Navigation control method and system for embodied robots and embodied robot

Abstract

This application relates to a navigation control method, system, and embodied robot. It includes: during the process of traveling to a target location corresponding to a task to be performed according to a global reference path, continuously performing local path planning; for each local path planning process, generating at least three candidate paths based on the current position of the embodied robot and the global reference path; if the distance between each candidate path and an obstacle is less than a first preset distance, determining the corresponding travel evaluation coefficient for each candidate path based on evaluation reference data selected from the at least three candidate paths; selecting the target path for this planning from the selected candidate paths based on the travel evaluation coefficients; controlling the embodied robot to travel according to the planned target path; and continuing the local path planning operation until the embodied robot reaches the target location. This method can avoid collisions between the embodied robot and obstacles.

Description

Technical Field

[0001] This application relates to the field of robot path planning technology, and in particular to a navigation control method, system, and embodied robot. Background Technology

[0002] With the rapid development of artificial intelligence and robotics, the application scenarios of embodied robots are becoming increasingly widespread. Wheeled embodied robots, due to their human-like interactive form, are gaining popularity in home, industrial and other scenarios.

[0003] However, in practical applications, when faced with complex ground environments, embodied robots may collide with obstacles due to insufficient reaction speed and decision-making ability. Traditional navigation and control methods for embodied robots are insufficient to find the optimal path for them, which urgently needs to be addressed. Summary of the Invention

[0004] Therefore, it is necessary to provide a navigation control method, system, and embodied robot to address the aforementioned technical problems, which can improve the navigation control flexibility and environmental adaptability of the embodied robot, thereby avoiding collisions between the embodied robot and obstacles.

[0005] In a first aspect, this application provides a navigation control method for an embodied robot, comprising:

[0006] During the process of moving to the target location corresponding to the task to be performed according to the global reference path, the embodied robot continuously performs local path planning, and for each local path planning process, generates at least three candidate paths based on the current position of the embodied robot and the global reference path; wherein, the global reference path is determined based on the initial movement position of the embodied robot and the target location corresponding to the task to be performed.

[0007] If the distance between at least three candidate paths and the obstacle is less than the first preset distance, then the travel evaluation coefficient corresponding to the candidate path is determined based on the evaluation reference data corresponding to each candidate path selected from the at least three candidate paths.

[0008] Based on the travel evaluation coefficients corresponding to each candidate path, the target path for this planning is selected from the selected candidate paths;

[0009] Control the embodied robot to move according to the target path planned in this project;

[0010] Continue executing the local path planning operation until the embodied robot reaches the target location.

[0011] In one embodiment, at least three candidate paths are generated based on the current position of the embodied robot and the global reference path, including:

[0012] Based on different prediction durations, at least one reference path point is selected from the global reference path; the path distance between the reference path point corresponding to different prediction durations and the current position of the embodied robot is different, and the path distance is positively correlated with the prediction duration.

[0013] For any reference path point, at least two candidate path endpoints corresponding to the reference path point are determined from the map data based on the lateral travel offset constraints of the embodied robot.

[0014] Starting from the current position of the embodied robot, at least three candidate paths are generated; among them, one candidate path at least partially overlaps with the global reference path, or the distance between it and the global reference path is less than a second preset distance; the remaining at least two candidate paths correspond to the end points of the candidate paths, and the end points of the candidate paths are the trajectory endpoints of the corresponding candidate paths.

[0015] In one embodiment, the lateral travel offset constraint includes a lateral distance offset constraint; correspondingly, based on the lateral travel offset constraint corresponding to the embodied robot, at least two candidate path endpoints corresponding to the reference path point are determined from the map data, including:

[0016] Each path point in the map data that is on the same horizontal reference line as the reference path point and meets the horizontal distance offset constraint is taken as at least two candidate path endpoints corresponding to the reference path point; wherein, the direction of the horizontal reference line is perpendicular to the tangent of the reference path point on the global reference path.

[0017] In one embodiment, the lateral travel offset constraint further includes a lateral motion offset constraint; correspondingly, the method further includes:

[0018] Based on the lateral motion offset constraint, target lateral configuration data is set for the end points of each candidate path;

[0019] Based on longitudinal motion offset constraints, target longitudinal configuration data is set for the end points of each candidate path.

[0020] The target lateral configuration data includes lateral velocity of 0, lateral acceleration of 0, and vertical distance between the candidate path endpoint and the global reference path.

[0021] Accordingly, based on the longitudinal motion offset constraint, target longitudinal configuration data is set for the end points of each candidate path, including:

[0022] With the desired longitudinal speed as the center, longitudinal speed is sampled within a preset speed range, and based on the sampling results, longitudinal speed is set for the end point of each candidate path.

[0023] Set the longitudinal acceleration corresponding to the end point of each candidate path to 0.

[0024] In one embodiment, starting from the current position of the embodied robot, at least three candidate paths are generated, including:

[0025] For any candidate path, starting from the current position of the embodied robot and ending at the end point of the candidate path, a corresponding candidate path is generated based on the initial configuration data corresponding to the current position and the target configuration data corresponding to the end point of the candidate path.

[0026] The target configuration data includes target lateral configuration data and target longitudinal configuration data; the initial configuration data includes initial lateral configuration data and initial longitudinal configuration data; the initial lateral configuration data includes the vertical distance between the current position and the global reference path, the lateral velocity and lateral acceleration of the embodied robot at the current position; the initial longitudinal configuration data includes the longitudinal velocity and longitudinal acceleration of the embodied robot at the current position.

[0027] In one embodiment, based on the initial configuration data corresponding to the current location and the target configuration data corresponding to the candidate path endpoint, a corresponding candidate path is generated, including:

[0028] Based on the initial configuration data corresponding to the current location and the target configuration data corresponding to the candidate path endpoint, determine the trajectory description data corresponding to the candidate path;

[0029] For the predicted duration corresponding to the candidate path, the trajectory description data is discretized according to the preset time sampling interval to obtain the lateral configuration data and the corresponding longitudinal configuration data at different times when the embodied robot moves on the candidate path. The lateral configuration data and the longitudinal configuration data at the corresponding time are then integrated to generate the corresponding candidate path.

[0030] In one embodiment, after generating at least three candidate paths based on the embodied robot's current position and a global reference path, the method further includes:

[0031] If the distance between each candidate path and the obstacle exceeds the first preset distance, then the candidate path that at least partially overlaps with the global reference path, or whose distance from the global reference path is less than the second preset distance, will be the target path for this planning.

[0032] In one embodiment, after at least three candidate paths are all less than a first preset distance from the obstacle, the method further includes:

[0033] Determine whether at least three candidate paths all intersect with the obstacle.

[0034] If some candidate paths intersect with obstacles, then the candidate paths with intersections are eliminated, and the remaining candidate paths are selected and their corresponding evaluation reference data is determined.

[0035] If at least three candidate paths intersect with the obstacle, or if at least three candidate paths do not intersect with the obstacle, then all candidate paths are selected and the corresponding evaluation reference data is determined.

[0036] In one embodiment, the evaluation reference data includes target configuration data and obstacle association data; accordingly, based on the evaluation reference data corresponding to each candidate path, the travel evaluation coefficient corresponding to the respective candidate path is determined, including:

[0037] For any candidate path, determine the motion evaluation coefficient corresponding to the candidate path based on the target configuration data corresponding to the end point of the candidate path.

[0038] Based on the obstacle association data corresponding to the candidate path, determine the obstacle evaluation coefficient corresponding to the candidate path;

[0039] Based on the motion evaluation coefficient and obstacle evaluation coefficient, the travel evaluation coefficient corresponding to the candidate path is determined.

[0040] In one embodiment, the motion evaluation coefficient includes a lateral motion evaluation coefficient and a longitudinal motion evaluation coefficient; determining the travel evaluation coefficient corresponding to the candidate path based on the motion evaluation coefficient and the obstacle evaluation coefficient includes:

[0041] The lateral motion evaluation coefficient, longitudinal motion evaluation coefficient, and obstacle evaluation coefficient are weighted and summed, and the sum is used as the travel evaluation coefficient corresponding to the candidate path.

[0042] Among them, the weights of the lateral motion evaluation coefficient and the obstacle evaluation coefficient are both greater than the weight of the longitudinal motion evaluation coefficient.

[0043] In one embodiment, for any candidate path, the obstacle association data corresponding to the candidate path is related to at least one of the following: the distance between each path point and the obstacle in the candidate path, the type of obstacle, or the method of collecting the obstacle.

[0044] Target configuration data includes horizontal configuration data and vertical configuration data; the horizontal motion evaluation coefficient is related to the horizontal configuration data, and the vertical motion evaluation coefficient is related to the vertical configuration data.

[0045] The lateral configuration data includes at least one of the following: the lateral velocity, lateral acceleration, and the change of lateral acceleration over time at the candidate path endpoint; and the vertical distance between the candidate path endpoint and the global reference path. The longitudinal configuration data includes at least one of the following: the difference between the longitudinal velocity and the longitudinal desired velocity at the candidate path endpoint; the longitudinal acceleration and the change of longitudinal acceleration over time; and the prediction duration.

[0046] In one embodiment, if the android collides with an obstacle during its movement, the method further includes:

[0047] Add obstacle information to the map data to update the map data;

[0048] The robot's current position is used as the initial motion position. Based on the initial motion position and the target position corresponding to the task to be performed, a new global reference path is determined, and the robot is controlled to move along the new global reference path. During the movement, local path planning is continuously performed.

[0049] Secondly, this application also provides a navigation and control system for an embodied robot, comprising:

[0050] The path generation module is used to continuously perform local path planning for the embodied robot during its journey to the target location corresponding to the task to be performed based on the global reference path. For each local path planning process, at least three candidate paths are generated based on the current position of the embodied robot and the global reference path. The global reference path is determined based on the initial movement position of the embodied robot and the target location corresponding to the task to be performed.

[0051] The coefficient determination module is used to determine the travel evaluation coefficient of the corresponding candidate path based on the evaluation reference data of each candidate path selected from the at least three candidate paths if the distance between the at least three candidate paths and the obstacle is less than the first preset distance.

[0052] The path selection module is used to select the target path for this planning based on the travel evaluation coefficients corresponding to each candidate path.

[0053] The robot control module is used to control the embodied robot to move according to the planned target path;

[0054] The path planning module is used to continue performing local path planning operations until the embodied robot reaches the target location.

[0055] Thirdly, this application also provides an embodied robot equipped with a computer device, the computer device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps:

[0056] During the process of moving to the target location corresponding to the task to be performed according to the global reference path, the embodied robot continuously performs local path planning, and for each local path planning process, generates at least three candidate paths based on the current position of the embodied robot and the global reference path; wherein, the global reference path is determined based on the initial movement position of the embodied robot and the target location corresponding to the task to be performed.

[0057] If the distance between at least three candidate paths and the obstacle is less than the first preset distance, then the travel evaluation coefficient corresponding to the candidate path is determined based on the evaluation reference data corresponding to each candidate path selected from the at least three candidate paths.

[0058] Based on the travel evaluation coefficients corresponding to each candidate path, the target path for this planning is selected from the selected candidate paths;

[0059] Control the embodied robot to move according to the target path planned in this project;

[0060] Continue executing the local path planning operation until the embodied robot reaches the target location.

[0061] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0062] During the process of moving to the target location corresponding to the task to be performed according to the global reference path, the embodied robot continuously performs local path planning, and for each local path planning process, generates at least three candidate paths based on the current position of the embodied robot and the global reference path; wherein, the global reference path is determined based on the initial movement position of the embodied robot and the target location corresponding to the task to be performed.

[0063] If the distance between at least three candidate paths and the obstacle is less than the first preset distance, then the travel evaluation coefficient corresponding to the candidate path is determined based on the evaluation reference data corresponding to each candidate path selected from the at least three candidate paths.

[0064] Based on the travel evaluation coefficients corresponding to each candidate path, the target path for this planning is selected from the selected candidate paths;

[0065] Control the embodied robot to move according to the target path planned in this project;

[0066] Continue executing the local path planning operation until the embodied robot reaches the target location.

[0067] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0068] During the process of moving to the target location corresponding to the task to be performed according to the global reference path, the embodied robot continuously performs local path planning, and for each local path planning process, generates at least three candidate paths based on the current position of the embodied robot and the global reference path; wherein, the global reference path is determined based on the initial movement position of the embodied robot and the target location corresponding to the task to be performed.

[0069] If the distance between at least three candidate paths and the obstacle is less than the first preset distance, then the travel evaluation coefficient corresponding to the candidate path is determined based on the evaluation reference data corresponding to each candidate path selected from the at least three candidate paths.

[0070] Based on the travel evaluation coefficients corresponding to each candidate path, the target path for this planning is selected from the selected candidate paths;

[0071] Control the embodied robot to move according to the target path planned in this project;

[0072] Continue executing the local path planning operation until the embodied robot reaches the target location.

[0073] The aforementioned navigation control method, system, and embodied robot, during their journey to the target location corresponding to the task to be performed according to the global reference path, continuously perform local path planning. For each local path planning process, based on the embodied robot's current position and the global reference path, at least three candidate paths are generated. When the distance between each candidate path and an obstacle is less than a first preset distance, a travel evaluation coefficient corresponding to the candidate path selected from the at least three candidate paths is determined based on the evaluation reference data. Then, based on the travel evaluation coefficients of each candidate path, the target path for this planning is selected from the selected candidate paths. The embodied robot is controlled to travel along the planned target path, and the local path planning operation is continuously executed during the journey until the embodied robot reaches the target location. In this process, on the one hand, continuous local path planning during the robot's movement enables it to respond in real time to new obstacles that appear after the map data is constructed, some dynamic obstacles, or obstacles that are difficult to avoid based on the global reference path, effectively avoiding collisions and improving the robot's safety in complex environments. On the other hand, by calculating the movement evaluation coefficient based on the evaluation reference data of each candidate path and selecting the target path accordingly, the safety, rationality, and smoothness of each candidate path can be comprehensively considered, so that the final selected target path not only meets the obstacle avoidance requirements but also adapts to the robot's stable movement characteristics. Attached Figure Description

[0074] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0075] Figure 1 This is an application environment diagram of the navigation and control method for an embodied robot in one embodiment;

[0076] Figure 2 This is a flowchart illustrating the navigation and control method of an embodied robot in one embodiment;

[0077] Figure 3A This is a flowchart illustrating the candidate path generation steps in one embodiment;

[0078] Figure 3B This is a schematic diagram showing the location of the candidate path endpoint in one embodiment;

[0079] Figure 4 This is a flowchart illustrating the candidate path generation step in another embodiment;

[0080] Figure 5 This is a flowchart illustrating the steps for determining the travel evaluation coefficient in one embodiment;

[0081] Figure 6 This is a structural block diagram of the navigation control device for an embodied robot in one embodiment;

[0082] Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0083] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0084] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0085] The navigation and control method for the android provided in this application embodiment can be applied to, for example... Figure 1 The application scenario shown includes an embodied robot 102 and a target environment, with the embodied robot 102 situated within the target environment. The embodied robot possesses perception, movement, and interaction capabilities, and can interact with its environment in real time. It can capture information about its surroundings through sensory organs such as cameras, LiDAR, or tactile sensors installed on the embodied robot. The embodied robot can be, but is not limited to, humanoid robots, cleaning robots, and companion robots. The cleaning robot can be, but is not limited to, robotic vacuum cleaners, robotic mops, and robotic vacuum and mop combos. The target environment can be any environment suitable for the application of the embodied robot, including, but is not limited to, home environments, industrial manufacturing environments, medical environments, warehousing and logistics environments, agricultural production environments, public service environments, educational and research environments, or entertainment performance environments.

[0086] In traditional technologies, the navigation control logic of embodied robots is usually generated before the embodied robot moves. Based on the map data of the target environment stored by the embodied robot, and according to the target location corresponding to the task to be performed by the embodied robot and the current location of the embodied robot, a global reference path is generated from the current location to the target location, and the embodied robot is controlled to move according to the global reference path.

[0087] However, since the map data is pre-built, the global reference path planned using the above path planning method is insufficient to avoid new obstacles that appear after the map data is built, or dynamic obstacles. This can easily lead to collisions between the embodied robot and obstacles, thus affecting task execution efficiency and completion rate. Therefore, there is an urgent need to provide a navigation and obstacle avoidance method that can perceive obstacle information in real time and dynamically update the global reference path during the embodied robot's movement. This method would ensure that when obstacles are updated or moved, a smooth, safe, and effective obstacle avoidance path and speed control can be quickly output, minimizing collisions between the embodied robot and obstacles and ensuring that the embodied robot can complete tasks continuously, stably, and efficiently.

[0088] In one exemplary embodiment, such as Figure 2 As shown, a navigation control method for an embodied robot is provided, which is applied to... Figure 1 Taking the embodied robot 102 as an example, the following steps are included:

[0089] S210, during the process of traveling to the target location corresponding to the task to be performed according to the global reference path, the embodied robot continuously performs local path planning, and for each local path planning process, generates at least three candidate paths based on the current position of the embodied robot and the global reference path.

[0090] The global reference path is determined based on the embodied robot's initial position and the target position corresponding to the task to be performed. It guides the embodied robot's overall direction of movement and route planning for completing the current task. The initial position can be understood as the embodied robot's location before receiving the task, such as its base station location, or its location after completing the previous task. The target position corresponds to the corresponding task; for example, when the task is laundry, the target position is the laundry area; when the task is retrieval, the target position is the location of the item to be retrieved.

[0091] Optionally, the global reference path can be planned based on common path planning algorithms. For example, taking the A* algorithm as an example, the process of global reference path planning is briefly explained: The embodied robot, using the constructed map data (grid map) as its starting point and the target position as its ending point, traverses each path point in the map data to determine the actual travel cost between each path point and the initial travel position, as well as the estimated travel cost between each path point and the target position. The sum of the actual and estimated travel costs for each path point is taken as the target cost for that path point. Subsequently, based on the target costs of each path point, the path points with smaller target costs (e.g., the smallest) are selected sequentially for path extension, forming multiple candidate paths from the initial travel position to the target position. The candidate path with the smallest total cost is then selected as the global reference path to guide the overall direction and route of the embodied robot in this task.

[0092] In this embodiment, after generating the global reference path, the embodied robot is controlled to start from its initial position and move along the global reference path, while continuously performing real-time local path planning during the movement. For example, using the current position of the embodied robot at different times during its movement as the starting point and the global reference path as a reference, local path planning is performed to avoid collisions caused by suddenly appearing dynamic obstacles.

[0093] To facilitate understanding by those skilled in the art, the embodiments of this application use a single local path planning process as an example for description.

[0094] In one alternative implementation, the embodied robot can receive information about obstacles detected by sensors in real time during its movement, and determine the current position of different obstacles based on the obstacle information. It then converts the obstacle information into two-dimensional positional information, projects it onto map data, and, based on the projected map data, plans at least three candidate paths—a left-side avoidance path, a middle straight path, and a right-side avoidance path—within the safe and passable areas on both sides and in the middle of the global reference path.

[0095] The sensors can include any sensor devices used to perceive the external environment, such as cameras, lidar, and position sensitive detectors (PSDs) mounted on the embodied robot. The safe, passable area can be influenced by the embodied robot's size, its mobility, and the size of obstacles. The specific determination method can be based on human experience or through extensive experimentation; this application does not impose any limitations on this.

[0096] Understandably, since the global reference path is a path used to guide the overall direction of movement of the embodied robot, all three candidate paths mentioned above are based on avoiding obstacles and continuing to move towards the target position. After completing obstacle avoidance, the robot can return to the global reference path through the selection of candidate paths and guide the embodied robot to the target position.

[0097] In another optional implementation, the real-time received obstacle information and global reference path can be input into a pre-trained candidate path generation model to generate at least three candidate paths. The candidate path generation model can be built based on common neural networks, which will not be elaborated upon here. Furthermore, during the training of the candidate path generation model, sample obstacle information and sample reference paths can be input into the model to predict candidate paths. Based on the difference between the predicted candidate paths and the pre-determined candidate path labels, the model parameters of the candidate path generation model are adjusted to improve the candidate path prediction accuracy.

[0098] S220, if the distance between at least three candidate paths and the obstacle is less than the first preset distance, then the travel evaluation coefficient corresponding to the corresponding candidate path is determined based on the evaluation reference data corresponding to each candidate path selected from the at least three candidate paths.

[0099] Among them, the distance between at least three candidate paths and obstacles is the distance between the candidate path and the nearest obstacle. The first preset distance can be understood as the safe distance between the embodied robot and the nearest obstacle during the robot's movement. It may be affected by the size of the embodied robot itself, or it may be determined based on human experience, or it may be determined through a large number of experiments. For example, the first preset distance is 1m. This application does not make any limitation on this.

[0100] The evaluation reference data refers to the relevant parameters that affect the final result of the travel evaluation coefficient during the determination of the travel evaluation coefficient of the candidate path. For example, the evaluation reference data may include at least one of the following: the offset distance between the corresponding candidate path and the global reference path, the presence of obstacles on the corresponding candidate path, and the driving state of the embodied robot on the corresponding candidate path, as described in the following embodiments.

[0101] Since a candidate path corresponds to a path trajectory, and the nearest obstacle corresponds to a location point, in some embodiments, during the process of determining the distance between each candidate path and an obstacle, the distance between the obstacle and each path point in the corresponding candidate path can be determined, and then the distance between the corresponding candidate path and the obstacle can be determined based on the distance between each path point and the obstacle. For example, the average distance between each path point and the obstacle can be used as the distance between the corresponding candidate path and the obstacle. Another example is using the smaller distance among the distances between each path point and the obstacle, for example, the minimum distance, as the distance between the corresponding candidate path and the obstacle.

[0102] In one optional implementation, if the distance between each candidate path and the nearest obstacle is less than a first preset distance, then the travel evaluation coefficient corresponding to the candidate path is determined according to the evaluation reference data corresponding to each candidate path.

[0103] In another alternative implementation, before determining the relationship between the distance between each candidate path and the nearest obstacle and the first preset distance, a trajectory rationality and stability check can be performed on each candidate path to eliminate invalid candidate paths with unreasonable or unstable trajectories. Furthermore, if the distance between each candidate path selected from at least three candidate paths and the nearest obstacle is less than the first preset distance, a travel evaluation coefficient corresponding to each candidate path is determined based on the evaluation reference data corresponding to each candidate path.

[0104] Optionally, the logic for performing trajectory rationality checks on each candidate path can be as follows: determine the distance between each path point in each candidate path; if the distance between adjacent path points does not exceed the distance threshold, then the trajectory rationality check of the corresponding candidate path is passed; otherwise, it is not passed.

[0105] Optionally, the logic for performing a stability check on each candidate path can be as follows: determine the motion state of the embodied robot at different times under each candidate path, such as motion speed and / or acceleration, and determine whether the corresponding motion state conforms to the embodied robot's own dynamic parameters. If it does, the stability check of the corresponding candidate path is passed; otherwise, it is not passed. Alternatively, it can be based on the changes in the motion state of the embodied robot at adjacent times under each path. If the changes in the motion state at adjacent times exceed a preset stability threshold, the stability check of the corresponding candidate path is passed; otherwise, it is not passed.

[0106] To improve the efficiency of target path selection, in some embodiments, after determining that the distance between each candidate path and the nearest obstacle is less than a first preset distance, it can be further determined whether at least three candidate paths intersect with obstacles. If some candidate paths intersect with obstacles, that is, if an obstacle falls on a candidate path, then the intersecting candidate paths are eliminated, and the remaining candidate paths are selected and their corresponding evaluation reference data is determined. In this way, invalid candidate paths can be eliminated, reducing the generation of subsequent evaluation reference data, thereby improving the efficiency of target path selection.

[0107] If at least three candidate paths intersect with the obstacle, or if at least three candidate paths do not intersect with the obstacle, then all candidate paths are selected and corresponding evaluation reference data is determined. Based on the evaluation reference data corresponding to each candidate path, the target path is selected from each candidate path.

[0108] In one alternative implementation, for the selected candidate paths, a process can be performed to determine the travel evaluation coefficient corresponding to each candidate path based on the evaluation reference data corresponding to each candidate path.

[0109] Taking the determination of the travel evaluation coefficient corresponding to a candidate path as an example, the following explanation is provided. A formula for determining the travel evaluation coefficient can be predetermined, and the evaluation reference data corresponding to the candidate path can be substituted into the formula to calculate the travel evaluation coefficient corresponding to the candidate path. Alternatively, the state information of the embodied robot at different path points on the candidate path can be predicted, and the travel evaluation parameters for each path point can be determined based on the state information. Furthermore, the sum or weighted sum of the travel evaluation parameters corresponding to each path point on the candidate path is used as the travel evaluation coefficient corresponding to the candidate path. The weights corresponding to different path points can be determined based on human experience or through extensive experimentation; this application does not impose any limitations on this.

[0110] To improve the efficiency of target path selection, in some embodiments, if the distance between each candidate path and the obstacle exceeds a first preset distance, that is, each candidate path can achieve obstacle avoidance, then the candidate path that at least partially overlaps with the global reference path, or whose distance from the global reference path is less than a second preset distance, is directly selected as the target path for this planning. The second preset distance is much smaller than the first preset distance, meaning that the candidate path that is close to the global reference path is selected as the target path for this planning. The second preset distance can be determined based on human experience or through extensive experimentation; this application does not impose any limitations on this.

[0111] In this process, when all candidate paths can achieve obstacle avoidance, the target path can be quickly located from the candidate paths, improving the efficiency of target path selection and the smoothness of the robot's navigation control.

[0112] S230, Based on the travel evaluation coefficients corresponding to each candidate path, select the target path for this planning from the selected candidate paths.

[0113] For example, the relative magnitudes of the travel evaluation coefficients corresponding to each candidate path can be determined, and the candidate path with the smaller travel evaluation coefficient, for example, the smallest evaluation coefficient, among the selected candidate paths can be taken as the target path for this planning.

[0114] S240 controls the embodied robot to move according to the planned target path.

[0115] For example, after selecting a target path, a travel control command can be generated based on the target path and sent to the embodied robot to instruct the embodied robot to travel based on the planned target path.

[0116] S250 continues to perform local path planning until the embodied robot reaches the target location.

[0117] For example, while controlling the embodied robot to move along the planned target path, the operation of local path planning continues to be executed in real time until the embodied robot reaches the target position. The operation process of local path planning has been described in the above process and will not be repeated here.

[0118] In the aforementioned navigation control method for the embodied robot, during the process of traveling to the target location corresponding to the task to be performed according to the global reference path, local path planning is continuously performed. For each local path planning process, at least three candidate paths are generated based on the current position of the embodied robot and the global reference path. When the distance between each candidate path and the obstacle is less than a first preset distance, the travel evaluation coefficient corresponding to the candidate path selected from the at least three candidate paths is determined based on the evaluation reference data. Then, based on the travel evaluation coefficients corresponding to each candidate path, the target path for this planning is selected from the selected candidate paths. The embodied robot is controlled to travel according to the target path for this planning, and the local path planning operation is continuously executed during the travel process until the embodied robot reaches the target location. In this process, on the one hand, continuous local path planning during the robot's movement enables it to respond in real time to new obstacles that appear after the map data is constructed, some dynamic obstacles, or obstacles that are difficult to avoid based on the global reference path, effectively avoiding collisions and improving the robot's safety in complex environments. On the other hand, by calculating the movement evaluation coefficient based on the evaluation reference data of each candidate path and selecting the target path accordingly, the safety, rationality, and smoothness of each candidate path can be comprehensively considered, so that the final selected target path not only meets the obstacle avoidance requirements but also adapts to the robot's stable movement characteristics.

[0119] Because dynamic obstacles typically change position rapidly and irregularly, although the aforementioned obstacle avoidance methods can avoid obstacles as much as possible, collisions between the android and the obstacle are still unavoidable in some special cases. In these cases, this application also provides subsequent control logic after a collision between the android and the obstacle: if the android collides with an obstacle during its movement, the information of the obstacle is added to the map data to update the map data; the current position of the android is used as the initial movement position; based on the initial movement position and the target position corresponding to the task to be performed, a new global reference path is determined, and the android is controlled to move along the new global reference path, while continuously performing local path planning during the movement.

[0120] The obstacle information can be obtained from the collision detection device of the embodied robot, and may include information such as the location and type of the obstacle.

[0121] For example, the location information of obstacles can be added to the map data to update the map data, and the corresponding locations can be marked as obstacles in the map data. The type information of obstacles can also be marked.

[0122] This process serves two purposes: firstly, it avoids interrupting or halting the robot's movement due to a single collision, ensuring the continuity of the robot's task execution; secondly, it adds obstacle information to the map data, allowing the robot to consider the obstacle during the next path planning process, thus preventing further collisions during subsequent movement.

[0123] Based on the technical solutions of the above embodiments, this application also provides an optional embodiment. In this optional embodiment, the process of generating at least three candidate paths based on the current position of the embodied robot and the global reference path is refined.

[0124] See Figure 3A Based on the current position of the embodied robot and the global reference path, at least three candidate paths are generated, including:

[0125] S310, select at least one reference path point from the global reference path based on different prediction durations.

[0126] The prediction duration can be understood as the length of time it takes for the embodied robot to predict its future trajectory. It can be visually represented as the distance between a reference path point and the robot's current location. It is understood that one prediction duration corresponds to at least one reference path point (e.g., one prediction duration corresponds to one reference path point).

[0127] The path distance between the reference path point and the current position of the embodied robot varies depending on the prediction duration, and the path distance is positively correlated with the prediction duration.

[0128] It is understandable that if the prediction duration is too short, the generated candidate paths may not be able to adequately handle unexpected situations; if the prediction duration is too long, the reliability of the generated candidate paths will decrease, and computational resources will be significantly wasted. Therefore, in this embodiment, the range of prediction duration can be set based on human experience or determined based on a large number of experiments, and this application does not impose any limitations on it. For example, the prediction duration can be 2 seconds or 5 seconds.

[0129] In one optional implementation, the current position of the embodied robot is on a global reference path, and the corresponding reference path point can be selected in the following way: starting from the current position of the embodied robot, it moves towards the target position along the global reference path, and the expected travel distance is determined according to the current travel speed of the embodied robot and different prediction durations; a point on the global reference path whose path length between it and the current position is equal to the expected travel distance is selected as the reference path point corresponding to the corresponding prediction duration.

[0130] In another alternative implementation, the current position of the embodied robot is not on the global reference path. The corresponding reference path point can be selected in the following way: find the projection point corresponding to the current position of the embodied robot on the global reference path (for example, the point closest to the current position on the global reference path), take the projection point as the starting point, and move towards the target position along the global reference path. Determine the expected travel distance based on the current travel speed of the embodied robot and different prediction durations; select a point on the global reference path whose path length from the current position is equal to the expected travel distance, and use it as the reference path point corresponding to the corresponding prediction duration.

[0131] S320: For any reference path point, determine at least two candidate path endpoints corresponding to the reference path point from the map data based on the lateral travel offset constraint corresponding to the embodied robot.

[0132] The lateral travel offset constraint can be understood as constraining the movement state of the embodied robot in the direction perpendicular to the global reference path during its movement. For example, it can be used to constrain the lateral offset range of the embodied robot in the direction perpendicular to the extension of the global reference path (corresponding to the lateral distance offset constraint below, which can be used to determine the safe passable area in the above embodiments); it can also be used to constrain the range of motion state of the embodied robot in the direction perpendicular to the global reference path (corresponding to the lateral motion offset constraint below).

[0133] The aforementioned lateral offset range can be affected by the robot's own volume, its own mobility, and the size of the obstacle. The specific determination method can be based on human experience or through a large number of experiments. This application does not impose any limitations on this.

[0134] In one optional implementation, the lateral travel offset constraint includes a lateral distance offset constraint; correspondingly, each path point in the map data that is on the same lateral reference line as the reference path point and meets the lateral distance offset constraint can be used as at least two candidate path end points corresponding to the reference path point; wherein, the direction of the lateral reference line is perpendicular to the tangent of the reference path point on the global reference path.

[0135] Typically, lateral distance offset constraints can be defined by setting upper limits for allowable offsets to the left and right sides, centered on the global reference path, thus constructing the range of the lateral distance offset constraint. These constraints can be set based on the volume of the embodied robot or based on human experience; this application does not impose any limitations in this regard. For example, the lateral distance offset constraint can be a 1-meter offset constraint to the left and right of the embodied robot.

[0136] See Figure 3BThe diagram shows the candidate path endpoint positions. l1 is the global reference path; l2 is the tangent line of the reference path point on the global reference path; l3 is perpendicular to l2 and is the lateral reference line; the distance between l4 and l5 is the lateral distance offset constraint corresponding to the embodied robot; point O is the current position of the embodied robot; and point A is a reference path point. Correspondingly, any point on the lateral reference line l3 and located between l4 and l5 can be considered a candidate path endpoint. Points A1, A2, A3, and A4 are examples of candidate path endpoints.

[0137] Optionally, the distance between the endpoints of different candidate paths can be unrestricted. To ensure a uniform distribution of candidate paths, the endpoints can be the same. The number of endpoints is related to the computing power of the embodied robot. The stronger the computing power, the more endpoints can be generated. When the computing power of the embodied robot is limited, the number of endpoints can be appropriately reduced to improve the accuracy of the candidate paths.

[0138] In another alternative implementation, the lateral travel offset constraint also includes a lateral motion offset constraint. Accordingly, target lateral configuration data can be set for each candidate path endpoint based on the lateral motion offset constraint; and target longitudinal configuration data can be set for each candidate path endpoint based on the longitudinal motion offset constraint.

[0139] The target configuration data is set based on the embodied robot's own motion capabilities (e.g., the maximum speed it can travel at, or the maximum sharp turn it can make), and is used to limit the number of motion states of the embodied robot. The target configuration data can include target lateral configuration data and target longitudinal configuration data. Target lateral configuration data is used to limit the embodied robot's motion states on the lateral reference line; target longitudinal lateral configuration data is used to limit the embodied robot's motion states on the longitudinal reference line. The target lateral configuration data is set with reference to lateral motion offset constraints; the target longitudinal configuration data is set with reference to longitudinal motion offset constraints.

[0140] Among them, the lateral motion offset constraint can be used to constrain the range of motion states of the embodied robot in the direction perpendicular to the global reference path. It can include lateral velocity and / or lateral acceleration. In order to enable the embodied robot to move as smoothly as possible, the lateral motion offset constraint usually includes lateral velocity and lateral acceleration.

[0141] In order to enable the embodied robot to travel along the global reference path as much as possible, in this embodiment, the target lateral configuration data includes a lateral velocity of 0, a lateral acceleration of 0, and the vertical distance between the candidate path end point and the global reference path.

[0142] Among them, the longitudinal motion offset constraint can be used to constrain the range of motion of the embodied robot in the tangential direction of the global reference path, and can include longitudinal velocity and / or longitudinal acceleration. In order to enable the embodied robot to move as smoothly as possible, the longitudinal motion offset constraint usually includes longitudinal velocity and longitudinal acceleration.

[0143] To further improve the trajectory smoothness of the candidate paths and the stability of the robot's movement, in this embodiment, longitudinal velocity sampling can be performed within a preset speed range with the desired longitudinal velocity as the center, and longitudinal velocity can be set for the end point of each candidate path based on the sampling results; the longitudinal acceleration corresponding to the end point of each candidate path can be set to 0.

[0144] The longitudinal desired speed and the preset speed range can both be determined based on human experience or through extensive experimentation; this application does not impose any limitations on either. For example, the longitudinal desired speed can be 0.3 m / s or 0.5 m / s; the preset speed range can be ±0.2 m / s.

[0145] S330 generates at least three candidate paths starting from the current position of the embodied robot.

[0146] Since reference path points can also be used as candidate path endpoints, among the generated candidate paths, one candidate path at least partially overlaps with the global reference path, or the distance between it and the global reference path is less than the second preset distance (empirical value); the remaining at least two candidate paths correspond to the candidate path endpoints determined after lateral travel offset constraints, and the candidate path endpoints are the trajectory endpoints of the corresponding candidate paths.

[0147] For example, taking the current position of the embodied robot as the starting point and the end point of each candidate path as the ending point, the continuous smooth trajectory between the starting point and the ending point is used as the candidate path corresponding to the end point of the corresponding candidate path. The target configuration data corresponding to the end point of the corresponding candidate path is used as a reference for the motion state of the embodied robot, generating at least three candidate paths. For example, the continuous smooth trajectory from the starting point to the ending point can be generated by curve fitting. The specific implementation method is described in detail in the following embodiments.

[0148] The number of candidate paths is determined by a combination of three factors: the lateral velocity, lateral acceleration, and longitudinal acceleration corresponding to each candidate path are all set to a fixed value of 0. For example, if the lateral distance constraint is [-1m, 1m] and the lateral distance sampling density is at intervals of 0.25m, then 9 sets of lateral distance sampling results can be obtained, representing 9 candidate path endpoints. If the longitudinal expected velocity is 0.4m / s and the longitudinal velocity sampling results are 0.3m / s and 0.5m / s, then 2 sets of longitudinal velocity sampling results can be obtained. If the prediction duration includes 2s and 5s, then 2 sets of prediction duration sampling results can be obtained. The final number of candidate paths is the product of the number of sets of lateral distance sampling results, the number of longitudinal velocity sampling results, and the number of prediction duration sampling results, i.e., 9*2*2 candidate paths.

[0149] In the above embodiments, selecting reference path points based on different prediction durations can take into account the planning needs of both near and far paths when planning candidate paths; selecting candidate path end points based on lateral travel distance offset constraints can smooth the offset between candidate paths and global reference paths when planning candidate paths; and based on lateral travel motion offset constraints and longitudinal motion offset constraints, the motion parameter limitations of the embodied robot can be taken into account when planning candidate paths. The whole process can ensure the rationality of each candidate path while improving the probability of candidate paths.

[0150] Based on the technical solutions of the above embodiments, this application also provides an optional embodiment. In this optional embodiment, the process of generating at least three candidate paths starting from the current position of the embodied robot will continue to be described.

[0151] See Figure 4 Starting from the current position of the embodied robot, generate at least three candidate paths, including:

[0152] S410: For any candidate path, starting from the current position of the embodied robot and ending at the end point of the candidate path, generate the corresponding candidate path based on the initial configuration data corresponding to the current position and the target configuration data corresponding to the end point of the candidate path.

[0153] The target configuration data includes target horizontal configuration data and target vertical configuration data. The target configuration data has been described in the above embodiments and will not be repeated here.

[0154] The initial configuration data includes initial lateral configuration data and initial longitudinal configuration data. The initial lateral configuration data includes the vertical distance between the current position and the global reference path, and the lateral velocity and acceleration of the embodied robot at its current position. The initial longitudinal configuration data includes the longitudinal velocity and acceleration of the embodied robot at its current position. Correspondingly, the parameters included in the target configuration data and the initial configuration data can be the same; the difference is that the target configuration data is the configuration data for the endpoints of the candidate path, while the initial configuration data is the configuration data for the current position.

[0155] In one optional implementation, trajectory description data corresponding to a candidate path can be determined based on the initial configuration data corresponding to the current location and the target configuration data corresponding to the end point of the candidate path. For the predicted duration corresponding to the candidate path, the trajectory description data is discretized according to a preset time sampling interval to obtain the lateral configuration data and the corresponding longitudinal configuration data at different times when the embodied robot moves on the candidate path. The lateral configuration data and the longitudinal configuration data at the corresponding times are then integrated to generate the corresponding candidate path.

[0156] The trajectory description data describes the continuous motion and change patterns of the corresponding candidate path within the prediction time, and may include both horizontal and vertical trajectory description data. The preset time sampling interval can be determined based on human experience or through extensive experimentation; this application does not impose any limitations on this.

[0157] The following describes the process of determining the trajectory description data corresponding to any candidate path: In an optional implementation, the trajectory description data corresponding to the candidate path can be calculated based on the initial configuration data and the target configuration data, using a pre-set trajectory description data determination function (including a lateral trajectory description data determination function and a longitudinal trajectory description data determination function). For example, the lateral trajectory description data determination function can be as follows:

[0158] ;

[0159] In the formula, D0 represents the initial lateral configuration data, where d(t0) represents the initial lateral offset, which is the vertical distance between the current position and the global reference path; Indicates the initial lateral velocity; d(t1) represents the initial lateral acceleration; D1 represents the target lateral configuration data, where d(t1) represents the target lateral offset, which is the vertical distance between the candidate path endpoint and the global reference path. Indicates the target's lateral velocity; c represents the target's lateral acceleration. d0 cd1 c d2 c d3 c d4 c d5 These represent the coefficients of each expression.

[0160] For example, the function for determining the longitudinal trajectory description data can be as follows:

[0161] ;

[0162] In the formula, S0 represents the initial longitudinal configuration data, where s(t0) represents the initial longitudinal offset; Indicates the initial longitudinal velocity; S1 represents the initial longitudinal acceleration; S2 represents the target longitudinal configuration data, where, Indicates the longitudinal velocity of the target; This represents the longitudinal acceleration of the target; a0, a1, a2, a3, and a4 represent the coefficients of each term.

[0163] Specifically, by substituting the initial lateral configuration data and the target lateral configuration data into the lateral trajectory description data determination function, the lateral correlation equation is obtained, and solving it yields c. d0 c d1 c d2 c d3 c d4 c d5 The value of is then used to determine the horizontal trajectory description data d(t). and Substituting the initial and target longitudinal configuration data into the longitudinal trajectory description data determination function yields the longitudinal correlation equation. Solving this equation provides the values ​​of a0, a1, a2, a3, and a4, thus determining the longitudinal trajectory description data s(t). and .

[0164] According to the preset time sampling interval, the prediction duration is divided into multiple sampling points, and the time corresponding to each sampling point is substituted into the horizontal trajectory description data d(t). and This allows us to determine the lateral configuration data of the embodied robot at the corresponding sampling point. Similarly, we determine the longitudinal configuration data of the embodied robot at the corresponding sampling point. Furthermore, for any sampling point, we integrate the lateral and longitudinal configuration data to determine the target configuration data for that sampling point. By connecting all sampling points sequentially in chronological order, we can generate a complete candidate path.

[0165] To illustrate with specific numerical examples, for instance, if the prediction duration is 2 seconds (initial time t0 = 0s, target time t1 = 2s), the initial horizontal configuration data will be... Target horizontal configuration data Substituting these values ​​into the function for determining the lateral trajectory description data, we obtain the following lateral correlation equation:

[0166] ;

[0167] Solving the above equation will yield c. d0 c d1 c d2 c d3 c d4 and c d5 The value of is then used to substitute the coefficients into the function for determining the lateral trajectory description data d(t), thereby determining the lateral trajectory description data d(t). and If the preset time sampling interval is 0.5s, and the sampling points are 0.5s, 1s, 1.5s, and 2s respectively, then the data d(t) can be described based on the horizontal trajectory. and Determine the horizontal configuration data corresponding to each sampling point.

[0168] Initial vertical configuration data Target vertical configuration data Substituting these values ​​into the longitudinal trajectory description data determination function, we obtain the following longitudinal correlation equation:

[0169] ;

[0170] Solving the above equation yields the values ​​of a0, a1, a2, a3, and a4. Substituting these coefficients into the longitudinal trajectory description data determination function, the longitudinal trajectory description data s(t) can be determined. and Continuing with the preset sampling interval of 0.5s, and taking sampling points of 0.5s, 1s, 1.5s, and 2s as examples, the data s(t) can be described based on the longitudinal trajectory. and Determine the longitudinal configuration data corresponding to each sampling point.

[0171] By integrating the horizontal and vertical configuration data under each sampling point, the target configuration data under the corresponding sampling point can be obtained, namely (s0, d0), (s0.5, d0.5), (s1, d1), and (s2, d2). By connecting each sampling point in chronological order, a complete candidate path can be generated.

[0172] In the above embodiments, the trajectory description data is divided into lateral trajectory description data and longitudinal trajectory description data. This allows for the use of different solution functions based on the actual constraints in the lateral and longitudinal directions. Specifically, the lateral direction requires simultaneous constraints on the lateral offset, velocity, and acceleration at both the initial and target times, resulting in more complex constraints and therefore more complex functions. The longitudinal direction, however, does not require constraints on the offset at the target time and can be solved using a simpler function. This approach improves the generation speed while ensuring the accuracy of candidate path generation.

[0173] Based on the technical solutions of the above embodiments, this application also provides an optional embodiment. In this optional embodiment, the evaluation reference data includes target configuration data and obstacle association data. Under this premise, the process of determining the travel evaluation coefficient corresponding to the corresponding candidate path based on the evaluation reference data corresponding to each candidate path will be described.

[0174] See Figure 5 Based on the evaluation reference data corresponding to each candidate path, the travel evaluation coefficient corresponding to the respective candidate path is determined, including:

[0175] S510: For any candidate path, determine the motion evaluation coefficient corresponding to the candidate path based on the target configuration data corresponding to the end point of the candidate path.

[0176] The motion evaluation coefficient is used to comprehensively measure the trajectory smoothness and motion rationality of candidate paths. The motion evaluation coefficient includes lateral motion evaluation coefficient and longitudinal motion evaluation coefficient. It can be understood that the lateral motion evaluation coefficient is related to the lateral configuration data, and the longitudinal motion evaluation coefficient is related to the longitudinal configuration data.

[0177] In some embodiments, the lateral configuration data includes at least one of the following: the lateral velocity, lateral acceleration, and the time-varying lateral acceleration of the candidate path endpoint, and the vertical distance between the candidate path endpoint and the global reference path. Taking the inclusion of the above items as an example, the lateral motion evaluation coefficient can be determined based on the following formula:

[0178] ;

[0179] In the formula, denoted by lateral motion evaluation coefficient; T represents the prediction duration corresponding to the candidate path; d represents the lateral offset, which is the vertical distance between the end point of the candidate path and the global reference path. Indicates lateral velocity; Indicates lateral acceleration; This indicates how lateral acceleration changes over time, also known as lateral jerk. These represent the weighting coefficients for lateral offset, lateral velocity, lateral acceleration, and lateral jerk, respectively. Each weighting coefficient can be determined based on human experience or through extensive experimentation; this application does not impose any limitations on this.

[0180] In some embodiments, the longitudinal configuration data includes at least one of the following: the difference between the longitudinal velocity and the desired longitudinal velocity at the candidate path endpoint, longitudinal acceleration, the variation of longitudinal acceleration over time, and the prediction duration. Taking the longitudinal configuration data including the above items as an example, the longitudinal motion evaluation coefficient can be determined based on the following formula:

[0181] ;

[0182] In the formula, Represents the longitudinal motion evaluation coefficient; T represents the prediction duration corresponding to the candidate path; v represents the initial longitudinal velocity; Indicates the longitudinal velocity of the target; Indicates longitudinal acceleration; It represents the change of longitudinal acceleration over time, i.e., longitudinal jerk. This represents the weighting coefficient corresponding to the deviation between the initial longitudinal velocity and the target longitudinal velocity; This represents the weighting coefficient corresponding to the prediction duration; This represents the weighting coefficient corresponding to the longitudinal acceleration; The weighting coefficients represent the longitudinal jerk. Each weighting coefficient can be determined based on human experience or through extensive experimentation; this application does not impose any limitations on this.

[0183] In one alternative implementation, the lateral motion evaluation coefficient, longitudinal motion evaluation coefficient, and obstacle evaluation coefficient can be weighted and summed, and the sum can be used as the travel evaluation coefficient corresponding to the candidate path.

[0184] It should be noted that because a preset speed range was considered during the setting of the target longitudinal configuration data, the sampled values ​​of the target longitudinal configuration data are limited to a reasonable safe driving range. Unreasonably large or small values ​​will not appear; only normal driving conditions need to be met. Therefore, longitudinal speed is not the focus of the evaluation. Instead, a certain weighting coefficient is set to guide the actual driving speed to closely match the expected longitudinal speed, while also considering driving efficiency. Therefore, the weights of the lateral motion evaluation coefficient and the obstacle evaluation coefficient are both greater than the weight of the longitudinal motion evaluation coefficient.

[0185] S520, Based on the obstacle association data corresponding to the candidate path, determine the obstacle evaluation coefficient corresponding to the candidate path.

[0186] Among them, obstacle association data is used to quantify the spatial relationship between candidate paths and obstacles in the environment. Generally, the closer the candidate path is to the obstacle, the larger the obstacle evaluation coefficient calculated by the obstacle association data, indicating that the candidate path has a higher collision risk; when the candidate path is far from the obstacle, the corresponding obstacle evaluation coefficient is smaller, indicating that the candidate path is safer.

[0187] In some embodiments, for any candidate path, the obstacle association data corresponding to the candidate path is related to at least one of the following: the distance between each path point and an obstacle in the candidate path, the type of obstacle, or the method of obstacle acquisition. The distance between each path point and an obstacle can be the distance between each path point and the nearest obstacle. Taking the obstacle association data including the above-mentioned top data as an example, the process of determining the obstacle evaluation coefficient is described. For example, the obstacle evaluation coefficient can be determined based on the following formula:

[0188] ;

[0189] In the formula, The obstacle evaluation coefficient is represented by T; the prediction time is represented by t; the sampling time corresponding to each path point on the candidate path is represented by d. obs (t) represents the physical distance between the embodied robot and the nearest obstacle at sampling time t, if d obs If the path is smaller than the robot's envelope radius, the path is directly determined to be invalid. This is a preset constant; OBS(type) represents the influence coefficient of different types of obstacles, which can be set based on the type of obstacle or the method of obstacle acquisition.

[0190] For example, regarding the type of obstacle, obstacles that the robot doesn't want to encounter, or that would significantly impact its movement, have a higher impact coefficient; conversely, obstacles that are less likely to cause problems have a lower impact coefficient. For instance, AI obstacles, pet feces, and cables can easily cause contamination, structural jamming, or walking malfunctions upon contact, thus requiring a higher impact coefficient; low, soft objects like slippers pose less of a risk of collision and cause less damage, thus requiring a lower impact coefficient. Regarding the method of obstacle detection, methods with high detection accuracy and requiring strict avoidance have a higher impact coefficient; conversely, methods with lower accuracy have a lower impact coefficient. For example, obstacles identified by cameras can accurately distinguish object types and are often targets that need to be avoided, thus requiring a higher impact coefficient; LiDAR can only detect outline distance and cannot identify the specific attributes of obstacles, thus requiring a lower impact coefficient. Furthermore, manually added no-go zones are considered areas that the user requires to be avoided, thus requiring a higher impact coefficient.

[0191] S530, determine the travel evaluation coefficient corresponding to the candidate path based on the motion evaluation coefficient and the obstacle evaluation coefficient.

[0192] For example, the weighted sum of the motion evaluation coefficient and the obstacle evaluation coefficient can be used as the travel evaluation coefficient for the corresponding candidate path.

[0193] It should be noted that during obstacle avoidance by the embodied robot, greater emphasis is placed on the distance between the candidate path and the nearest obstacle, as well as the stability of lateral movement. If the candidate path is too close to the nearest obstacle, the embodied robot will be at risk of collision as it grazes the obstacle, and the corresponding obstacle evaluation coefficient should be significantly increased. If the deviation from the global reference path is too large during the avoidance process, or if the lateral velocity and acceleration change too much per unit time, it will lead to a decrease in driving stability, and the corresponding lateral motion evaluation coefficient should also be increased accordingly. Therefore, in the weighting configuration of the motion evaluation coefficients, the weights of the lateral motion evaluation coefficient and the obstacle evaluation coefficient are both greater than the weight of the longitudinal motion evaluation coefficient, in order to prioritize the optimization of lateral driving stability while ensuring safe obstacle avoidance.

[0194] In the above embodiments, the motion evaluation coefficient and obstacle evaluation coefficient are calculated separately and then fused to obtain the travel evaluation coefficient. This not only constrains the motion state of the embodied robot, ensuring smooth travel and a reasonable trajectory, but also quantifies the danger level of obstacles, effectively avoiding collision risks, further improving the rationality of the embodied robot's candidate path selection, and ensuring the safety of the embodied robot's travel process.

[0195] Based on the technical solutions of the above embodiments, this application also provides an optional embodiment. In this optional embodiment, the navigation and control method for the android provided by this application will be described in detail.

[0196] S601, based on the initial motion position of the embodied robot and the target position corresponding to the task to be performed, determine the global reference path;

[0197] S602, based on the global reference path, controls the embodied robot to move to the target position corresponding to the task to be performed, and continuously performs local path planning during the movement of the embodied robot;

[0198] S603, for each local path planning process, select at least one reference path point from the global reference path based on different prediction durations;

[0199] Among them, the path distance between the reference path point corresponding to different prediction durations and the current position of the embodied robot is different, and the path distance is positively correlated with the prediction duration;

[0200] S604. Each path point in the map data that is on the same horizontal reference line as the reference path point and meets the horizontal distance offset constraint is taken as at least two candidate path end points corresponding to the reference path point.

[0201] The direction of the horizontal reference line is perpendicular to the tangent of the reference path point on the global reference path;

[0202] S605, based on lateral motion offset constraints, sets target lateral configuration data for the end points of each candidate path;

[0203] The target lateral configuration data includes lateral velocity of 0, lateral acceleration of 0, and vertical distance between the candidate path endpoint and the global reference path.

[0204] S606, with the desired longitudinal speed as the center, performs longitudinal speed sampling within a preset speed range, and sets the longitudinal speed for the end point of each candidate path based on the sampling results;

[0205] S607, set the longitudinal acceleration corresponding to the end point of each candidate path to 0;

[0206] S608, for any candidate path, taking the current position of the embodied robot as the starting point and the end point of the candidate path as the trajectory endpoint, based on the initial configuration data corresponding to the current position and the target configuration data corresponding to the end point of the candidate path, determine the trajectory description data corresponding to the candidate path.

[0207] S609, for the prediction duration corresponding to the candidate path, the trajectory description data is discretized according to the preset time sampling interval to obtain the lateral configuration data and the corresponding longitudinal configuration data at different times when the embodied robot moves on the candidate path;

[0208] S610 integrates the lateral and longitudinal configuration data of the embodied robot at different times according to the order of each time to generate corresponding candidate paths;

[0209] S611, determine the relationship between the distance between each candidate path and the nearest obstacle and the first preset distance;

[0210] S612, if the distance between each candidate path and the nearest obstacle is less than the first preset distance, then execute S613; if the distance between each candidate path and the obstacle exceeds the first preset distance, then execute S618.

[0211] S613, determine whether each candidate path intersects with an obstacle; if some candidate paths intersect with obstacles, execute S614; if all candidate paths intersect with obstacles, or if none of the candidate paths intersect with obstacles, execute S615.

[0212] S614, Eliminate candidate paths that have overlap, and based on the remaining candidate paths, execute S615;

[0213] S615, Determine the motion evaluation coefficient corresponding to the candidate path based on the target configuration data corresponding to the end point of the candidate path;

[0214] The target configuration data includes lateral configuration data and longitudinal configuration data; the lateral motion evaluation coefficient is related to the lateral configuration data, and the longitudinal motion evaluation coefficient is related to the longitudinal configuration data; the lateral configuration data includes at least one of the following: the lateral velocity, lateral acceleration, and the change of lateral acceleration over time at the candidate path endpoint, and the vertical distance between the candidate path endpoint and the global reference path; the longitudinal configuration data includes at least one of the following: the difference between the longitudinal velocity and the longitudinal desired velocity at the candidate path endpoint, the longitudinal acceleration, the change of longitudinal acceleration over time, and the prediction duration.

[0215] S616, Determine the obstacle evaluation coefficient corresponding to the candidate path based on the obstacle association data corresponding to the candidate path;

[0216] For any candidate path, the obstacle association data corresponding to the candidate path is related to at least one of the following: the distance between each path point and the obstacle in the candidate path, the type of obstacle, or the method of collecting the obstacle.

[0217] S617, the lateral motion evaluation coefficient, longitudinal motion evaluation coefficient and obstacle evaluation coefficient are weighted and summed, and the sum is used as the travel evaluation coefficient corresponding to the candidate path;

[0218] S618, select the candidate path that at least partially overlaps with the global reference path, or whose distance from the global reference path is less than the second preset distance, as the target path for this planning, and execute S620;

[0219] S619, Based on the travel evaluation coefficients corresponding to each candidate path, select the target path for this planning from the selected candidate paths;

[0220] S620 controls the embodied robot to move according to the target path planned in this project;

[0221] S621, continue performing local path planning operations until the embodied robot reaches the target location.

[0222] It should be noted that if the embodied robot collides with an obstacle during its movement, the obstacle's information is added to the map data to update the map data. The embodied robot's current position is used as the initial movement position. Based on the initial movement position and the target position corresponding to the task to be performed, a new global reference path is determined, and the embodied robot is controlled to move along the new global reference path. During the movement, local path planning is continuously performed.

[0223] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0224] Based on the same inventive concept, this application also provides a navigation control device for a android that implements the navigation control method for the android described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the navigation control device for androids provided below can be found in the limitations of the navigation control method for androids described above, and will not be repeated here.

[0225] In one exemplary embodiment, such as Figure 6 As shown, a navigation control system for an embodied robot is provided, including: a path generation module 610, a coefficient determination module 620, a path selection module 630, a robot control module 640, and a path planning module 650, wherein:

[0226] The path generation module 610 is used to continuously perform local path planning for the embodied robot during its journey to the target location corresponding to the task to be performed based on the global reference path. For each local path planning process, at least three candidate paths are generated based on the current position of the embodied robot and the global reference path. The global reference path is determined based on the initial movement position of the embodied robot and the target location corresponding to the task to be performed.

[0227] The coefficient determination module 620 is used to determine the travel evaluation coefficient of the corresponding candidate path based on the evaluation reference data of each candidate path selected from the at least three candidate paths if the distance between at least three candidate paths and the obstacle is less than the first preset distance.

[0228] The path selection module 630 is used to select the target path for this planning from the selected candidate paths based on the travel evaluation coefficients corresponding to each candidate path.

[0229] The robot control module 640 is used to control the embodied robot to move according to the planned target path.

[0230] The path planning module 650 is used to continue performing local path planning operations until the embodied robot reaches the target position.

[0231] In some embodiments, the path generation module 610 includes a reference path point selection unit, configured to select at least one reference path point from a global reference path based on different prediction durations; the path distance between the reference path point corresponding to different prediction durations and the current position of the embodied robot is different, and the path distance is positively correlated with the prediction duration; a first determination unit, configured to determine at least two candidate path end points corresponding to any reference path point from map data according to the lateral movement offset constraint corresponding to the embodied robot; and a first generation unit, configured to generate at least three candidate paths starting from the current position of the embodied robot; wherein, one candidate path at least partially overlaps with the global reference path, or the distance between it and the global reference path is less than a second preset distance; and the remaining at least two candidate paths each correspond to a candidate path end point, and the candidate path end point is the trajectory endpoint of the corresponding candidate path.

[0232] In one embodiment, the lateral travel offset constraint includes a lateral distance offset constraint; correspondingly, the first determining unit is specifically used to select each path point in the map data that is on the same lateral reference line as the reference path point and meets the lateral distance offset constraint as at least two candidate path end points corresponding to the reference path point; wherein, the direction of the lateral reference line is perpendicular to the tangent of the reference path point on the global reference path.

[0233] In one embodiment, the lateral travel offset constraint further includes a lateral motion offset constraint; correspondingly, the first determining unit is further configured to: set target lateral configuration data for each candidate path end point based on the lateral motion offset constraint; sample longitudinal velocity within a preset velocity range centered on the longitudinal desired velocity, and set longitudinal velocity for each candidate path end point based on the sampling results; set the longitudinal acceleration corresponding to each candidate path end point to 0; wherein, the target lateral configuration data includes a lateral velocity of 0, a lateral acceleration of 0, and a vertical distance between the candidate path end point and the global reference path.

[0234] In one embodiment, the path generation module 610 is specifically used to generate a corresponding candidate path for any candidate path, taking the current position of the embodied robot as the starting point and the end point of the candidate path as the trajectory endpoint, based on the initial configuration data corresponding to the current position and the target configuration data corresponding to the end point of the candidate path; wherein, the target configuration data includes target lateral configuration data and target longitudinal configuration data; the initial configuration data includes initial lateral configuration data and initial longitudinal configuration data; the initial lateral configuration data includes the vertical distance between the current position and the global reference path, the lateral velocity and lateral acceleration of the embodied robot at the current position; the initial longitudinal configuration data includes the longitudinal velocity and longitudinal acceleration of the embodied robot at the current position.

[0235] In one embodiment, the path generation module 610 includes a second determining unit, used to determine trajectory description data corresponding to a candidate path based on the initial configuration data corresponding to the current position and the target configuration data corresponding to the end point of the candidate path; and a second generating unit, used to discretize the trajectory description data according to a preset time sampling interval for the predicted duration corresponding to the candidate path, to obtain the lateral configuration data and the corresponding longitudinal configuration data at different times when the embodied robot moves on the candidate path, and to integrate the lateral configuration data and the longitudinal configuration data at the corresponding times to generate the corresponding candidate path.

[0236] In one embodiment, the navigation control device of the embodied robot further includes a path designation module, which is used to select the candidate path that at least partially overlaps with the global reference path or has a distance of less than a second preset distance from the global reference path as the target path for this planning if the distance between each candidate path and the obstacle exceeds a first preset distance.

[0237] In one embodiment, the navigation control device of the embodied robot further includes an intersection judgment module, used to determine whether at least three candidate paths all intersect with obstacles; and if some candidate paths intersect with obstacles, to eliminate the intersecting candidate paths, and to select the remaining candidate paths and determine the corresponding evaluation reference data; and if at least three candidate paths all intersect with obstacles, or at least three candidate paths do not intersect with obstacles, to select all candidate paths and determine the corresponding evaluation reference data.

[0238] In one embodiment, the evaluation reference data includes target configuration data and obstacle association data; correspondingly, the coefficient determination module 620 includes a third determination unit, which determines the motion evaluation coefficient corresponding to any candidate path based on the target configuration data corresponding to the end point of the candidate path; a fourth determination unit, which determines the obstacle evaluation coefficient corresponding to the candidate path based on the obstacle association data corresponding to the candidate path; and a fifth determination unit, which determines the travel evaluation coefficient corresponding to the candidate path based on the motion evaluation coefficient and the obstacle evaluation coefficient.

[0239] In one embodiment, the motion evaluation coefficient includes a lateral motion evaluation coefficient and a longitudinal motion evaluation coefficient; correspondingly, the fifth determining unit is used to perform a weighted summation of the lateral motion evaluation coefficient, the longitudinal motion evaluation coefficient, and the obstacle evaluation coefficient, and use the sum as the travel evaluation coefficient corresponding to the candidate path; wherein, the weight of the lateral motion evaluation coefficient and the weight of the obstacle evaluation coefficient are both greater than the weight of the longitudinal motion evaluation coefficient.

[0240] In one embodiment, for any candidate path, the obstacle association data corresponding to the candidate path is related to at least one of the following: the distance between each path point and the obstacle in the candidate path, the type of obstacle, or the method of collecting the obstacle.

[0241] Target configuration data includes horizontal configuration data and vertical configuration data; the horizontal motion evaluation coefficient is related to the horizontal configuration data, and the vertical motion evaluation coefficient is related to the vertical configuration data.

[0242] The lateral configuration data includes at least one of the following: the lateral velocity, lateral acceleration, and the change of lateral acceleration over time at the candidate path endpoint; and the vertical distance between the candidate path endpoint and the global reference path. The longitudinal configuration data includes at least one of the following: the difference between the longitudinal velocity and the longitudinal desired velocity at the candidate path endpoint; the longitudinal acceleration and the change of longitudinal acceleration over time; and the prediction duration.

[0243] In one embodiment, the navigation control device of the embodied robot further includes a map update module, which is used to add the information of the obstacle to the map data to update the map data when the embodied robot collides with the obstacle during its movement; and to take the current position of the embodied robot as the initial movement position, determine a new global reference path based on the initial movement position and the target position corresponding to the task to be performed, and control the embodied robot to move along the new global reference path, and continuously perform local path planning during the movement.

[0244] The various modules in the navigation and control device of the aforementioned embodied robot can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0245] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 7 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a navigation control method for an embodied robot. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0246] Those skilled in the art will understand that Figure 7The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0247] In one exemplary embodiment, a body robot is provided, on which a computer device is provided, the computer device including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps in the above method embodiment.

[0248] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0249] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the method embodiments described above.

[0250] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0251] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0252] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A navigation and control method for an embodied robot, characterized in that, The method includes: During the process of moving to the target location corresponding to the task to be performed according to the global reference path, the embodied robot continuously performs local path planning, and for each local path planning process, generates at least three candidate paths based on the current position of the embodied robot and the global reference path; wherein, the global reference path is determined based on the initial movement position of the embodied robot and the target location corresponding to the task to be performed. If the distance between the at least three candidate paths and the obstacle is less than the first preset distance, then the travel evaluation coefficient corresponding to the corresponding candidate path is determined according to the evaluation reference data corresponding to each candidate path selected from the at least three candidate paths. Based on the travel evaluation coefficients corresponding to each of the candidate paths, the target path for this planning is selected from the selected candidate paths. Control the embodied robot to move according to the target path planned in this operation; Continue executing the local path planning operation until the embodied robot reaches the target location.

2. The method according to claim 1, characterized in that, Based on the current position of the android and the global reference path, at least three candidate paths are generated, including: Based on different prediction durations, at least one reference path point is selected from the global reference path; the path distance between the reference path point corresponding to different prediction durations and the current position of the embodied robot is different, and the path distance is positively correlated with the prediction duration; For any reference path point, based on the lateral movement offset constraint corresponding to the embodied robot, at least two candidate path endpoints corresponding to the reference path point are determined from the map data; Starting from the current position of the embodied robot, at least three candidate paths are generated; wherein, one of the candidate paths at least partially overlaps with the global reference path, or the distance between the candidate path and the global reference path is less than a second preset distance; the remaining at least two candidate paths correspond to the end points of the candidate paths, and the end points of the candidate paths are the trajectory endpoints of the corresponding candidate paths.

3. The method according to claim 2, characterized in that, The lateral travel offset constraint includes a lateral distance offset constraint; correspondingly, determining at least two candidate path endpoints corresponding to the reference path point from the map data based on the lateral travel offset constraint corresponding to the android includes: Each path point in the map data that is on the same horizontal reference line as the reference path point and meets the horizontal distance offset constraint is taken as at least two candidate path end points corresponding to the reference path point; wherein, the direction of the horizontal reference line is perpendicular to the tangent of the reference path point in the global reference path.

4. The method according to claim 3, characterized in that, The lateral travel offset constraint further includes a lateral motion offset constraint; correspondingly, the method further includes: Based on the lateral motion offset constraint, target lateral configuration data is set for the end points of each candidate path; Based on the longitudinal motion offset constraint, target longitudinal configuration data is set for the end points of each candidate path; The target lateral configuration data includes a lateral velocity of 0, a lateral acceleration of 0, and the vertical distance between the candidate path endpoint and the global reference path. Accordingly, setting target longitudinal configuration data for each candidate path endpoint based on longitudinal motion offset constraints includes: With the desired longitudinal speed as the center, longitudinal speed sampling is performed within a preset speed range, and based on the sampling results, longitudinal speed is set for the end points of each candidate path. Set the longitudinal acceleration corresponding to the end point of each candidate path to 0.

5. The method according to claim 4, characterized in that, The process involves generating at least three candidate paths, starting from the current location of the embodied robot, including: For any candidate path, starting from the current position of the embodied robot and ending at the end point of the candidate path, a corresponding candidate path is generated based on the initial configuration data corresponding to the current position and the target configuration data corresponding to the end point of the candidate path. The target configuration data includes the target lateral configuration data and the target longitudinal configuration data; the initial configuration data includes initial lateral configuration data and initial longitudinal configuration data; the initial lateral configuration data includes the vertical distance between the current position and the global reference path, the lateral velocity and lateral acceleration of the embodied robot at the current position; the initial longitudinal configuration data includes the longitudinal velocity and longitudinal acceleration of the embodied robot at the current position.

6. The method according to claim 5, characterized in that, The step of generating a corresponding candidate path based on the initial configuration data corresponding to the current location and the target configuration data corresponding to the candidate path endpoint includes: Based on the initial configuration data corresponding to the current location and the target configuration data corresponding to the candidate path endpoint, determine the trajectory description data corresponding to the candidate path; For the predicted duration corresponding to the candidate path, the trajectory description data is discretized according to a preset time sampling interval to obtain the lateral configuration data and the corresponding longitudinal configuration data at different times when the robot moves on the candidate path. The lateral configuration data and the longitudinal configuration data at the corresponding times are then integrated to generate the corresponding candidate path.

7. The method according to claim 2, characterized in that, After generating at least three candidate paths based on the current position of the android and the global reference path, the method further includes: If the distance between each candidate path and the obstacle exceeds the first preset distance, then the candidate path that at least partially overlaps with the global reference path, or whose distance from the global reference path is less than the second preset distance, will be the target path for this planning.

8. The method according to claim 1, characterized in that, After the distance between the at least three candidate paths and the obstacle is less than a first preset distance, the method further includes: Determine whether all three candidate paths intersect with the obstacle; If some of the candidate paths intersect with the obstacle, then the candidate paths that intersect are eliminated, and the remaining candidate paths are selected and the corresponding evaluation reference data is determined. If all three candidate paths intersect with the obstacle, or if none of the three candidate paths intersect with the obstacle, then all the candidate paths are selected and the corresponding evaluation reference data is determined.

9. The method according to claim 2, characterized in that, The evaluation reference data includes target configuration data and obstacle association data; correspondingly, determining the travel evaluation coefficient corresponding to each candidate path based on the evaluation reference data for each candidate path includes: For any candidate path, the motion evaluation coefficient corresponding to the candidate path is determined based on the target configuration data corresponding to the endpoint of the candidate path. Based on the obstacle association data corresponding to the candidate path, determine the obstacle evaluation coefficient corresponding to the candidate path; Based on the motion evaluation coefficient and the obstacle evaluation coefficient, the travel evaluation coefficient corresponding to the candidate path is determined.

10. The method according to claim 9, characterized in that, The motion evaluation coefficient includes a lateral motion evaluation coefficient and a longitudinal motion evaluation coefficient; determining the travel evaluation coefficient corresponding to the candidate path based on the motion evaluation coefficient and the obstacle evaluation coefficient includes: The lateral motion evaluation coefficient, the longitudinal motion evaluation coefficient, and the obstacle evaluation coefficient are weighted and summed, and the sum is used as the travel evaluation coefficient corresponding to the candidate path. The weights of the lateral motion evaluation coefficient and the obstacle evaluation coefficient are both greater than the weight of the longitudinal motion evaluation coefficient.

11. The method according to claim 10, characterized in that, For any of the candidate paths, the obstacle association data corresponding to the candidate path is related to at least one of the following: the distance between each path point and the obstacle in the candidate path, the type of obstacle, or the method of collecting the obstacle. The target configuration data includes horizontal configuration data and vertical configuration data; the horizontal motion evaluation coefficient is related to the horizontal configuration data, and the vertical motion evaluation coefficient is related to the vertical configuration data; The lateral configuration data includes at least one of the following: the lateral velocity, lateral acceleration, and the change of the lateral acceleration over time at the candidate path endpoint; and the vertical distance between the candidate path endpoint and the global reference path. The longitudinal configuration data includes at least one of the following: the difference between the longitudinal velocity and the longitudinal desired velocity at the candidate path endpoint, the longitudinal acceleration, and the change of the longitudinal acceleration over time; and the prediction duration.

12. The method according to any one of claims 1-11, characterized in that, If the android collides with an obstacle during its movement, the method further includes: The information about the obstacles is added to the map data to update the map data; The current position of the embodied robot is used as the initial movement position. A new global reference path is determined based on the initial movement position and the target position corresponding to the task to be performed. The embodied robot is then controlled to move along the new global reference path, and local path planning is continuously performed during the movement.

13. A navigation control system for an embodied robot, characterized in that, The system includes: The path generation module is used to continuously perform local path planning during the process of the embodied robot traveling to the target location corresponding to the task to be executed according to the global reference path, and to generate at least three candidate paths for each local path planning process based on the current position of the embodied robot and the global reference path; wherein, the global reference path is determined based on the initial movement position of the embodied robot and the target location corresponding to the task to be executed. The coefficient determination module is used to determine the travel evaluation coefficient corresponding to the corresponding candidate path based on the evaluation reference data corresponding to each candidate path selected from the at least three candidate paths if the distance between the at least three candidate paths and the obstacle is less than the first preset distance. The path selection module is used to select the target path for this planning from the selected candidate paths according to the travel evaluation coefficients corresponding to each candidate path. The robot control module is used to control the movement of the embodied robot according to the planned target path. The path planning module is used to continue executing the local path planning operation until the embodied robot reaches the target position.

14. A body-worn robot, on which a computer device is mounted, said computer device including a memory and a processor, said memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-12.

Images