Foot Landing Control Method and System for Legged Robot with Automatic Gait Switching

Abstract

The present invention provides a foot landing control method and system for a legged robot with automatic gait switching. The method classifies the environment in front of the robot according to front image data to obtain different road condition areas. Based on the correspondence between pre-calibrated image data and point cloud data, point cloud data corresponding to each road condition area is obtained, and the point cloud normal vector of each area is extracted. Point clouds with consistent normal vectors or differences within a set range in each road condition area are clustered, and the elevation information of the clustered area is calculated, and the point cloud data with elevation information is stored in a point cloud feature data queue. During the movement of the robot, the point cloud data in the point cloud feature queue are iteratively matched and merged to obtain a local environment feature map around the robot's foot end. The plane area in the local environment feature map is extracted, and the landing point is obtained by iteratively calculating the normals of different areas in the local map. The present invention realizes adaptive switching and correction of gait and landing point based on preset rules.

Description

Technical Field

[0001] The present invention relates to the technical field of legged robots, and in particular to a foot placement control method and system for a legged robot with automatic gait switching. Background Art

[0002] The statements in this section merely provide background art related to the present invention and do not necessarily constitute prior art.

[0003] Existing legged robot body control system (such as Figure 1 The controller, shown in the dashed box in Figure 3, inputs the desired input (torso posture, gait, foot trajectory, foot landing point, etc.). The controller then combines the state estimator's 3D posture, velocity, and foot contact information, derived from the robot's current torso posture, gait, and leg joint motor position / velocity / torque data. Using control algorithms such as virtual model control (VMC) and model predictive control (MPC), the controller calculates the output of the robot's leg joint motors. These outputs are then executed by the motor drivers of each joint, tracking the desired input and enabling motion under complex terrain conditions. If the robot experiences abnormal motion, a manual emergency stop input disconnects the motor drivers from controlling the joint motors, protecting the robot.

[0004] The inventors have discovered that the current legged robot walking control has the following problems:

[0005] (1) Due to the complexity of the environment, the current robot control system does not introduce various environmental factors into the control process. This also leads to the robot's insufficient adaptability when facing complex situations such as continuous irregular steps and sudden changes in terrain, affecting the safety of the robot and the equipment installed on it;

[0006] (2) Conventional foot-end touchdown detection methods include installing a touchdown switch at the foot end and a one-dimensional force sensor at the calf. Due to the strong impact between the foot and the ground, the touchdown switch is easily damaged and needs to be replaced frequently. The one-dimensional force sensor at the calf solves the problem of easy damage due to mechanical impact, but requires the calf to enter the ground nearly vertically. When the robot overcomes obstacles, the calf may be close to horizontal. At this time, the one-dimensional force sensor cannot detect the touchdown force in the vertical direction, which will lead to missed detection. Summary of the Invention

[0007] In order to address the shortcomings of the existing technology, the present invention collects the robot's body state and posture data in flat road conditions as well as gravel, grass, steps and other road environments, and analyzes the protection needs of the robot body and equipment when it falls. Based on manually set protection rules, a foot landing control method and system with automatic gait switching for a legged robot is proposed, which realizes active detection of the ground environment at the robot's foot end and adaptive switching and correction of gait and landing point based on preset rules.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] A first aspect of the present invention provides a foot landing control method for a legged robot with automatic gait switching.

[0010] A foot landing control method for a legged robot with automatic gait switching includes the following steps:

[0011] Based on the image data in front of the robot, the environment in front of the robot is classified to obtain point cloud data corresponding to different road conditions.

[0012] Cluster the point cloud normal vectors in each road condition area, calculate the elevation information of the clustered area, and store the point cloud data with elevation information in the point cloud feature data queue;

[0013] The point cloud data in the point cloud feature queue are iteratively matched and merged to obtain the local environment feature map around the robot foot end. The landing point is obtained by iteratively calculating the normals of different plane areas in the local map.

[0014] As an optional implementation of the first aspect of the present invention, based on the correspondence between pre-calibrated image data and point cloud data, point cloud data corresponding to each road condition area is obtained, and the point cloud normal vector of each area is extracted;

[0015] The point clouds with the same normal vectors or differences within the set range in each road condition area are clustered, and the elevation information of the clustered area is calculated. The point cloud data with elevation information is stored in the point cloud feature data queue.

[0016] As a further limitation of the first aspect of the present invention, a planar region in a local environment feature map is extracted based on a RANSAC algorithm, and the consistency of the normals in each plane is evaluated by iteratively calculating the normals of different regions in the local map;

[0017] If the normal direction of a plane area is vertical or the inclination angle is within the set range, this plane area can be used as an alternative landing point for the robot to walk, and the corresponding point cloud can be marked.

[0018] As a further limitation of the first aspect of the present invention, by projecting the local map point cloud onto a plane grid and filling the grid with the average height of the points contained in the grid, ground grid data that can be used for landing is obtained.

[0019] As an optional implementation of the first aspect of the present invention, if there is an obstacle on the bottom surface and the obstacle size is larger than a set value, there may be a landing point on the surface, and the landing point selection on the obstacle includes:

[0020] If there is a slope ahead but the inclination angle is less than the first set threshold, you can land; if there is a gravel road ahead, the inclination angle is greater than the first set threshold and the range of change is greater than the second set threshold, you cannot land; if there is a gravel road ahead, the inclination angle is less than the first set threshold and the range of change is less than the second set threshold, you can land.

[0021] As an optional implementation method of the first aspect of the present invention, according to the different road conditions identified, the ratio of the robot's foot's air time to the ground contact time, the height of the sole of the foot from the ground, and the step length are adjusted to configure gait rule parameters suitable for the corresponding environment.

[0022] As an optional implementation of the first aspect of the present invention, the gait of the current environment is dynamically selected based on the current perceived environment characteristics and matching with known environment characteristics.

[0023] As an optional implementation method of the first aspect of the present invention, based on the local environment feature map, based on the principle of being closest to the expected landing point and minimizing the reaction force after landing, and based on the evaluation of the landing area within the artificially set neighborhood of the expected landing point of the current gait, the landing points of each foot end of the current gait are selected and the gait input of the controller is dynamically corrected.

[0024] A second aspect of the present invention provides a foot landing control system for a legged robot with automatic gait switching.

[0025] A foot control system for a legged robot with automatic gait switching, comprising:

[0026] The region division module is configured to: classify the environment in front of the robot based on the image data in front of the robot, and obtain point cloud data corresponding to different road condition areas;

[0027] The point cloud clustering module is configured to: cluster the point cloud normal vectors in each road condition area, calculate the elevation information of the clustering area, and store the point cloud data with the elevation information into the point cloud feature data queue;

[0028] The footfall acquisition module is configured to iteratively match and merge the point cloud data in the point cloud feature queue to obtain a local environment feature map around the robot foot end, and obtain the footfall point by iteratively calculating the normals of different plane areas in the local map.

[0029] A third aspect of the present invention provides a method for detecting the foot contact of a legged robot with automatic gait switching.

[0030] A method for detecting foot contact of a legged robot with automatic gait switching includes the following steps:

[0031] The foot landing control method for a legged robot with automatic gait switching according to the first aspect of the present invention is used to obtain a foot landing position, and a planning result of a foot end trajectory in a swing phase is obtained based on the foot landing position;

[0032] According to the planning results of the landing point position and the foot end trajectory in the swing phase, touchdown detection is performed based on the joint torque to determine whether it is one of the following: early touchdown, expected touchdown, and delayed touchdown. If so, it is determined to be a touchdown and the legged robot switches to the support phase; if not, it is determined to be a step on the air and switches to quadruped support.

[0033] As an optional implementation of the second aspect of the present invention, a ground contact detection is performed based on the joint torque to determine whether the ground contact is premature. If so, it is determined to be a ground contact, and the legged robot switches to the support phase; if not, the ground contact detection is continued;

[0034] Determine whether the foot touches the ground as expected. If so, it is determined to be touching the ground and the legged robot switches to the support phase. If not, the foot continues to probe down a set distance and continues to perform touchdown detection.

[0035] Determine whether the touchdown is delayed. If so, it is determined to be touchdown and the legged robot switches to the support phase; if not, it is determined to be stepping on air and switches to quadruped support.

[0036] As an optional implementation of the second aspect of the present invention, performing ground touchdown detection includes:

[0037] During the leg swinging process, the joint torque is detected in real time, and the external force on the foot end is calculated based on the joint torque. When the external force on the foot end is greater than or equal to a first preset threshold, it is determined that the leg has touched the ground.

[0038] As a further limitation of the second aspect of the present invention, calculating the external force applied to the foot end based on the joint torque includes:

[0039]

[0040] Among them, q is the angle of the three joints of a leg, is the angular acceleration of the three joints of one leg, is the angular velocity of the three joints of one leg, M is the mass matrix, C is the Gordian force matrix and centripetal force matrix, G is the gravity matrix, τ is the joint torque vector, J is the Jacobian matrix of the single leg, and F is the external force on the foot.

[0041] A fourth aspect of the present invention provides a foot-end ground contact detection system for a legged robot with automatic gait switching.

[0042] A foot-end touchdown detection system for a legged robot with automatic gait switching, comprising:

[0043] The foothold selection module is configured to: select a foothold based on the pre-processed point cloud data of the robot's forward movement, and obtain the planning results of the foothold position and the foot end trajectory of the swing phase;

[0044] The touchdown detection module is configured to: perform touchdown detection based on the joint torque according to the planning results of the foot landing point position and the foot end trajectory in the swing phase, and determine whether it is one of the following: early touchdown, expected touchdown, and delayed touchdown. If so, it is determined to be a touchdown and the legged robot switches to the support phase; if not, it is determined to be a step on the air and switches to quadruped support.

[0045] A fifth aspect of the present invention is a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of the foot landing control method for a legged robot with automatic gait switching as described in the first aspect of the present invention or the foot-end touchdown detection method for a legged robot with automatic gait switching as described in the third aspect of the present invention.

[0046] The sixth aspect of the present invention provides an electronic device, comprising a memory, a processor, and a program stored in the memory and runnable on the processor, wherein when the processor executes the program, it implements the steps in the foot landing control method of a legged robot with automatic gait switching as described in the first aspect of the present invention or the foot-end touchdown detection method of a legged robot with automatic gait switching as described in the third aspect of the present invention.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] 1. The present invention innovatively proposes a foot landing control method for a legged robot with automatic gait switching. By collecting the robot's body state and posture data on flat roads and road environments such as gravel, grass, and steps, and analyzing the protection needs of the robot body and equipment when it falls, on the basis of manually set protection rules, the method realizes active detection of the ground environment at the robot's foot end and adaptive switching and correction of gait and landing points based on preset rules; a landing point and gait feedback control loop based on environmental perception is added to the outer layer of the original robot body control system. By actively switching the robot's gait and correcting the landing point in different operating road environments, the stability and safety of the robot's operation in complex environments are improved.

[0049] 2. The present invention creatively proposes a method for detecting foot contact with the ground for a legged robot with automatic gait switching. It adopts an "external sensor + internal sensor" solution, which avoids the defects of various previous ground contact detection methods. No sensors need to be installed on the calves and feet, which reduces the weight and moment of inertia of the legs, extends the life of the equipment, and improves the wading ability.

[0050] Advantages of additional aspects of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0052] Figure 1 Schematic diagram of the principle of the foot landing control method of a legged robot with automatic gait switching provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0053] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0054] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs.

[0055] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0056] In the absence of conflict, the embodiments of the present invention and the features thereof may be combined with each other.

[0057] Example 1:

[0058] like Figure 1 As shown, embodiment 1 of the present invention provides a foot landing control method for a legged robot with automatic gait switching, comprising the following process:

[0059] S1: Environmental Perception

[0060] It is mainly composed of a solid-state laser sensor, a binocular camera and a 3D laser radar installed on the front of the robot. Through the combination of different sensors, the robot's blind spots in environmental detection are reduced, and the robot can effectively collect environmental information at different distances, near, medium and far.

[0061] S2: Feature Recognition

[0062] Based on the current environment image, point cloud and other information output by the environmental perception module, the feature recognition and classification algorithm is used to extract the current environmental characteristics of the robot (surface texture, normal vector, elevation, etc.), realize the classification of the current environmental characteristics of the robot, and at the same time construct a local feature map of the robot's foot end to provide a basis for subsequent foot landing and gait control.

[0063] S2.1: Obtain image data of the robot's forward direction and use the LBP method (Local Binary Patterns) to extract texture features in front of the robot. Then, use a convolutional neural network to classify the environment in front of the robot and obtain information about possible areas such as flat roads, steps, and water surfaces.

[0064] S2.2: Based on the pre-calibrated correspondence between the image and the point cloud area, obtain the point cloud of the flat road surface, steps, water surface, etc. in the above steps, and use the normal vector extraction function in the PCL Point Cloud Library to extract the normal vector of each point in the area.

[0065] S2.3: Using the K-Means clustering algorithm, the point clouds in the area with consistent normal vectors are clustered, and the elevation information of the clustered area is calculated, which realizes the quantitative processing of the visual extraction area and compresses the point cloud data with elevation information into a fixed-length feature data queue.

[0066] S2.4: During robot movement, the ICP algorithm (Iterative Closest Point) is used to iteratively match and merge the point cloud data in the point cloud feature queue to construct a feature map of the local environment around the robot's foot, thereby providing a basis for subsequent foot placement and gait control.

[0067] S3: Footfall Control

[0068] The front downward-tilted solid-state laser sensor scans the ground information on the soles of the foot-type mobile chassis in real time and transmits it back to the gait control module of the foot-type mobile chassis, thereby adjusting the landing point of the foot-type mobile chassis to avoid stepping on empty space or obstacles.

[0069] S3.1: Use a plane extraction algorithm to extract plane regions from the local environment feature map. This embodiment is based on the RANSAC algorithm (Random Sample Consensus). By iteratively calculating the normals of different regions in the local map, the consistency of the normals within the plane is evaluated. If the normal direction is vertical or the tilt angle is within a set range, it indicates that the plane is not tilted significantly and can be used as an alternative landing point for the robot to walk. The normal is then marked within the point cloud.

[0070] S3.2: After that, the local map point cloud is projected onto a plane grid and the grid is filled with the mean height of the points contained in the grid, thereby forming ground grid data that can be used for landing.

[0071] For obstacles on the ground, if the obstacle is large, there may be landing points on its surface. In addition, the landing points are extracted using an adjustable threshold:

[0072] If there is a slope ahead but the inclination angle is less than the first set threshold, you can land your foot;

[0073] If there is a gravel road ahead, and the inclination angle is greater than the first set threshold and the change is greater than the second set threshold, you cannot land.

[0074] If there is a gravel road ahead, and the inclination angle is less than the first set threshold and the change is less than the second set threshold, a landing point can be provided. At this time, the robot mainly relies on the balance and stability control algorithm to maintain stability. The selection of the landing point cannot determine whether a certain part of the gravel is stepped on to ensure instability.

[0075] In this embodiment, the inclination of the landing area cannot be too large, and the inclination change in the area cannot be too drastic. The inclination change is measured by calculating the degree of height change between grids. If it exceeds the set threshold, it is considered to be a drastic change and not suitable for landing.

[0076] S4: Gait Control

[0077] This module can accept manually configured gait rules in different environments, select a safe gait suitable for the current environment based on the robot's body state and environmental feature information, and then search for the landing area output by the foot control module in the current foot-end neighborhood. Combined with the expected gait information, it realizes dynamic correction of the controller's gait input to ensure the stable operation of the robot in complex environments.

[0078] S4.1: Based on different environmental characteristics (flat ground, stairs, water surface, etc.), manually adjust variables such as the ratio of the robot's foot's airborne time to the ground contact time, the height of the sole of the foot from the ground, and the stride length (an empirical value obtained through manual testing in the relevant environment, and the landing point will be selected later to achieve gait correction), and configure gait rule parameters applicable to the corresponding environment.

[0079] S4.2: Based on the current perceived environmental features and matching them with known environmental features, dynamically select a gait for the current environment (by visually identifying the ground environmental features and then selecting the above-mentioned set gait based on the features).

[0080] S4.3: Based on the established local environment map, and according to the principle of proximity to the desired footfall point and minimum reaction force after footfall, the footfall points of each foot end of the current gait are selected by evaluating the possible footfall areas within the manually set neighborhood of the desired footfall point of the current gait, thereby achieving dynamic correction of the controller gait input.

[0081] The set of all landing grids in the local environment map is:

[0082] P={P1,P2,…P n}

[0083]

[0084] P Q is the selected foothold grid, is the reaction force in each candidate footfall point (can also be considered as the output force of the robot leg), P C is the current foothold grid position, It is calculated through the dynamic equation of the legged robot.

[0085] Example 2:

[0086] Embodiment 2 of the present invention provides a foot landing control system for a legged robot with automatic gait switching, comprising:

[0087] The region division module is configured to: classify the environment in front of the robot based on the image data in front of the robot, and obtain point cloud data corresponding to different road condition areas;

[0088] The point cloud clustering module is configured to: cluster the point cloud normal vectors in each road condition area, calculate the elevation information of the clustering area, and store the point cloud data with the elevation information into the point cloud feature data queue;

[0089] The footfall acquisition module is configured to iteratively match and merge the point cloud data in the point cloud feature queue to obtain a local environment feature map around the robot foot end, and obtain the footfall point by iteratively calculating the normals of different plane areas in the local map.

[0090] The working method of the system is the same as the foot landing control method of the legged robot with automatic gait switching provided in Example 1, and will not be repeated here.

[0091] Example 3:

[0092] Embodiment 3 of the present invention provides a method for detecting foot contact of a legged robot with automatic gait switching, comprising the following steps:

[0093] The foot landing point position is obtained using the foot landing control method for a legged robot with automatic gait switching described in Example 1 of the present invention, and a planning result of the foot end trajectory of the swing phase is obtained based on the foot landing point position;

[0094] Based on the planned results of the foot landing point position and the foot end trajectory in the swing phase, the ground contact detection is performed based on the joint torque to determine whether it has touched the ground in advance. If so, it is determined to be touching the ground and the legged robot switches to the support phase; if not, the next step is executed;

[0095] The ground contact detection is performed based on the joint torque to determine whether the ground contact is as expected. If so, it is determined to be ground contact and the legged robot switches to the support phase; if not, the next step is executed;

[0096] The foot continues to probe down a set distance, and performs ground contact detection based on the joint torque to determine whether it is delayed. If so, it is determined to be ground contact, and the legged robot switches to the support phase; if not, it is determined to be stepping on air, and switches to the support phase.

[0097] Specifically, during leg swing, the joint torque is detected in real time, and the external force on the foot end is calculated using the following equation:

[0098]

[0099] Among them, q is the angle of the three joints of a leg (3×1 matrix), is the angular acceleration of the three joints of a leg (3×1 matrix), is the angular velocity of the three joints of one leg (a 3×1 matrix), M is the mass matrix (a 3×3 matrix), C is the Coriolis force matrix and the centripetal force matrix, G is the gravity matrix (a 3×1 matrix), τ is the joint torque vector (a 3×1 matrix), J is the Jacobian matrix of a single leg (a 3×3 matrix), and F is the external force on the foot (a 3×1 matrix, the three elements are the forces in the x-direction, y-direction, and z-direction respectively).

[0100] Due to modeling errors and sensor noise, the estimated value of F will fluctuate within a small range, but it is far from the support force brought by the foot touching the ground. Therefore, it can be determined as a ground contact state when it exceeds a threshold.

[0101] For non-foot contact situations (such as the calf hitting a raised area on the ground), the F estimated by this method will also exceed the threshold. Although the foot is suspended in the air and the estimated F has a large deviation, it can still be determined as a touchdown because the leg contact area has provided sufficient support.

[0102] According to the above method, the swinging leg is controlled to move to the desired landing point. If the touchdown is detected in advance, it is judged as touching the ground; if the touchdown is not detected on time, it continues to move down for a preset distance; if the touchdown is still not detected after exceeding the preset distance, it is judged as stepping on air, and the state of forced switching to four-leg support is tried to regain balance.

[0103] Example 4:

[0104] Embodiment 4 of the present invention provides a foot-end ground contact detection system for a legged robot with automatic gait switching, comprising:

[0105] The foothold selection module is configured to: select a foothold based on the pre-processed point cloud data of the robot's forward movement, and obtain the planning results of the foothold position and the foot end trajectory of the swing phase;

[0106] The touchdown detection module is configured to: perform touchdown detection based on the joint torque according to the planning results of the foot landing point position and the foot end trajectory in the swing phase, and determine whether it is one of the following: early touchdown, expected touchdown, and delayed touchdown. If so, it is determined to be a touchdown and the legged robot switches to the support phase; if not, it is determined to be a step on the air and switches to quadruped support.

[0107] Specifically, during leg swing, the joint torque is detected in real time, and the external force on the foot end is calculated using the following equation:

[0108]

[0109] Among them, q is the angle of the three joints of a leg (3×1 matrix), is the angular acceleration of the three joints of a leg (3×1 matrix), is the angular velocity of the three joints of one leg (a 3×1 matrix), M is the mass matrix (a 3×3 matrix), C is the Coriolis force matrix and the centripetal force matrix, G is the gravity matrix (a 3×1 matrix), τ is the joint torque vector (a 3×1 matrix), J is the Jacobian matrix of a single leg (a 3×3 matrix), and F is the external force on the foot (a 3×1 matrix, the three elements are the forces in the x-direction, y-direction, and z-direction respectively).

[0110] Due to modeling errors and sensor noise, the estimated value of F will fluctuate within a small range, but it is far from the support force brought by the foot touching the ground. Therefore, it can be determined as a ground contact state when it exceeds a threshold.

[0111] For non-foot contact situations (such as the calf hitting a raised area on the ground), the F estimated by this method will also exceed the threshold. Although the foot is suspended in the air and the estimated F has a large deviation, it can still be determined as a touchdown because the leg contact area has provided sufficient support.

[0112] According to the above method, the swinging leg is controlled to move to the desired landing point. If the touchdown is detected in advance, it is judged as touching the ground; if the touchdown is not detected on time, it continues to move down for a preset distance; if the touchdown is still not detected after exceeding the preset distance, it is judged as stepping on air, and the state of forced switching to four-leg support is tried to regain balance.

[0113] Example 5:

[0114] Embodiment 5 of the present invention provides a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, the program implements the steps in the foot landing control method of a legged robot with automatic gait switching as described in Embodiment 1 of the present invention or the foot-end touchdown detection method of a legged robot with automatic gait switching as described in Embodiment 3 of the present invention.

[0115] Example 6:

[0116] Embodiment 6 of the present invention provides an electronic device, comprising a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, the processor implements the steps in the foot landing control method for a legged robot with automatic gait switching as described in Embodiment 1 of the present invention or the foot-end touchdown detection method for a legged robot with automatic gait switching as described in Embodiment 3 of the present invention.

[0117] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage and optical storage, etc.) containing computer-usable program code.

[0118] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0119] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0120] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0121] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware through a computer program. The program can be stored in a computer-readable storage medium, and when executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0122] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for controlling the landing of a legged robot with automatic gait switching, characterized in that: The following processes are included: Based on the image data in front of the robot, the environment in front of the robot is classified to obtain point cloud data corresponding to different road conditions. Cluster the point cloud normal vectors in each road condition area, calculate the elevation information of the clustered area, and store the point cloud data with elevation information in the point cloud feature data queue; Iteratively match and merge the point cloud data in the point cloud feature queue to obtain a local environmental feature map around the robot's foot. The foothold is obtained by iteratively calculating the normals of different plane areas in the local map. The consistency of the normals in each plane is evaluated by iteratively calculating the normals of different areas in the local map. If the normal direction of a plane area is vertical or the tilt angle is within the set range, this plane area can be used as an alternative foothold for the robot to walk, and the corresponding point cloud is marked. Adjust the ratio of the robot's foot's airborne time to ground contact time, the height of the foot from the ground, and the stride length according to different road conditions and configure gait rule parameters suitable for the corresponding environment. According to the local environment feature map, the principle of being closest to the expected landing point and minimizing the reaction force after landing, and based on the evaluation of the landing area within the manually set neighborhood of the expected landing point of the current gait, the landing points of each foot end of the current gait are selected, and the gait input of the controller is dynamically corrected.

2. The method for controlling the footfall of a legged robot with automatic gait switching according to claim 1, wherein: Based on the correspondence between the pre-calibrated image data and the point cloud data, the point cloud data corresponding to each road condition area is obtained, and the point cloud normal vector of each area is extracted; The point clouds with the same normal vectors or differences within the set range in each road condition area are clustered, and the elevation information of the clustered area is calculated. The point cloud data with elevation information is stored in the point cloud feature data queue.

3. The method for controlling the footfall of a legged robot with automatic gait switching according to claim 2, wherein: The planar area in the local environment feature map is extracted based on the RANSAC algorithm.

4. The method for controlling the footfall of a legged robot with automatic gait switching according to claim 3, wherein: By projecting the local map point cloud onto a plane grid and filling the grid with the mean height of the points contained in the grid, ground grid data that can be used for landing is obtained.

5. The method for controlling the footfall of a legged robot with automatic gait switching according to claim 1, wherein: If there is an obstacle on the bottom surface and the obstacle size is larger than the set value, there may be a landing point on the surface. The landing point selection on the obstacle includes: If there is a slope ahead but the inclination angle is less than the first set threshold, you can land; if there is a gravel road ahead, the inclination angle is greater than the first set threshold and the range of change is greater than the second set threshold, you cannot land; if there is a gravel road ahead, the inclination angle is less than the first set threshold and the range of change is less than the second set threshold, you can land.

6. The method for controlling the footfall of a legged robot with automatic gait switching according to claim 1, wherein: Based on the current perceived environment characteristics and matching them with known environment characteristics, the gait of the current environment is dynamically selected.

7. A foot control system for a legged robot with automatic gait switching, characterized in that: include: The region division module is configured to: classify the environment in front of the robot based on the image data in front of the robot, and obtain point cloud data corresponding to different road condition areas; The point cloud clustering module is configured to: cluster the point cloud normal vectors in each road condition area, calculate the elevation information of the clustering area, and store the point cloud data with the elevation information into the point cloud feature data queue; The foothold acquisition module is configured to: iteratively match and merge the point cloud data in the point cloud feature queue to obtain a local environmental feature map around the robot's foot end, and obtain the foothold by iteratively calculating the normals of different plane areas in the local map; iteratively calculate the normals of different areas in the local map to evaluate the consistency of the normals in each plane. If the normal direction of a plane area is vertical or the tilt angle is within a set range, this plane area can be used as an alternative foothold for the robot to walk, and the corresponding point cloud is marked; Adjust the ratio of the robot's foot's airborne time to ground contact time, the height of the foot from the ground, and the stride length according to different road conditions and configure gait rule parameters suitable for the corresponding environment. According to the local environment feature map, the principle of being closest to the expected landing point and minimizing the reaction force after landing, and based on the evaluation of the landing area within the manually set neighborhood of the expected landing point of the current gait, the landing points of each foot end of the current gait are selected, and the gait input of the controller is dynamically corrected.

8. A method for detecting foot contact of a legged robot with automatic gait switching, characterized in that: The following processes are included: The foot landing control method for a legged robot with automatic gait switching according to any one of claims 1 to 6 is used to obtain a foot landing point position, and a planning result of a foot end trajectory of a swing phase is obtained based on the foot landing point position; According to the planning results of the landing point position and the foot end trajectory in the swing phase, touchdown detection is performed based on the joint torque to determine whether it is one of the following: early touchdown, expected touchdown, and delayed touchdown. If so, it is determined to be a touchdown and the legged robot switches to the support phase; if not, it is determined to be a step on the air and switches to quadruped support.

9. The method for detecting foot contact of a legged robot with automatic gait switching according to claim 8, wherein: Perform ground contact detection based on joint torque to determine whether the robot has touched the ground in advance. If so, the robot is deemed to have touched the ground and switches to the support phase. If not, the robot continues to perform ground contact detection. Determine whether the foot touches the ground as expected. If so, it is determined to be touching the ground and the legged robot switches to the support phase. If not, the foot continues to probe down a set distance and continues to perform touchdown detection. Determine whether the touchdown is delayed. If so, it is determined to be touchdown and the legged robot switches to the support phase; if not, it is determined to be stepping on air and switches to quadruped support.

10. The method for detecting foot contact of a legged robot with automatic gait switching according to claim 8 or 9, wherein: Perform touchdown detection, including: During the leg swinging process, the joint torque is detected in real time, and the external force on the foot end is calculated based on the joint torque. When the external force on the foot end is greater than or equal to a first preset threshold, it is determined that the leg has touched the ground.

11. The method for detecting foot contact of a legged robot with automatic gait switching according to claim 10, wherein: Calculate the external force on the foot based on the joint torque, including: in, is the angle of the three joints of a leg, is the angular acceleration of the three joints of one leg, is the angular velocity of the three joints of one leg, is the mass matrix, are the Gordian force matrix and the centripetal force matrix, is the gravity matrix, is the joint torque vector, is the single-leg Jacobian matrix, The external force applied to the foot end.

12. A foot-end touchdown detection system for a legged robot with automatic gait switching, using the foot-end touchdown control method for a legged robot with automatic gait switching according to any one of claims 1 to 6, characterized in that: include: The foothold selection module is configured to: select a foothold based on the pre-processed point cloud data of the robot's forward movement, and obtain the planning results of the foothold position and the foot end trajectory of the swing phase; The touchdown detection module is configured to: perform touchdown detection based on the joint torque according to the planning results of the foot landing point position and the foot end trajectory in the swing phase, and determine whether it is one of the following: early touchdown, expected touchdown, and delayed touchdown. If so, it is determined to be a touchdown and the legged robot switches to the support phase; if not, it is determined to be a step on the air and switches to quadruped support.

13. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, the steps of the foot landing control method of a legged robot with automatic gait switching as described in any one of claims 1 to 6 or the foot touchdown detection method of a legged robot with automatic gait switching as described in any one of claims 8 to 11 are implemented.

14. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein: When the processor executes the program, it implements the steps in the foot landing control method of a legged robot with automatic gait switching as described in any one of claims 1-6 or the foot end touchdown detection method of a legged robot with automatic gait switching as described in any one of claims 8-11.

Images