Driving robot and control method thereof

Abstract

The present disclosure relates to a driving robot capable of driving to a destination without map and location information and a control method thereof. The driving robot may comprise: a sensor unit for acquiring surrounding environment information; an input unit for input of a movement command; a drive motor unit for moving the robot; and a control unit for controlling the drive motor unit on the basis of the surrounding environment information and the movement command, wherein the control unit generates road structure information by recognizing a road structure on the basis of the surrounding environment information, selects a movement target point and an action of the robot on the basis of the road structure information and the movement command, determines a driving operation of the robot by inputting the surrounding environment information and selection information including the movement target point and the action of the robot into a pre-trained neural network model, and controls the drive motor unit to move the robot on the basis of the determined driving operation.

Description

Driving robot and its control method

[0001] The present disclosure relates to a driving robot capable of driving to a destination based on artificial intelligence without a map or location information, and a control method thereof.

[0002] In general, a robot is a machine that automatically processes or operates a given task with its own abilities, and the application fields of robots are generally classified into various fields such as industrial, medical, space, and underwater.

[0003] Recently, there has been an increasing trend of robots that can communicate or interact with humans through voice or gestures.

[0004] These robots can include various types of robots, such as guide robots that are placed in specific locations and provide various information to users, or home robots that are installed in homes.

[0005] In order for a robot to drive, it must find its location on a map, have a path to the destination location on the map, and be able to control its motors to follow that path.

[0006] The location of the robot on the map can be determined through the SLAM (Simultaneous Localization and Mapping) function, which compares the robot's sensor information, including camera images, lidar, and distances to obstacles around the robot, with map information.

[0007] In addition, the path generation to the destination is divided into a process of generating a global path that generates a path from the robot location to the destination location on the map and a local path to avoid obstacles around the robot, and the robot can control the motor through a motion controller to follow the local path that is ultimately generated.

[0008] However, the robot had a problem in that it could not know the path to the destination if it failed to recognize its own location or failed to generate a path such as a global path or a local path, and thus stopped for safety reasons until it received assistance from an outsider.

[0009] In this case, there was the inconvenience of having to indefinitely suspend services such as robot delivery and guidance until outside assistance was available, or having to move the robot to another location and restart it to re-recognize its location.

[0010] Therefore, in the future, there is a need to develop a driving robot that can drive to a destination based on a given path movement command and road structure recognition without a map or location information.

[0011] The present disclosure aims to solve the above-mentioned problems and other problems.

[0012] The present disclosure provides a driving robot and a control method thereof that can drive stably to a destination while recognizing the structure of a road without a map or location information by selecting a moving target point and action of the robot based on road structure information and movement commands and determining the driving action of the robot based on artificial intelligence.

[0013] According to one embodiment of the present disclosure, a driving robot includes a sensor unit for acquiring information about a surrounding environment, an input unit for inputting a movement command, a drive motor unit for moving the robot, and a control unit for controlling the drive motor unit based on the information about the surrounding environment and the movement command, wherein the control unit recognizes a road structure based on the information about the surrounding environment to generate road structure information, selects a movement target point and an action of the robot based on the road structure information and the movement command, inputs selection information including the information about the surrounding environment and the movement target point and action of the robot into a pre-learned neural network model to determine a driving action of the robot, and controls the drive motor unit so that the robot moves based on the determined driving action.

[0014] A control method of a driving robot according to one embodiment of the present disclosure is a control method of a driving robot including a drive motor unit, the control method including a step of obtaining surrounding environment information and a movement command, a step of recognizing a road structure based on the surrounding environment information and generating road structure information, a step of selecting a movement target point and an action of the robot based on the road structure information and the movement command, a step of inputting selection information including the surrounding environment information and the movement target point and action of the robot into a pre-learned neural network model to determine a driving direction and a speed of the robot, and a step of controlling the drive motor unit so that the robot moves based on the determined driving direction and speed.

[0015] According to one embodiment of the present disclosure, a driving robot can drive stably to a destination while recognizing the structure of the road without a map and location information by selecting a moving target point and action of the robot based on road structure information and a movement command and determining the driving action of the robot based on artificial intelligence.

[0016] In this way, since the present disclosure enables the robot to drive even without a map and location information, scanning and mapping work for robot driving are not required, thereby saving the cost and time consumed therefor.

[0017] Additionally, the present disclosure can be applied even when the robot temporarily loses its location information.

[0018] That is, the present disclosure can be operated as a relatively lightweight system because the robot does not need to have map and location information, and is easy to apply to new locations.

[0019] FIG. 1 is a drawing for explaining a driving robot according to one embodiment of the present disclosure.

[0020] FIG. 2 is a diagram for explaining a process of acquiring information on the surrounding environment of a driving robot according to an embodiment of the present disclosure.

[0021] FIG. 3 and FIG. 4 are diagrams for explaining a process of generating a measurement distance map of a driving robot according to an embodiment of the present disclosure.

[0022] FIG. 5 is a diagram for explaining a process of generating road structure information of a driving robot according to an embodiment of the present disclosure.

[0023] FIG. 6 is a diagram for explaining a process for calculating an intersection collision area of ​​a driving robot according to an embodiment of the present disclosure.

[0024] FIGS. 7 to 9 are drawings for explaining path structure information of a driving robot according to an embodiment of the present disclosure.

[0025] FIG. 10 and FIG. 11 are drawings for explaining a control unit of a driving robot according to an embodiment of the present disclosure.

[0026] FIG. 12 and FIG. 13 are diagrams for explaining a global map analysis process of a driving robot according to one embodiment of the present disclosure.

[0027] FIG. 14 is a diagram for explaining a global path analysis process of a driving robot according to an embodiment of the present disclosure.

[0028] FIG. 15 is a drawing for explaining a movement process based on a movement command script of a driving robot according to one embodiment of the present disclosure.

[0029] FIG. 16 is a diagram for explaining a process of generating a map-based movement command script for a driving robot according to one embodiment of the present disclosure.

[0030] Figure 17 is a flowchart for explaining a control process of a driving robot according to one embodiment of the present disclosure.

[0031] Hereinafter, embodiments disclosed in this specification will be described in detail with reference to the attached drawings. Regardless of the drawing numbers, identical or similar components will be given the same reference numbers and redundant descriptions thereof will be omitted. The suffixes "module" and "part" used for components in the following description are assigned or used interchangeably only for the convenience of writing the specification, and do not in themselves have distinct meanings or roles. In addition, when describing the embodiments disclosed in this specification, if it is determined that a specific description of a related known technology may obscure the gist of the embodiments disclosed in this specification, a detailed description thereof will be omitted. In addition, the attached drawings are only intended to facilitate easy understanding of the embodiments disclosed in this specification, and the technical ideas disclosed in this specification are not limited by the attached drawings, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and technical scope of the present disclosure.

[0032] Terms that include ordinal numbers, such as first, second, etc., may be used to describe various components, but the components are not limited by these terms. These terms are used solely to distinguish one component from another.

[0033] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components intervening. Conversely, when a component is referred to as being "directly connected" or "connected" to another component, it should be understood that there are no other components intervening.

[0034] Additionally, throughout this specification, the terms "neural network," "neural network," and "network function" may be used interchangeably. A neural network may be comprised of a set of interconnected computational units, which may generally be referred to as "nodes." These "nodes" may also be referred to as "neurons." A neural network comprises at least two or more nodes. The nodes (or neurons) constituting the neural networks may be interconnected by one or more "links."

[0035] Artificial Intelligence (AI) refers to the field of study that studies artificial intelligence or the methodologies for creating it, while Machine Learning (ML) refers to the field that defines various problems in the field of AI and studies the methodologies for solving them. Machine learning is also defined as an algorithm that improves performance on a task through consistent experience.

[0036] An artificial neural network (ANN) is a model used in machine learning. It can refer to a model with problem-solving capabilities, comprised of artificial neurons (nodes) formed by the connection of synapses. An ANN can be defined by the connection patterns between neurons in different layers, the learning process that updates model parameters, and the activation function that generates output values.

[0037] An artificial neural network may include an input layer, an output layer, and optionally one or more hidden layers. Each layer may contain one or more neurons, and the artificial neural network may include synapses connecting neurons. In an artificial neural network, each neuron can output a function value of an activation function based on input signals, weights, and biases received through the synapses.

[0038] Model parameters are parameters determined through learning, including synaptic connection weights and neuron biases. Hyperparameters are parameters that must be set before learning in machine learning algorithms, including the learning rate, number of iterations, mini-batch size, and initialization function.

[0039] The goal of artificial neural network training can be seen as determining model parameters that minimize a loss function. The loss function can be used as an indicator for determining optimal model parameters during the artificial neural network training process.

[0040] Machine learning can be classified into supervised learning, unsupervised learning, and reinforcement learning depending on the learning method.

[0041] Supervised learning refers to a method for training an artificial neural network given labels for training data. Labels can refer to the correct answer (or output) that the artificial neural network must infer when the training data is input to the artificial neural network. Unsupervised learning refers to a method for training an artificial neural network without given labels for the training data. Reinforcement learning refers to a learning method in which an agent defined within a given environment is trained to select actions or action sequences that maximize cumulative rewards in each state.

[0042] Among artificial neural networks, machine learning implemented with a deep neural network (DNN) containing multiple hidden layers is sometimes called deep learning, and deep learning is a subset of machine learning. Hereinafter, "machine learning" is used to encompass deep learning.

[0043] FIG. 1 is a drawing for explaining a driving robot according to one embodiment of the present disclosure.

[0044] As illustrated in FIG. 1, the driving robot (10) of the present disclosure may include a sensor unit (100) that acquires surrounding environment information, an input unit (200) that inputs a movement command, a drive motor unit (300) that moves the robot, and a control unit (400) that controls the drive motor unit (300) based on the surrounding environment information and the movement command.

[0045] Here, the sensor unit (100) may include at least one of a 3D camera, a lidar sensor, an ultrasonic sensor, a distance sensor, and a virtual sensor.

[0046] In addition, the control unit (400) recognizes the road structure based on the surrounding environment information to generate road structure information, selects the movement target point and action of the robot based on the road structure information and the movement command, inputs selection information including the surrounding environment information and the movement target point and action of the robot into a pre-learned neural network model to determine the driving direction and speed of the robot, and controls the drive motor unit (300) to move the robot based on the determined driving direction and speed.

[0047] When generating road structure information, the control unit (400) collects surrounding environment information from the sensor unit (100), generates a measurement distance map based on the surrounding environment information, and inputs the measurement distance map into a pre-learned neural network model to generate road structure recognition and road structure information.

[0048] Here, the control unit (400) can generate a measurement distance map based on the surrounding environment information measured from the sensor unit (100) including the robot's front lidar sensor, side lidar sensor, front ultrasonic and distance sensor, side ultrasonic and distance sensor, front depth camera, rear depth camera, visual object recognition sensor for object detection, and rear virtual sensor when generating a measurement distance map.

[0049] For example, when generating a measurement distance map, the control unit (400) may generate a first measurement distance map based on surrounding environment information measured from a lidar sensor, generate a second measurement distance map based on surrounding environment information measured from a depth camera, and generate a third measurement distance map based on surrounding environment information measured from a visual object recognition sensor.

[0050] At this time, the control unit (400) can generate a measurement distance map that includes the distance to the obstacle located at the closest distance from the robot by fusing the first measurement distance map, the second measurement distance map, and the third measurement distance map when generating the measurement distance map.

[0051] For example, when generating a measurement distance map, the control unit (400) can generate a measurement distance map including a total of 360 measurement distance values ​​at 1-degree intervals from -180 degrees to +180 degrees.

[0052] In addition, when generating road structure information, the control unit (400) scales one-dimensional sensor input data including a measured distance map to a preset size, classifies the scaled input data by feature extraction item, and inputs the classified feature extraction items into a plurality of pre-learned neural network models, thereby generating road structure information by feature extraction item.

[0053] Here, the neural network model may include a neural network set structure that combines a convolutional neural network and a neural network.

[0054] For example, a neural network model can have multiple sets of neural networks arranged in parallel, and the multiple sets of neural networks can input different feature extraction items as input data and simultaneously output path structure information for each different feature extraction item.

[0055] For example, among the plurality of neural network sets, a first neural network set may output first road structure information including a road direction angle, a second neural network set among the plurality of neural network sets may output second road structure information including a collision area by intersection direction based on the current location, and a third neural network set among the plurality of neural network sets may output third road structure information including a collision area by intersection direction ahead at a specific distance.

[0056] Additionally, the control unit (400) can individually train and update multiple sets of neural networks.

[0057] And, when generating road structure information, the control unit (400) recognizes a straight line longer than a certain distance as a corridor wall if it is detected, and if the straight line of the wall is located on the left and right of the robot, it can calculate the road direction and corridor direction based on the straight line of the wall.

[0058] In some cases, when generating road structure information, the control unit (400) may recognize a straight line longer than a certain distance as a corridor wall, and recognize the location as an intersection if the straight line on the wall is positioned so as to face the road.

[0059] In another case, when generating road structure information, the control unit (400) can recognize a straight line longer than a certain distance as a corridor wall if detected, and can recognize the corresponding location as a vertex or corner if the straight line on the wall meets the road at a right angle.

[0060] Next, when selecting a movement target point and action of the robot, the control unit (400) can check the type of movement command when a movement command is input, generate a movement command script corresponding to the type of movement command, and select a movement target point and action of the robot based on the movement command script and path structure information.

[0061] Here, when checking the type of movement command, the control unit (400) can use a script containing the movement command content as a movement command script if the movement command is a script type containing the movement command content from the starting point to the destination.

[0062] In some cases, when checking the type of movement command, the control unit (400) may generate a movement command script based on the analysis information if the movement command is a type of analysis information that analyzed the map and path given before the failure of the robot's position recognition.

[0063] Here, the control unit (400) may include a global path analyzer that provides path analysis information by analyzing a global path based on a map given before the robot's location recognition failure, and a global map analyzer that provides map analysis information by analyzing a global map around the path based on a map given before the robot's location recognition failure.

[0064] For example, a global map analyzer can analyze a global map around a route to generate movement commands that include intersection information and road structure information.

[0065] That is, the global map analyzer can determine the map structure around the robot position on the estimated path and the map structure around the estimated target waypoint.

[0066] Additionally, the global map analyzer can determine the map structure around the robot position based on the distance to the left wall and the distance to the right wall on the estimated path, and the direction angle of the left wall and the direction angle of the right wall on the global path.

[0067] Here, the global map analyzer can determine the map structure around the target waypoint based on intersection type and intersection direction information.

[0068] For example, a global map analyzer can generate a movement command script that includes the distance the robot will move, the action direction after the movement, intersection type information, and direction information at the intersection.

[0069] At this time, the global map analyzer can provide direction information at intersections in the direction of the hour hand.

[0070] Additionally, the global path analyzer can calculate the robot's position on the path through odometry position estimation when a robot's position error occurs, and can analyze the current movement path to generate a movement command for the remaining path to the destination.

[0071] Here, the global path analyzer can generate a movement command script that includes the distance the robot will move and the action direction after the movement.

[0072] In another case, when checking the type of movement command, the control unit (400) may generate a movement command script based on a map including a movement mark if the movement command is a map type including a movement mark from a starting point to a destination.

[0073] Here, the control unit (400) can check whether the straight distance of the straight road and the straight distance between the destination and the intersection are set within the schematic diagram based on the block structure including the movement mark, and if both the straight distance of the straight road and the straight distance between the destination and the intersection are set, a movement command script can be generated.

[0074] That is, when generating a movement command script, the control unit (400) can automatically recognize the width of the intersection while the robot is driving, automatically recognize the return driving path of the robot, and reflect the recognized width of the intersection and the return driving path in the movement command script.

[0075] As another case, when checking the type of the movement command, the control unit (400) may generate a movement command script based on the map editing map if the movement command is of the map editing map type.

[0076] Here, the control unit (400) can generate a map editing map by performing a test that provides a floor plan of the map, adjusts transparency for the floor plan of the map, inputs the scale and information of the map, inputs the route information of the map, inputs the destination of the map, and selects the starting point and destination of the map to display the route when generating a map editing map.

[0077] For example, when inputting the scale of a map, the control unit (400) may set the scale of the map by selecting two points within the map and inputting the distance between the two points, when inputting map information, input map information including at least one of the region name and the number of floors of the map, and when inputting map road information, input corridor information including the location and direction of the road and the corridor width.

[0078] In addition, when interpreting a movement command script, the control unit (400) can sequentially interpret, in the arrangement order of the unit action commands, a unit action command including information on the robot's driving direction, a distance to be moved, an action completion condition, and the number of times to skip a designated intersection exists in the movement command script.

[0079] Next, the control unit (400) may include a movement command generator that generates a movement command script, a path structure recognizer that recognizes a path structure based on surrounding environment information and generates path structure information, an action controller that selects a movement target point and an action of the robot based on the path structure information and the movement command, and a driving controller that inputs selection information including surrounding environment information and the movement target point and action of the robot into a pre-learned neural network model to determine a driving action of the robot, and controls a drive motor unit (300) so that the robot moves based on the determined driving action.

[0080] In this way, the present disclosure determines the driving motion of the robot based on artificial intelligence by selecting the moving target point and action of the robot based on road structure information and movement commands, thereby enabling the robot to drive stably while recognizing the road structure to the destination even without map and location information.

[0081] In this way, since the present disclosure enables the robot to drive even without a map and location information, scanning and mapping work for robot driving are not required, thereby saving the cost and time consumed therefor.

[0082] Additionally, the present disclosure can be applied even when the robot temporarily loses its location information.

[0083] That is, the present disclosure can be operated as a relatively lightweight system because the robot does not need to have map and location information, and is easy to apply to new locations.

[0084] FIG. 2 is a diagram for explaining a process of acquiring information on the surrounding environment of a driving robot according to an embodiment of the present disclosure.

[0085] As illustrated in FIG. 2, the driving robot (10) of the present disclosure must be able to understand the structure of the currently given path in order to drive according to a movement command.

[0086] The driving robot (10) of the present disclosure may include a plurality of sensors for obtaining information about the surrounding environment, and the plurality of sensors may include at least one of a 3D camera, a lidar sensor, an ultrasonic sensor, a distance sensor, and a virtual sensor.

[0087] Here, the driving robot (10) can recognize the road structure based on the surrounding environment information and generate road structure information.

[0088] A driving robot (10) can obtain information about its surrounding environment from a front lidar sensor, a side lidar sensor, a front ultrasonic and distance sensor, a side ultrasonic and distance sensor, a front depth camera, a rear depth camera, a visual object recognition sensor for object detection, and a rear virtual sensor.

[0089] FIG. 3 and FIG. 4 are diagrams for explaining a process of generating a measurement distance map of a driving robot according to an embodiment of the present disclosure.

[0090] As illustrated in FIGS. 3 and 4, the driving robot of the present disclosure can collect surrounding environment information from a sensor, generate a measurement distance map based on the surrounding environment information, and input the measurement distance map into a pre-learned neural network model to recognize a road structure and generate road structure information.

[0091] Here, as shown in FIG. 3, the driving robot can generate a measurement distance map based on the surrounding environment information measured from the robot's front lidar sensor, side lidar sensor, front ultrasonic and distance sensor, side ultrasonic and distance sensor, front depth camera, rear depth camera, visual object recognition sensor for object detection, and rear virtual sensor.

[0092] For example, when generating a measurement distance map, a driving robot may generate a first measurement distance map based on surrounding environment information measured from a lidar sensor, a second measurement distance map based on surrounding environment information measured from a depth camera, and a third measurement distance map based on surrounding environment information measured from a visual object recognition sensor.

[0093] At this time, when generating a measurement distance map, the driving robot can generate a measurement distance map including the distance to an obstacle located at the closest distance from the robot by fusing the first measurement distance map, the second measurement distance map, and the third measurement distance map.

[0094] For example, a driving robot can generate a measurement distance map containing a total of 360 measurement distance values ​​at 1 degree intervals from -180 degrees to +180 degrees.

[0095] As shown in Fig. 4, the driving robot can generate a first measurement distance map for sensing a wall based on the surrounding environment information measured from the lidar sensor.

[0096] In addition, as shown in FIG. 4, the driving robot can generate a first measurement distance map for sensing a wall based on the surrounding environment information measured from the lidar sensor, a second measurement distance map for sensing a chair based on the surrounding environment information measured from the depth camera, and a third measurement distance map for sensing an object on the floor based on the surrounding environment information measured from the visual object recognition sensor, and fuse these to recognize the distance to the obstacle located at the closest distance from the robot.

[0097] FIG. 5 is a diagram for explaining a process of generating road structure information of a driving robot according to an embodiment of the present disclosure.

[0098] As illustrated in FIG. 5, the present disclosure scales one-dimensional sensor input data including a measured distance map to a preset size when generating road structure information, classifies the scaled input data by feature extraction items, and inputs the classified feature extraction items into a plurality of pre-learned neural network models, thereby generating road structure information by feature extraction item.

[0099] Here, the neural network model may include a neural network set structure that combines a convolutional neural network and a neural network.

[0100] For example, a neural network model can have multiple sets of neural networks arranged in parallel, and the multiple sets of neural networks can input different feature extraction items as input data and simultaneously output path structure information for each different feature extraction item.

[0101] For example, among the plurality of neural network sets, a first neural network set may output first road structure information including a road direction angle, a second neural network set among the plurality of neural network sets may output second road structure information including a collision area by intersection direction based on the current location, and a third neural network set among the plurality of neural network sets may output third road structure information including a collision area by intersection direction ahead at a specific distance.

[0102] Additionally, the present disclosure can individually train and update multiple sets of neural networks.

[0103] FIG. 6 is a diagram for explaining a process for calculating an intersection collision area of ​​a driving robot according to an embodiment of the present disclosure.

[0104] As illustrated in FIG. 6, when generating road structure information, the present disclosure recognizes a straight line longer than a certain distance as a corridor wall (15) and, if the straight line of the wall (15) is located on the left and right of the robot (10), the road direction and corridor direction can be calculated based on the straight line of the wall (15).

[0105] In addition, the present disclosure can pre-train a neural network model to check the length and width of the intersection check area based on the intersection start and end points when the intersection check area is 90 degrees to the right, and to calculate the area of ​​collision with the wall by assuming the intersection space for each direction.

[0106] FIGS. 7 to 9 are drawings for explaining path structure information of a driving robot according to an embodiment of the present disclosure.

[0107] As illustrated in FIG. 7, when generating road structure information, the present disclosure recognizes a straight line (21) of a certain distance or more as a corridor wall if detected, and if the straight line (21) of the wall is located on the left and right of the robot (10), the road direction and corridor direction can be calculated based on the straight line (21) of the wall.

[0108] In addition, as illustrated in FIG. 8, when generating road structure information, the present disclosure recognizes a straight line (21) of a certain distance or more as a corridor wall if it is detected, and recognizes the location as an intersection (24) if the straight line (21) of the wall is positioned so as to face the road (22).

[0109] In addition, as illustrated in FIG. 9, when generating road structure information, the present disclosure recognizes a straight line (21) of a certain distance or longer as a corridor wall if detected, and recognizes the corresponding location as a vertex and corner if the straight line (21) of the wall meets the road (25) at a right angle.

[0110] FIG. 10 and FIG. 11 are drawings for explaining a control unit of a driving robot according to an embodiment of the present disclosure.

[0111] As illustrated in FIG. 10, the control unit of the driving robot may include a movement command generator (110) that generates a movement command script, a road structure recognizer (120) that recognizes a road structure based on surrounding environment information and generates road structure information, an action controller (130) that selects a movement target point and an action of the robot based on the road structure information and the movement command, and a driving controller (140) that inputs selection information including surrounding environment information and the movement target point and action of the robot into a pre-learned neural network model to determine a driving action of the robot, and controls a drive motor so that the robot moves based on the determined driving action.

[0112] Here, the present disclosure can determine the type of the movement command when a movement command is input, generate a movement command script corresponding to the type of the movement command, and select a movement target point and action of the robot based on the movement command script and path structure information.

[0113] In the present disclosure, when checking the type of a movement command, if the movement command is a script type that includes movement command contents from a starting point to a destination, a script including the movement command contents can be used as a movement command script.

[0114] In addition, as illustrated in FIG. 11, when confirming the type of movement command, the present disclosure can generate a movement command script based on the analysis information if the movement command is a type of analysis information that analyzed a map and path given before the failure of the robot's position recognition.

[0115] Here, the present disclosure may include a global path analyzer (150) that provides path analysis information by analyzing a global path based on a given map before a failure in position recognition of the robot, and a global map analyzer (160) that provides map analysis information by analyzing a global map around the path based on a given map before a failure in position recognition of the robot.

[0116] For example, a global map analyzer (160) can analyze a global map around a route to generate a movement command that includes intersection information and road structure information.

[0117] That is, the global map analyzer (160) can determine the map structure around the robot position on the estimated path and determine the map structure around the estimated target waypoint.

[0118] In addition, when determining the map structure around the robot position, the global map analyzer (160) can determine the map structure based on the distance from the left wall and the distance from the right wall on the estimated path, and the direction angle of the left wall and the direction angle of the right wall on the global path.

[0119] Here, the global map analyzer (160) can determine the map structure based on intersection type and intersection direction information when determining the map structure around the target waypoint.

[0120] For example, the global map analyzer (160) can generate a movement command script that includes the distance the robot will move, the action direction after movement, intersection type information, and direction information at the intersection.

[0121] At this time, the global map analyzer (160) can provide direction information at an intersection in the direction of the hour hand.

[0122] In addition, the global path analyzer (150) can calculate the robot's position on the path through odometry position estimation when a robot's position error occurs, and analyze the current movement path to generate a movement command for the remaining path to the destination.

[0123] Here, the global path analyzer (150) can generate a movement command script including the distance the robot will move and the action direction after movement.

[0124] FIG. 12 and FIG. 13 are diagrams for explaining a global map analysis process of a driving robot according to one embodiment of the present disclosure.

[0125] As illustrated in FIGS. 12 and 13, the global map analyzer can analyze a global map around a path to generate a movement command including intersection information and road structure information.

[0126] That is, the global map analyzer can determine the map structure around the robot position on the estimated path and the map structure around the estimated target waypoint.

[0127] In addition, the global map analyzer, as shown in FIG. 12, can determine the map structure around the location of the robot (10) based on the distance to the left wall and the distance to the right wall on the estimated path, and the direction angle of the left wall and the direction angle of the right wall on the global path.

[0128] Additionally, the global map analyzer can determine the map structure around the target waypoint based on intersection type and intersection direction information, as shown in FIG. 13.

[0129] For example, a global map analyzer can generate a movement command script that includes the distance the robot will move, the action direction after the movement, intersection type information, and direction information at the intersection.

[0130] At this time, the global map analyzer can provide direction information at intersections in the direction of the hour hand.

[0131] For example, a movement command script that includes the distance the robot will move, the action direction after movement, intersection type information, and direction information at the intersection can be generated as follows.

[0132] The move command script <action direction=""-90"" distance=""2.1”" xroad=""LFR"" xroad_directions_hr=""12.0," 9.0, 3.0" / >In this case, the move command script can be interpreted as turning 90 degrees to the left 2.1m ahead, the intersection type is LFR(+), and the intersection directions are 12 o'clock, 9 o'clock, and 3 o'clock.

[0133] FIG. 14 is a diagram for explaining a global path analysis process of a driving robot according to an embodiment of the present disclosure.

[0134] As illustrated in Fig. 14, the global path analyzer can calculate the robot's position on the path through odometry position estimation when a position error occurs in the robot (10), and analyze the current movement path to generate a movement command for the remaining path to the destination.

[0135] Here, the global path analyzer can generate a movement command script that includes the distance the robot (10) will move and the action direction after movement.

[0136] For example, a movement command script that includes the distance the robot will move and the action direction after the movement can be generated as follows.

[0137] The move command script <action direction=""-90"" distance=""2.1”" / > In this case, the move command script can be interpreted as a turn 90 degrees to the left from 2.1m in front.

[0138] FIG. 15 is a drawing for explaining a movement process based on a movement command script of a driving robot according to one embodiment of the present disclosure.

[0139] As illustrated in FIG. 15, when interpreting a movement command script, the present disclosure can sequentially interpret, according to the arrangement order of the unit action commands, if there are unit action commands within the movement command script that include information on the robot's driving direction, distance to be moved, action completion conditions, and number of times to skip a designated intersection.

[0140] For example, the move command script <action direction="STRAIGHT" distance="1.0" stop_cond="LF" skip_num="1" / > In this case, the movement command script can be interpreted as the driving direction is straight, the first intersection is skipped once after moving forward at least 1m, and straight until the second LF(ㅓ) intersection appears.

[0141] Next, the move command script <action direction="LEFT" distance="3.0" / > In this case, the move command script can be interpreted as making a left turn at the intersection and then moving to a point 3 meters away.

[0142] FIG. 16 is a diagram for explaining a process of generating a map-based movement command script for a driving robot according to one embodiment of the present disclosure.

[0143] As illustrated in FIG. 16, the present disclosure can generate a movement command script based on a map (50) including a movement mark if the movement command is a map type including a movement mark from a starting point to a destination.

[0144] Here, the present disclosure can check whether the straight distance of a straight road and the straight distance between a destination and an intersection are set within the schematic diagram (50) of a block structure including a movement mark, and if both the straight distance of the straight road and the straight distance between the destination and the intersection are set, a movement command script can be generated.

[0145] As shown in Fig. 16, when a robot (10) moves in the direction of the arrow in front of an elevator (E) and inputs a map (50) with a destination (T) 3 m ahead, a corresponding movement command script can be generated based on this.

[0146] The present disclosure can automatically recognize the width of an intersection during robot driving when generating a movement command script, automatically recognize the return driving path of the robot, and reflect the recognized width of the intersection and the return driving path in the movement command script.

[0147] In another case, the present disclosure may generate a move command script based on a map edit map if the move command is of a map edit map type.

[0148] Here, the present disclosure can generate a map editing map by performing a test that provides a floor plan of the map, adjusts transparency for the floor plan of the map, inputs scale and information of the map, inputs route information of the map, inputs a destination of the map, and displays a route by selecting a starting point and a destination of the map.

[0149] For example, in the present disclosure, when inputting the scale of a map, the scale of the map can be set by selecting two points within the map and inputting the distance between the two points, when inputting information of the map, information of the map including at least one of the region name and the number of floors of the map can be input, and when inputting road information of the map, corridor information including the location and direction of the road and the width of the corridor can be input.

[0150] Figure 17 is a flowchart for explaining a control process of a driving robot according to one embodiment of the present disclosure.

[0151] As illustrated in FIG. 17, the present disclosure can obtain surrounding environment information and movement commands when a robot drives (S10), and recognize a road structure based on the surrounding environment information to generate road structure information (S30).

[0152] Next, the present disclosure recognizes the current movement command step based on the road structure information and the movement command (S20), and can select the movement target point and action of the robot.

[0153] Here, the present disclosure can determine the type of the movement command when a movement command is input, generate a movement command script corresponding to the type of the movement command, and select a movement target point and action of the robot based on the movement command script and path structure information.

[0154] In the present disclosure, when checking the type of a movement command, if the movement command is a script type that includes movement command contents from a starting point to a destination, a script including the movement command contents can be used as a movement command script.

[0155] In some cases, the present disclosure may generate a movement command script based on the analysis information when determining the type of movement command, if the movement command is a type of analysis information that analyzed a map and path given before the robot's location recognition failure.

[0156] In another case, the present disclosure may generate a movement command script based on a map including a movement mark when determining the type of a movement command, if the movement command is a map type including a movement mark from a starting point to a destination.

[0157] As another case, the present disclosure may generate a movement command script based on a map edit map when checking the type of the movement command, if the movement command is a map edit map type.

[0158] Next, the present disclosure can determine the driving motion of the robot by inputting selection information including surrounding environment information and the robot's movement target point and action into a pre-learned neural network model (S40).

[0159] And, the present disclosure can generate a robot control command to move the robot based on the determined driving motion (S50).

[0160] In this way, the present disclosure determines the driving motion of the robot based on artificial intelligence by selecting the moving target point and action of the robot based on road structure information and movement commands, thereby enabling the robot to drive stably while recognizing the road structure to the destination even without map and location information.

[0161] In this way, since the present disclosure enables the robot to drive even without a map and location information, scanning and mapping work for robot driving are not required, thereby saving the cost and time consumed therefor.

[0162] Additionally, the present disclosure can be applied even when the robot temporarily loses its location information.

[0163] That is, the present disclosure can be operated as a relatively lightweight system because the robot does not need to have map and location information, and is easy to apply to new locations.

[0164] According to the driving robot according to the present disclosure, the robot's driving motion is determined based on artificial intelligence by selecting the robot's moving target point and actions based on road structure information and movement commands, thereby enabling the robot to drive stably while recognizing the road structure to a destination without a map or location information, and thus has significant industrial applicability.

Claims

1. Sensor unit that obtains information about the surrounding environment; Input section where movement commands are entered; A drive motor that moves the robot; and It includes a control unit that controls the driving motor unit based on the above-mentioned surrounding environment information and movement commands, The above control unit, A driving robot characterized in that it recognizes a road structure based on the surrounding environment information to generate road structure information, selects a movement target point and an action of the robot based on the road structure information and the movement command, inputs selection information including the surrounding environment information and the movement target point and action of the robot into a pre-learned neural network model to determine a driving direction and speed of the robot, and controls the drive motor unit so that the robot moves based on the determined driving direction and speed.

2. In paragraph 1, The above control unit, A driving robot characterized in that, when generating the above road structure information, the surrounding environment information is collected from the sensor unit, a measurement distance map is generated based on the surrounding environment information, and the measurement distance map is input into a pre-learned neural network model to recognize the road structure and generate road structure information.

3. In paragraph 2, The above control unit, A driving robot characterized in that, when generating the above road structure information, one-dimensional sensor input data including the measured distance map is scaled to a preset size, the scaled input data is classified by feature extraction items, and the classified feature extraction items are each input into a plurality of pre-learned neural network models to generate road structure information by feature extraction item.

4. In paragraph 2, The above control unit, A driving robot characterized in that when generating the above road structure information, if a straight line longer than a certain distance is detected, it is recognized as a corridor wall, and if the straight line of the wall is located on the left and right of the robot, the road direction and corridor direction are calculated based on the straight line of the wall.

5. In paragraph 2, The above control unit, A driving robot characterized in that, when generating the above road structure information, if a straight line longer than a certain distance is detected, it is recognized as a corridor wall, and if the straight line of the wall is positioned so as to face the road, the corresponding location is recognized as an intersection.

6. In paragraph 2, The above control unit, A driving robot characterized in that when generating the above road structure information, if a straight line longer than a certain distance is detected, it is recognized as a corridor wall, and if the straight line of the wall intersects the road at a right angle, the corresponding location is recognized as a vertex and corner.

7. In paragraph 1, The above control unit, A mobile robot characterized in that when selecting a movement target point and action of the robot, when the movement command is input, the type of the movement command is checked, a movement command script is generated corresponding to the type of the movement command, and the movement target point and action of the robot are selected based on the movement command script and the path structure information.

8. In paragraph 7, The above control unit, A mobile robot characterized in that, when checking the type of the movement command, if the movement command is a script type that includes movement command contents from a starting point to a destination, a script including the movement command contents is used as the movement command script.

9. In paragraph 7, The above control unit, A mobile robot characterized in that, when checking the type of the movement command, if the movement command is a type of analysis information that analyzed a map and path given before the failure of the robot's location recognition, the movement command script is generated based on the analysis information.

10. In paragraph 7, The above control unit, A mobile robot characterized in that when checking the type of the movement command, if the movement command is a map type including a movement mark from a starting point to a destination, the movement command script is generated based on the map including the movement mark.

11. In paragraph 10, The above control unit, A mobile robot characterized in that it checks whether the straight distance of a straight road and the straight distance between a destination and an intersection are set within the map based on a block structure including the above movement mark, and if both the straight distance of the straight road and the straight distance between the destination and the intersection are set, it generates the movement command script.

12. In paragraph 7, The above control unit, A mobile robot characterized in that when checking the type of the movement command, if the movement command is a map editing map type, the movement command script is generated based on the map editing map.

13. In paragraph 12, The above control unit, A driving robot characterized in that, when generating the above-mentioned map editing map, the map editing map is generated by providing a floor plan of the map, adjusting transparency of the floor plan of the map, entering scale and information of the map, entering path information of the map, entering a destination of the map, and performing a test to display a path by selecting a starting point and a destination of the map.

14. In paragraph 1, The above control unit, A move command generator that generates move command scripts; A road structure recognizer that recognizes a road structure based on the above surrounding environment information and generates road structure information; An action controller that selects a movement target point and action of the robot based on the above-mentioned path structure information and the above-mentioned movement command; and A driving robot characterized in that it includes a driving controller that inputs the surrounding environment information and selection information including the movement target point and action of the robot into a pre-learned neural network model to determine the driving motion of the robot, and controls the driving motor unit so that the robot moves based on the determined driving motion.

15. A method for controlling a driving robot including a driving motor unit, Step of acquiring surrounding environment information and movement commands; A step of recognizing a road structure based on the above surrounding environment information and generating road structure information; A step of selecting a movement target point and action of the robot based on the above-mentioned path structure information and the above-mentioned movement command; A step of inputting the above-mentioned surrounding environment information and selection information including the movement target point and action of the robot into a pre-learned neural network model to determine the driving action of the robot; and, A control method for a driving robot, characterized in that it includes a step of controlling the driving motor unit so that the robot moves based on the determined driving motion.

Images