

[0029] Otherwise, judge The pixel points are static areas, and the obstacle is a static type obstacle;

[0030] In summary, the obstacles and their types on the initial path of the factory transport robot are obtained.

[0031] Furthermore, the specific analysis method of each information data of the static type obstacle on the path is as follows:

[0032] Detect objects based on each frame of the video stream of the factory transport robot to obtain the width of the occupied path of static obstacles , the width of the path where the static type obstacle is located .

[0033] Furthermore, the obstacle avoidance judgment value of the factory transport robot is calculated and the static obstacle avoidance response measures are implemented. The specific analysis method is as follows:

[0034] The width of the occupied path based on static type obstacles , the width of the factory transport robot , the width of the path where the static type obstacle is located , through the formula: , calculate the initial path passability judgment value ;

[0035] like , then it is judged that the initial path is allowed to pass, and the transport robot moves and deviates to avoid static obstacles;

[0036] like , then it is determined that the initial path is not allowed to pass, and a temporary obstacle avoidance path is executed.

[0037] Furthermore, the specific analysis method of the temporary obstacle avoidance path is as follows:

[0038] The coordinates of the target position are extracted from the database, and a temporary obstacle avoidance path is generated based on the position coordinates of the point that the transport robot cannot pass through. The instruction information of the temporary obstacle avoidance path is sent to the control center of the transport robot, and the transport robot executes the temporary obstacle avoidance path.

[0039] Furthermore, the probability of each future path of the dynamic type obstacle on the path is specifically analyzed by:

[0040] Through the fuzzy matching algorithm, the future paths of dynamic obstacles are generated, and through the probability cloud model: , calculate the probability value of each future path ,in represents a point coordinate vector on each future path, represents the coordinate vector of each future position, represents discreteness, and π represents pi.

[0041] Furthermore, the specific analysis method for implementing the dynamic obstacle avoidance response measure is as follows:

[0042] Based on the probability values ​​of various future paths, the factory transport robot chooses a path that avoids obstacles.

[0043] Compared with the prior art, the present invention has the following benefits: First, the present invention divides the risk area on the initialization path according to the obstacle information on the historical initialization path. When the transport robot reaches the risk area, it adopts more sophisticated image recognition to determine whether there are obstacles and their specific types, so as to facilitate timely path obstacle avoidance, meet the intelligence and efficiency of factory transport robots, and improve factory production efficiency.

[0044] Second, the present invention adopts fuzzy recognition and the imprecision of fuzzy logic to deal with uncertainties and ambiguities in practical applications, which can significantly improve the accuracy of obstacle recognition, tracking and classification in complex scenes.

[0045] Third, the present invention uses probabilistic methods in path planning and obstacle avoidance problems, and can assign a probability value to different paths. This value represents the feasibility or risk of the path. Through cloud computing, a large number of possible paths can be calculated and evaluated, and the path with the lowest risk and the best cost can be selected, which can promptly respond to the problem of re-planning the obstacle path. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 The figure is a schematic diagram of the method flow chart of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] For ease of understanding, some specific terms of the present invention are explained below:

[0050] Image recognition: refers to the convolutional neural network (CNN) in deep learning. CNN gradually extracts high-level features of the image through multi-layer convolution and pooling operations to achieve recognition and classification of complex images.

[0051] Video stream fuzzy matching: refers to a technique in the field of video processing and computer vision that is used to identify and process objects or patterns in video streams, even if these objects or patterns do not match exactly or have a certain degree of noise, blur or changes.

[0052] Probabilistic cloud computing refers to video stream fuzzy matching, a technology in the field of video processing and computer vision that is used to identify and process objects or patterns in video streams, even if these objects or patterns do not match exactly or have a certain degree of noise, blur or change.

[0053] Reference Figure 1 As shown, a path obstacle avoidance method based on fuzzy matching camera video stream and probabilistic cloud computing includes:

[0054] Step 1: Initial path planning: Obtain the working data of the factory transport robot and the information data of each candidate path, calculate the matching index of each candidate path, and plan the initial path of the factory transport robot;

[0055] Step 2: Predict the probability of obstacles on the initial path: Obtain the locations of each obstacle that has appeared in history on the initial path, analyze the probability of each obstacle appearing, and divide the obstacle risk area on the initial path;

[0056] Step 3: Analyze the video stream to identify obstacles: Obtain the video stream of the factory transport robot during its operation, identify obstacles on the path through fuzzy matching, and classify the obstacles on the path to obtain static obstacles and dynamic obstacles.

[0057] Step 4. Implement path obstacle avoidance response measures: obtain various information data of static type obstacles on the path, calculate the obstacle avoidance judgment value of the factory transport robot based on the various information data of the factory transport robot, implement static type obstacle avoidance response measures, obtain the probability values ​​of each future path of dynamic type obstacles on the path, and implement dynamic type obstacle avoidance response measures.

[0058] In a specific embodiment of the present invention, the matching index of each candidate path is calculated, and the specific analysis method is as follows:

[0059] The working data of the factory transport robot, including the width of the factory transport robot , Transportation time required ;

[0060] The information data of each path to be selected includes the maximum width of each path to be selected , minimum width , the number of obstacles in each transportation process in history , Historical transportation duration ,in , Indicates the number of each candidate path, represents the total number of paths to be selected, , Indicates the number of each historical transportation process, Indicates the total number of historical transportation processes;

[0061] The number of obstacles in each transportation process of each candidate path Perform mean processing, and the mean result is used as the historical average number of obstacles on each candidate path. ;

[0062] By formula: , calculate the width of each candidate path to meet the state value , through the formula: , calculate the matching index of each candidate path ,in Indicates the acceptable number of pre-set path obstacles. represents the exponential function identifier, Represents the logarithmic function identifier.

[0063] In a specific embodiment of the present invention, the initial path of the factory transport robot is planned, and the specific analysis method is as follows:

[0064] Based on the matching index of each candidate path Sort them and select the candidate path corresponding to the subscript of the maximum matching index, which is recorded as the initial path of the factory transport robot;

[0065] The command information of the initial path of the factory transport robot is obtained from the database and sent to the control center of the factory transport robot, and the factory transport robot performs the transportation work.

[0066] In a specific embodiment of the present invention, the obstacle risk area on the initial path is divided, and the specific analysis method is as follows:

[0067] The starting position of the factory transport robot is set as the origin coordinate, the vertical direction is the longitudinal axis, and the horizontal right direction is the horizontal axis, and a two-dimensional coordinate system of the initial path of the factory transport robot is established;

[0068] The positions of each obstacle in the history of the initial path are extracted from the database, and mapped to the two-dimensional coordinate system after position conversion to obtain the position coordinates of each obstacle in the history of the initial path. ,in , Indicates the number of obstacles that appeared in the history of the initial path. Indicates the total number of historical obstacles that occurred on the initial path;

[0069] The location coordinates of each obstacle occurrence in the history of the initial path , statistics are located in Area where the secondary obstacle appears Total number of obstacles within Similarly, the total number of obstacles in the area where each obstacle appears is obtained , through the formula: , calculate the first The probability of obstacles in the coordinate area of ​​the location where the obstacle appears ,in, A mathematical symbol for a plane representing the area where obstacles appear. Indicates the preset judgment range scale;

[0070] like , then the first The area where the secondary obstacle appears is the risk area, where Indicates the obstacle probability judgment threshold;

[0071] Otherwise, the initial path is divided into The area where the secondary obstacle appears is the general area.

[0072] The present invention divides the risk area on the initialization path according to the obstacle information on the historical initialization path. When the transport robot reaches the risk area, it adopts more sophisticated image recognition to determine whether there are obstacles and the specific type, so as to facilitate timely path obstacle avoidance, meet the intelligence and efficiency of the factory transport robot, and improve the factory production efficiency.

[0073] In a specific embodiment of the present invention, the obstacles on the path are identified and classified, and the specific analysis method is as follows:

[0074] Get the video stream of the factory transport robot and perform frame processing. For the video stream of the risk area, set the frame processing extraction interval to , for video streams in the general area, set the frame processing extraction interval to , obtain each frame image of the video stream of the factory transport robot, perform target detection on each frame image of the video stream of the factory transport robot through the deep learning model, and segment the obstacles in the scene.

[0075] Analyze the obstacle pixels in each frame of the video stream of the factory transport robot, and use the formula: , calculate the difference between consecutive frames ,in The current frame of the video stream representing the factory transport robot Pixel coordinates, The previous frame of the video stream of the factory transport robot Pixel coordinates, , represents the number of each pixel in the frame of the video stream of the factory transport robot, The total number of pixels in the frame of the video stream representing the factory transport robot;

[0076] like , then determine the The pixels are dynamic areas, and the obstacle is a dynamic obstacle. represents the difference threshold of consecutive frames obtained from the database;

[0077] Otherwise, judge The pixel points are static areas, and the obstacle is a static type obstacle;

[0078] In summary, the obstacles and their types on the initial path of the factory transport robot are obtained.

[0079] The present invention adopts a fuzzy recognition method and the imprecision of fuzzy logic to deal with uncertainties and ambiguities in practical applications, which can significantly improve the accuracy of obstacle recognition, tracking and classification in complex scenes.

[0080] In a specific embodiment of the present invention, the specific analysis method of each information data of the static type obstacle on the path is:

[0081] Detect objects based on each frame of the video stream of the factory transport robot to obtain the width of the occupied path of static obstacles , the width of the path where the static type obstacle is located ;

[0082] In a specific embodiment of the present invention, the obstacle avoidance judgment value of the factory transport robot is calculated and the static obstacle avoidance response measures are implemented. The specific analysis method is as follows:

[0083] The width of the occupied path based on static type obstacles , the width of the factory transport robot , the width of the path where the static type obstacle is located , through the formula: , calculate the initial path passability judgment value ;

[0084] like , then it is judged that the initial path is allowed to pass, and the transport robot moves and deviates to avoid static obstacles;

[0085] like , then it is determined that the initial path is not allowed to pass, and a temporary obstacle avoidance path is executed.

[0086] In a specific embodiment of the present invention, the temporary obstacle avoidance path is specifically analyzed by:

[0087] The coordinates of the target position are extracted from the database, and a temporary obstacle avoidance path is generated based on the position coordinates of the point that the transport robot cannot pass through. The instruction information of the temporary obstacle avoidance path is sent to the control center of the transport robot, and the transport robot executes the temporary obstacle avoidance path.

[0088] In a specific embodiment of the present invention, the probability of each future path of the dynamic type obstacle on the path is analyzed by:

[0089] Through the fuzzy matching algorithm, the future paths of dynamic obstacles are generated, and through the probability cloud model: , calculate the probability value of each future path ,in represents a point coordinate vector on each future path, represents the coordinate vector of each future position, represents discreteness, and π represents pi.

[0090] In a specific embodiment of the present invention, the implementation of the dynamic obstacle avoidance response measure is specifically analyzed as follows:

[0091] Based on the probability values ​​of various future paths, the factory transport robot chooses a path that avoids obstacles.

[0092] The present invention uses a probabilistic method in path planning and obstacle avoidance problems, and can assign a probability value to different paths. This value represents the feasibility or risk of the path. Through cloud computing, a large number of possible paths can be calculated and evaluated, and the path with the lowest risk and the best cost can be selected, so that the problem of obstacle path replanning can be dealt with in a timely manner.

[0093] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they shall all fall within the protection scope of the present invention.

Claims

1. A path obstacle avoidance method based on fuzzy matching camera video stream and probabilistic cloud computing, characterized in that: The method comprises: Step 1: Initial path planning: Obtain the working data of the factory transport robot and the information data of each candidate path, calculate the matching index of each candidate path, and plan the initial path of the factory transport robot; Step 2: Predict the probability of obstacles on the initial path: Obtain the locations of each obstacle that has appeared in history on the initial path, analyze the probability of each obstacle appearing, and divide the obstacle risk area on the initial path; Step 3: Analyze the video stream to identify obstacles: Obtain the video stream of the factory transport robot during its operation, identify obstacles on the path through fuzzy matching, and classify the obstacles on the path to obtain static obstacles and dynamic obstacles. Step 4: Implement path obstacle avoidance response measures: obtain information data of static obstacles on the path, calculate the obstacle avoidance judgment value of the factory transport robot based on the information data of the factory transport robot, implement static obstacle avoidance response measures, obtain the probability values ​​of each future path of dynamic obstacles on the path, and implement dynamic obstacle avoidance response measures; The working data of the factory transport robot, including the width of the factory transport robot , Transportation time required ; The information data of each path to be selected includes the maximum width of each path to be selected , minimum width , the number of obstacles in each transportation process in history , Historical transportation duration ,in , Indicates the number of each candidate path, represents the total number of paths to be selected, , Indicates the number of each historical transportation process, Indicates the total number of historical transportation processes; The number of obstacles in each transportation process of each candidate path Perform mean processing, and the mean result is used as the historical average number of obstacles on each candidate path. ; By formula: , calculate the width of each candidate path to meet the state value , through the formula: , calculate the matching index of each candidate path ,in Indicates the acceptable number of pre-set path obstacles. represents the exponential function identifier, Represents the logarithmic function identifier; Based on the matching index of each candidate path Sort them and select the candidate path corresponding to the subscript of the maximum matching index, which is recorded as the initial path of the factory transport robot; Obtaining instruction information of the initial path of the factory transport robot from the database, and sending it to the control center of the factory transport robot, so that the factory transport robot performs the transport work; The starting position of the factory transport robot is set as the origin coordinate, the vertical direction is the longitudinal axis, and the horizontal right direction is the horizontal axis, and a two-dimensional coordinate system of the initial path of the factory transport robot is established; The positions of each obstacle in the history of the initial path are extracted from the database, and mapped to the two-dimensional coordinate system after position conversion to obtain the position coordinates of each obstacle in the history of the initial path. ,in , Indicates the number of obstacles that appeared in the history of the initial path. Indicates the total number of historical obstacles that occurred on the initial path; The location coordinates of each obstacle occurrence in the history of the initial path , statistics are located in Area where the secondary obstacle appears Total number of obstacles within Similarly, the total number of obstacles in the area where each obstacle appears is obtained , through the formula: , calculate the first The probability of obstacles in the coordinate area of ​​the location where the obstacle appears ,in, A mathematical symbol for a plane representing the area where obstacles appear. Indicates the preset judgment range scale; like , then the first The area where the secondary obstacle appears is the risk area, where Indicates the obstacle probability judgment threshold; Otherwise, the first The area where the secondary obstacle appears is the general area; Get the video stream of the factory transport robot and perform frame processing. For the video stream of the risk area, set the frame processing extraction interval to , for video streams in the general area, set the frame processing extraction interval to , obtain each frame image of the video stream of the factory transport robot, perform target detection on each frame image of the video stream of the factory transport robot through a deep learning model, and segment obstacles in the scene; Analyze the obstacle pixels in each frame of the video stream of the factory transport robot, and use the formula: , calculate the difference between consecutive frames ,in The current frame of the video stream representing the factory transport robot Pixel coordinates, The previous frame of the video stream of the factory transport robot Pixel coordinates, , represents the number of each pixel in the frame of the video stream of the factory transport robot, The total number of pixels in the frame of the video stream representing the factory transport robot; like , then determine the The pixels are dynamic areas, and the obstacle is a dynamic obstacle. represents the difference threshold of consecutive frames obtained from the database; Otherwise, judge The pixel points are static areas, and the obstacle is a static type obstacle; In summary, the obstacles and their types on the initial path of the factory transport robot are obtained; Detect objects based on each frame of the video stream of the factory transport robot to obtain the width of the occupied path of static obstacles , the width of the path where the static type obstacle is located ; The width of the occupied path based on static type obstacles , the width of the factory transport robot , the width of the path where the static type obstacle is located , through the formula: , calculate the initial path passability judgment value ; like , then it is judged that the initial path is allowed to pass, and the transport robot moves and deviates to avoid static obstacles; like , then it is determined that the initial path is not allowed to pass, and a temporary obstacle avoidance path is executed; Extract the coordinates of the target position from the database, generate a temporary obstacle avoidance path based on the position coordinates of the point that the transport robot cannot pass through, send the instruction information of the temporary obstacle avoidance path to the control center of the transport robot, and the transport robot executes the temporary obstacle avoidance path; Through the fuzzy matching algorithm, the future paths of dynamic obstacles are generated, and through the probability cloud model: , calculate the probability value of each future path ,in represents a point coordinate vector on each future path, represents the coordinate vector of each future position, represents discreteness, and π represents pi.

2. The path obstacle avoidance method based on fuzzy matching camera video stream and probabilistic cloud computing according to claim 1 is characterized in that: The specific analysis method for implementing the dynamic obstacle avoidance response measures is as follows: Based on the probability values ​​of various future paths, the factory transport robot chooses a path that avoids obstacles.

Images