AMR robot path control method and system

Abstract

The invention relates to the technical field of AMR robot path control, and provides an AMR robot path control method and system, and the method comprises the steps: obtaining the motion state information of an AMR robot and the current consumption information of a drive motor; the turning radius of the AMR robot is calculated according to the motion state information, whether the AMR robot is in a high-risk positioning working condition or not is judged according to the turning radius and driving motor current consumption information, and a high-risk judgment result is obtained; adjusting a matching quality threshold of the positioning system for receiving the position information according to the high-risk judgment result; based on the adjusted matching quality threshold, when a matching score corresponding to the laser radar data is received, whether the pose of the AMR robot is updated or not is judged, and a control action is executed according to a judgment result; and after the high-risk positioning working condition is relieved, the matching quality threshold of the positioning system is recovered. The method has the advantages that the positioning robustness and safety of the AMR robot are improved, and traffic conflicts or secondary collision with facilities are prevented.

Description

Technical Field

[0001] The present invention relates to the technical field of AMR robot path control, and in particular to an AMR robot path control method and system. Background Art

[0002] In modern smart warehousing environments, autonomous mobile robot (AMR) systems are widely used for material handling, and accurate path control is crucial. However, long-term operation of AMR robots on mixed floors with varying friction coefficients can cause asymmetric wear on their drive wheels. To maintain surface tracking accuracy, the AMR's path control system continuously outputs small compensation control variables to the drive motors to offset this physical deviation. This can make the robot appear to be accurately following the path, but its internal motor control is actually in an asymmetric correction state.

[0003] When an AMR robot performs such challenging turns, the combined effects of compensatory output and extreme path planning can cause a slight "overswing" in the robot's actual turning trajectory—a subtle deviation from the planned trajectory. In this overswing, the onboard LiDAR sensor's scanning data mismatches the angles predicted by the positioning system model, leading to "feature mismatches" in the positioning algorithm. This mismatch causes the positioning system to instantly update the AMR's position to an inaccurate position and posture, even without triggering a physical collision warning. The AMR robot, carrying this "error of confidence," continues its subsequent mission, its actual position significantly deviating from its correct position on the map. This can easily lead to potential traffic conflicts or secondary collisions with other vehicles, severely impacting the safety and efficiency of the AMR system.

[0004] In view of the above problems, the existing technology is in urgent need of improvement. Summary of the Invention

[0005] The present application discloses an AMR robot path control method and system, which aims to solve the technical problem that under complex working conditions, the AMR robot may experience feature mismatch in the positioning system due to deviation between the motion posture and the planned expectations, thereby causing incorrect posture updates and ultimately affecting the robot's safety and operating efficiency.

[0006] The technical solution of this application is as follows: In a first aspect, the present application discloses an AMR robot path control method, comprising: Obtain the motion status information of the AMR robot and the current consumption information of the drive motor; The turning radius of the AMR robot is calculated based on the motion state information. Based on the turning radius and the current consumption information of the drive motor, it is determined whether the AMR robot is in a high-risk positioning condition and a high-risk judgment result is obtained; Adjust the matching quality threshold for the positioning system to accept location information based on high-risk judgment results; Based on the adjusted matching quality threshold, when receiving the matching score corresponding to the lidar data, it determines whether to update the AMR robot's posture and performs control actions based on the judgment result; After the high-risk positioning condition is resolved, the matching quality threshold of the positioning system is restored.

[0007] Through the technical solution, this application can dynamically identify the positioning risks that the AMR robot may face, and adjust the matching strategy of the positioning system accordingly, effectively avoiding accepting erroneous posture updates under risky working conditions, thereby significantly improving the positioning accuracy and operational safety of the AMR robot, and overcoming the problem of feature mismatching caused by "over-swing" in the existing technology.

[0008] Furthermore, based on the above method, the steps of determining whether the AMR robot is in a high-risk positioning condition and obtaining a high-risk determination result include: Calculate the turning radius of the AMR robot based on the motion state information; Get the AMR robot's travel speed and / or robot load; Adjust and determine the turning radius threshold based on driving speed and / or robot load; Obtain the friction characteristics of the floor where the AMR robot is located; Adjusting and determining a current consumption threshold value according to driving speed and / or friction characteristics; Based on the turning radius, the current consumption information of the driving motor, the turning radius threshold, and the current consumption threshold, it is determined whether the AMR robot is in a high-risk positioning condition and a high-risk judgment result is obtained.

[0009] Through the technical solution, this application comprehensively considers multiple factors such as turning radius, driving speed, robot load, floor friction characteristics, and drive motor current consumption, and dynamically adjusts the corresponding thresholds, making the judgment of high-risk positioning conditions more comprehensive, accurate and adaptive, thereby providing a more reliable basis for subsequent matching quality threshold adjustments.

[0010] In some preferred embodiments, based on the above method, the step of adjusting the matching quality threshold for the positioning system to accept location information includes: Obtain the turning radius and drive motor current consumption information of the AMR robot; Determine the turning risk level based on the turning radius; Determine the load risk level based on the current consumption information of the drive motor; Determine the comprehensive risk level of the high-risk judgment result based on the turning risk level and the load risk level; According to the comprehensive risk level, a matching quality threshold is selected as the matching quality threshold for the positioning system to accept location information.

[0011] Through technical solutions, this application refines the high-risk judgment results into turning risk level and load risk level, and further determines the comprehensive risk level, so that the adjustment of the matching quality threshold can be finely selected according to the type and degree of risk, thereby achieving more flexible and effective positioning system control.

[0012] Furthermore, based on the above method, the steps of selecting a matching quality threshold according to the comprehensive risk level as the matching quality threshold for the positioning system to accept location information include: Get the positioning feature quality of the AMR robot's current environment; Obtain the accuracy requirements of the AMR robot's current task; Based on the comprehensive risk level, positioning feature quality and accuracy requirements, the matching quality threshold is calculated as the matching quality threshold for the positioning system to accept location information.

[0013] Through technical solutions, this application not only considers the comprehensive risk level when selecting the matching quality threshold, but also incorporates the positioning feature quality of the current environment and the accuracy requirements of the task, making the setting of the matching quality threshold more intelligent and scenario-based, ensuring that the best positioning performance can be maintained under different environments and task requirements.

[0014] Based on the above, the present application further proposes that, based on the above method, based on the adjusted matching quality threshold, upon receiving the matching score corresponding to the lidar data, the steps of determining whether to update the posture of the AMR robot and executing the control action according to the determination result include: When receiving the matching score corresponding to the lidar data, the current matching score is compared with the adjusted matching quality threshold to obtain the matching comparison result; If the matching comparison result indicates that the current matching score is higher than or equal to the matching quality threshold, the position of the AMR robot is updated; If the matching comparison result indicates that the current matching score is lower than the matching quality threshold, the AMR robot's posture will not be updated; When the AMR robot's posture is not updated, the positioning confidence value of the AMR robot is evaluated based on the duration of posture update failure and the uncertainty of the current posture estimate; Adjust the maximum driving speed of the AMR robot according to the positioning confidence value; Adjust the AMR robot's path planning strategy based on the positioning confidence value to guide the AMR robot to areas with rich positioning features; If the positioning confidence value is lower than the preset threshold and reaches the preset time, the AMR robot's safe parking behavior is triggered or an assistance request is sent to the central dispatch system.

[0015] Through the technical solution, this application does not blindly update the posture when the matching score is lower than the threshold, but further evaluates the positioning confidence value and takes multi-level response measures based on the confidence value, including adjusting the driving speed, guiding to feature-rich areas, and even triggering safe parking or requesting assistance, thereby building a comprehensive risk response mechanism, effectively avoiding "confidence errors" and improving the robot's autonomous safety capabilities.

[0016] Preferably, based on the above method, when the pose of the AMR robot is not updated, the step of evaluating the positioning confidence value of the AMR robot according to the duration of the pose update failure and the uncertainty of the current pose estimation includes: Get the count of failed pose updates; Get the covariance matrix of the current pose estimate; Get the particle diffusion degree of the positioning system particle filter; The positioning confidence value of the AMR robot is comprehensively evaluated based on the count of pose update failures, covariance matrix, and particle diffusion degree.

[0017] Through the technical solution, this application comprehensively evaluates the positioning confidence value of the AMR robot by combining multiple indicators such as the pose update failure count, covariance matrix and particle diffusion degree, making the judgment of the confidence value more accurate and robust, and providing a more reliable basis for subsequent risk response measures.

[0018] In one embodiment, based on the above method, the step of adjusting the maximum driving speed of the AMR robot according to the positioning confidence value includes: According to the positioning confidence value, the maximum driving speed of the AMR robot is calculated through a preset function relationship.

[0019] Through the technical solution, this application can dynamically and smoothly adjust the maximum driving speed of the AMR robot according to the changes in the positioning confidence value, thereby automatically reducing the speed when the positioning uncertainty is high, effectively reducing the potential collision risk, and improving operational safety.

[0020] In another embodiment, based on the above method, the step of adjusting the path planning strategy of the AMR robot according to the positioning confidence value to guide the AMR robot to an area rich in positioning features includes: Adjust the cost weights of different path segments in the path planning algorithm based on the positioning confidence value; Based on the adjusted cost weights, a path planning strategy is generated for the AMR robot to move to areas with rich positioning features.

[0021] Through the technical solution, this application can actively adjust path planning according to the positioning confidence value, guide the AMR robot to prioritize areas with richer positioning features, thereby increasing the probability of successful repositioning and helping the robot to return to a high-confidence positioning state more quickly.

[0022] As an optional solution, based on the above method, the steps of adjusting the cost weights of different path segments in the path planning algorithm according to the positioning confidence value include: Read the positioning confidence value; According to the positioning confidence value, the cost weights of different path segments are calculated through preset function rules.

[0023] Through the technical solution, this application establishes a quantitative relationship between the positioning confidence value and the cost weight of the path segment through preset function rules, making the adjustment of path planning more systematic and controllable, ensuring that the robot can be effectively guided to areas with rich positioning features.

[0024] In a second aspect, the present application further discloses an AMR robot path control system for performing AMR robot path control, comprising: The state information acquisition module is used to obtain the motion state information of the AMR robot and the current consumption information of the drive motor; The state risk judgment module is used to calculate the turning radius of the AMR robot based on the motion state information, and determine whether the AMR robot is in a high-risk positioning condition based on the turning radius and the current consumption information of the drive motor, thereby obtaining a high-risk judgment result; A matching threshold adjustment module is used to adjust the matching quality threshold of the positioning system for receiving location information based on high-risk judgment results; The pose update control module is used to determine whether to update the pose of the AMR robot based on the adjusted matching quality threshold when receiving the matching score corresponding to the lidar data, and perform control actions based on the judgment result; The matching threshold recovery module is used to restore the matching quality threshold of the positioning system after the high-risk positioning condition is resolved.

[0025] Through the technical solution, this application provides a system that can implement the above-mentioned AMR robot path control method. Through modular design, the identification of high-risk positioning conditions, dynamic adjustment of matching thresholds, and control of posture updates can be effectively integrated and automatically executed, providing hardware and software support for the safe and stable operation of AMR robots in complex environments.

[0026] Beneficial effects

[0027] The AMR robot path control method disclosed in this application proposes an innovative solution to the existing problem of AMR robots overswinging when performing challenging turns due to the combination of compensatory output and extreme path planning, which in turn leads to "feature mismatching" and incorrect position updates in the positioning system. This method dynamically calculates the turning radius by acquiring real-time information about the AMR robot's motion state and drive motor current consumption. Combined with this current consumption information, it accurately determines whether the AMR robot is in a high-risk positioning condition. Once a high-risk condition is identified, the system immediately adjusts the matching quality threshold for accepting position information, thereby increasing the matching score requirement for lidar data. This means that in risky conditions with high positioning uncertainty or a high risk of mismatching, the system will accept position updates more cautiously, effectively avoiding erroneous updates to the AMR robot's position and posture, effectively preventing "confidence errors." Once the high-risk condition is resolved, the system can promptly restore the matching quality threshold to ensure normal operation and efficiency. Through this dynamic and adaptive matching quality threshold adjustment mechanism, this application can significantly improve the positioning robustness and safety of AMR robots in complex and dynamic environments, effectively prevent traffic conflicts or secondary collisions with facilities caused by posture errors, thereby ensuring the stable operation and operating efficiency of the AMR system. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a flow chart of a method for controlling an AMR robot path in one embodiment of the present invention; Figure 2 This is one of the method flow charts of a path control method for an AMR robot in another embodiment of the present invention; Figure 3 This is a second flow chart of a method for controlling an AMR robot path in another embodiment of the present invention; Figure 4 This is a third flow chart of a method for controlling an AMR robot path in another embodiment of the present invention; Figure 5 This is a fourth flow chart of a method for controlling an AMR robot path in another embodiment of the present invention; Figure 6 This is a system block diagram of an AMR robot path control system in another embodiment of the present invention; Description of reference numerals: 1. AMR robot path control system; 11. Status information acquisition module; 12. Status risk judgment module; 13. Matching threshold adjustment module; 14. Posture update control module; 15. Matching threshold recovery module. DETAILED DESCRIPTION

[0029] The technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. The components of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0030] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0031] In modern smart warehousing environments, autonomous mobile robot (AMR) systems are widely used for material handling, and the accuracy of their path control is crucial. When conventional AMR robots operate on mixed floors with varying friction coefficients for extended periods, their drive wheels may experience asymmetric wear, causing the path control system to continuously output compensatory control variables. When an AMR robot performs challenging turns in narrow paths, the combined effect of compensatory output and extreme path planning can cause a slight "overswing" in the actual turning trajectory, leading to "feature mismatches" in the positioning algorithm. This mismatch causes the positioning system to incorrectly update the AMR robot's position, causing its actual position to deviate significantly from its correct position on the map. This can easily lead to potential traffic conflicts or secondary collisions with equipment, seriously impacting the safety and operational efficiency of the AMR system.

[0032] In this regard, this application proposes an AMR robot path control method, combining Figure 1 Shown, including: S1, obtain the motion state information of the AMR robot and the current consumption information of the driving motor; S2, calculating the turning radius of the AMR robot based on the motion state information, and judging whether the AMR robot is in a high-risk positioning condition based on the turning radius and the current consumption information of the drive motor, thereby obtaining a high-risk judgment result; S3, based on the high-risk judgment result, adjust the matching quality threshold of the positioning system to accept the location information; S4, based on the adjusted matching quality threshold, upon receiving the matching score corresponding to the lidar data, determines whether to update the position and posture of the AMR robot, and executes a control action based on the determination result; S5, after the high-risk positioning condition is resolved, the matching quality threshold of the positioning system is restored.

[0033] To facilitate a clearer and easier understanding of the technical solutions of this application, some key terms are explained below. An AMR robot is an autonomous mobile robot, a mobile platform capable of autonomously navigating, avoiding obstacles, and performing tasks in complex environments. Motion state information refers to real-time data such as the speed, acceleration, angular velocity, and heading of an AMR robot during operation. Drive motor current consumption information refers to the actual current consumed by the AMR robot's drive motor during operation. This data reflects the impact of factors such as motor load, ground friction, and wheel wear on motor output. Turning radius refers to the radius of the circular arc trajectory followed by the AMR robot's center of mass during a turn, reflecting the abruptness of the turn. A high-risk positioning condition refers to a risky state in which, under specific operating conditions, the AMR robot's positioning system is prone to feature mismatching, resulting in inaccurate pose estimation. The positioning system refers to the system used by the AMR robot to determine its precise position and pose in the environment. It typically integrates data from multiple sensors, such as lidar, an inertial measurement unit (IMU), and an odometry meter. The matching quality threshold refers to the minimum acceptable score used by the positioning system to assess the degree of match between current sensor data and map features when processing sensor data. LiDAR data refers to the three-dimensional point cloud data generated by the LiDAR sensor after scanning the surrounding environment, which contains information about the geometric features of the environment. The matching score is a quantitative indicator used by the positioning algorithm to evaluate the degree of fit after matching the current LiDAR data with pre-built map features. Posture refers to a comprehensive description of the position and posture of the AMR robot in three-dimensional space. Control action refers to the corresponding behavior taken by the AMR robot based on the positioning results and risk assessment. This method is commonly used for AMR robots in scenarios such as smart warehousing and factory automation. These scenarios may include mixed floors, narrow passages, and high-density storage areas, which place high demands on the path control accuracy and safety of the AMR robot.

[0034] The AMR robot path control method proposed in this application may include the following steps in its specific implementation.

[0035] First, obtain the AMR robot's motion state information and drive motor current consumption information. Motion state information can include the AMR robot's real-time speed, acceleration, angular velocity, and heading data. This information can be collected using the AMR robot's internal inertial measurement unit, encoder, or visual sensor. For example, speed information can be calculated by combining the drive wheel encoder pulse count with the wheel radius; angular velocity and heading information can be provided by the inertial measurement unit. Drive motor current consumption information refers to the actual current consumed by the AMR robot's drive motor during operation. This information can be monitored and obtained in real time by installing current sensors in the drive motor circuit. These sensors convert the collected analog signals into digital signals and transmit them to the AMR robot's main controller for processing.

[0036] Next, the AMR robot's turning radius is calculated based on the motion state information. Based on the turning radius and the drive motor current consumption information, the AMR robot is judged to be in a high-risk positioning condition, resulting in a high-risk judgment result. After obtaining the motion state information, the AMR robot's turning radius can be calculated based on this information. For example, the instantaneous turning radius can be calculated using the AMR robot's real-time linear and angular velocity. The turning radius is equal to the linear velocity divided by the angular velocity. Subsequently, the calculated turning radius and the drive motor current consumption information are combined to determine whether the AMR robot is in a high-risk positioning condition. One implementation method is to preset a series of turning radius and current consumption thresholds. When the calculated turning radius is less than a preset threshold (indicating a sharp turn) and the drive motor current consumption information is higher than a preset threshold (indicating high load or abnormal friction), the AMR robot is judged to be in a high-risk positioning condition. The high-risk judgment result can be a Boolean value (yes / no) or a risk level.

[0037] Furthermore, based on the high-risk judgment result, the matching quality threshold for the positioning system to accept location information is adjusted. Once a high-risk judgment result is obtained, the matching quality threshold for the positioning system to accept location information will be adjusted accordingly. For example, if the high-risk judgment result indicates that the AMR robot is in a high-risk positioning condition, the matching quality threshold can be raised. This means that the positioning system will become more "picky" when accepting new location information, and only data that has a very high degree of match with map features will be adopted. Conversely, if it is judged to be a non-high-risk condition, the matching quality threshold can be maintained or lowered. This adjustment can be achieved by consulting a preset lookup table or applying a simple functional relationship, which maps different risk levels to different matching quality threshold values.

[0038] Therefore, based on the adjusted matching quality threshold, when the matching score corresponding to the LiDAR data is received, the system determines whether to update the AMR robot's pose and executes control actions based on the judgment result. After the matching quality threshold is adjusted, when the AMR robot receives LiDAR data and calculates the corresponding matching score, the system will determine whether to update the AMR robot's pose based on this adjusted threshold. Specifically, if the current matching score is higher than or equal to the adjusted matching quality threshold, the positioning data is considered reliable, and the AMR robot's pose will be updated. If the current matching score is lower than the adjusted matching quality threshold, the positioning data is considered unreliable, and the AMR robot's pose will not be updated to avoid introducing incorrect pose estimates. Based on this judgment result, the AMR robot will execute the corresponding control action. For example, if the pose is not updated, a series of safety measures can be triggered, such as deceleration and warnings.

[0039] Finally, after the high-risk positioning condition is resolved, the positioning system's matching quality threshold is restored. Once the AMR robot exits the high-risk positioning condition, the positioning system's matching quality threshold is restored to its normal or preset default value. For example, this can be achieved by continuously monitoring the AMR robot's motion state information and drive motor current consumption information. When this information no longer meets the judgment criteria for the high-risk condition, the high-risk positioning condition is considered to have been resolved. At this point, the system will automatically adjust the matching quality threshold back to a lower level that allows for more relaxed matching, ensuring that the AMR robot can efficiently update its position under normal operating conditions and maintain its normal navigation and task execution capabilities.

[0040] Optional, combined Figure 2 As shown, S2 calculates the turning radius of the AMR robot based on the motion state information, and determines whether the AMR robot is in a high-risk positioning condition based on the turning radius and the current consumption information of the drive motor. The steps of obtaining the high-risk judgment result include: S21, calculating the turning radius of the AMR robot according to the motion state information; S22, obtaining the driving speed and / or robot load of the AMR robot; S23, adjusting and determining a turning radius threshold according to the driving speed and / or robot load; S24, obtaining the friction characteristics of the floor where the AMR robot is located; S25, adjusting and determining a current consumption threshold value according to the driving speed and / or friction characteristics; S26, judging whether the AMR robot is in a high-risk positioning condition based on the turning radius, the driving motor current consumption information, the turning radius threshold, and the current consumption threshold, and obtaining a high-risk judgment result.

[0041] Specifically, when determining whether an AMR robot is in a high-risk positioning condition, in addition to considering the turning radius and drive motor current consumption information, the AMR robot's driving speed and / or robot load, as well as the friction characteristics of the floor on which the AMR robot is located, are also introduced. Obtaining the AMR robot's driving speed and / or robot load is intended to reflect the dynamic characteristics and load-bearing capacity of the AMR robot's current operation. Driving speed directly affects the AMR robot's inertia, while the robot load affects its overall mass and center of gravity, both of which have a significant impact on the AMR robot's stability when turning. Based on this, the turning radius threshold can be dynamically adjusted and determined based on the driving speed and / or robot load. For example, when the AMR robot's driving speed is high or the load is heavy, its safe turning radius will increase accordingly. In this case, the turning radius threshold should be set to a higher value to more strictly assess the turning risk.

[0042] Furthermore, the friction characteristics of the floor on which the AMR robot is located are acquired to more accurately assess the risk reflected by the drive motor current draw. Floor friction characteristics, such as the friction coefficient, directly impact the driving force required for the AMR robot to navigate and turn, as well as the potential for slipping or drifting. For example, on a slippery floor with a low friction coefficient, even a small turn or load can cause an abnormal increase in the drive motor current draw, indicating a potential risk of instability. Therefore, the current draw threshold can be adjusted and determined based on the driving speed and / or friction characteristics. For example, on floors with a low friction coefficient, the current draw threshold should be appropriately lowered to more sensitively detect abnormal current draw, thereby identifying high-risk operating conditions earlier. Ultimately, by comprehensively considering turning radius, drive motor current draw information, and dynamically adjusting the turning radius threshold and current draw threshold, a more comprehensive and accurate assessment of whether the AMR robot is in a high-risk positioning condition can be achieved, resulting in a more reliable high-risk determination.

[0043] In some preferred embodiments, a specific example is provided below. Assume that an AMR robot is performing material handling tasks in a warehouse. At a certain moment, the AMR robot is traveling at a speed of 5 m / s and carrying 500 kg of cargo. The floor it is on is made of epoxy resin with a friction coefficient of 0.6.

[0044] First, the system calculates the current turning radius as 2 meters based on the AMR robot's motion state information (such as wheel speed difference and heading angle change rate).

[0045] The system then determines the AMR's speed of 5 meters per second and its payload of 500 kilograms. Based on a pre-set function, the system dynamically adjusts the turning radius threshold to 3 meters, taking into account the high speed and heavy load.

[0046] At the same time, the system detects that the floor's friction characteristic is 0.6. Based on the driving speed, the system dynamically adjusts the drive motor current consumption threshold to 15 amps.

[0047] At this point, the system detects that the AMR robot's drive motor current consumption is 20 amps.

[0048] By comparison, the current turning radius of 2 meters is smaller than the adjusted turning radius threshold of 3 meters, indicating that the turn is relatively sharp; at the same time, the drive motor current consumption of 20 amperes is higher than the adjusted current consumption threshold of 15 amperes, indicating that the motor load is heavy or there is a risk of slipping.

[0049] Based on the above analysis, the system determines that the AMR robot is currently in a high-risk positioning condition. Based on this high-risk judgment, the positioning system will further adjust the matching quality threshold for the position information it receives to address potential positioning uncertainties and take appropriate control measures, such as reducing driving speed or adjusting path planning strategies, to ensure the safe and stable operation of the AMR robot.

[0050] Optional, combined Figure 3 As shown, S3 adjusts the matching quality threshold of the positioning system for accepting location information based on the high-risk judgment result, including the following steps: S31, obtaining the turning radius and driving motor current consumption information of the AMR robot; S32, determining a turning risk level based on the turning radius; S33, determining a load risk level based on the current consumption information of the drive motor; S34, determining a comprehensive risk level of the high risk judgment result based on the turning risk level and the load risk level; S35, selecting a matching quality threshold based on the comprehensive risk level as the matching quality threshold for the positioning system to accept location information.

[0051] Specifically, obtaining the AMR robot's turning radius and drive motor current consumption information involves the system continuously monitoring the AMR robot's real-time motion data, such as wheel speed and angular velocity, and calculating the current turning radius. Simultaneously, drive motor current consumption information, such as motor current and power, is also collected in real time to reflect the AMR robot's load status. This information forms the basis for subsequent risk assessment. Determining the turning risk level based on the turning radius can be understood as categorizing the AMR robot into different risk levels based on its current turning radius. For example, a smaller turning radius indicates a sharper turn and a higher risk of positioning misalignment, thus assigning a higher turning risk level. This level can be determined based on a preset threshold range or functional relationship. In practical applications, determining the load risk level based on drive motor current consumption information involves analyzing the drive motor's current consumption to assess the load level currently borne by the AMR robot. For example, a significant increase in current consumption may indicate that the AMR robot is carrying a heavy load or traveling on uneven surfaces, which increases positioning difficulty and uncertainty, thus assigning a higher load risk level. This level can also be determined based on a preset threshold range or model. Furthermore, the comprehensive risk level of the high-risk judgment result is determined based on the turning risk level and the load risk level, with the aim of fusing the two independent risk assessment results to obtain a more comprehensive and accurate risk measurement. For example, the turning risk level and the load risk level can be combined into a comprehensive risk value or level through weighted averaging, table lookup, or fuzzy logic. Thus, a matching quality threshold is selected based on the comprehensive risk level as the matching quality threshold for the positioning system to accept position information. This means that the positioning system no longer simply adjusts the threshold with a "yes" or "no" answer, but dynamically selects a suitable matching quality threshold based on the level of the comprehensive risk level. For example, the higher the comprehensive risk level, the higher the selected matching quality threshold, requiring the quality of the positioning match to be more stringent to ensure the reliability of positioning under risky conditions.

[0052] Some preferred embodiments are described below using a specific example. Suppose an AMR robot is making a sharp turn in a narrow passage while carrying a heavy load. First, the system obtains the AMR robot's motion state information and calculates its turning radius to be 0.5 meters. According to preset rules, a turning radius less than 1 meter is defined as a high turning risk, and thus the turning risk level is determined to be "high." Simultaneously, the system obtains information about the drive motor's current consumption and finds that the current consumption reaches 90% of the rated current, significantly higher than the unloaded current. According to preset rules, a current consumption exceeding 80% is defined as a high load risk, and thus the load risk level is determined to be "high." Next, based on the "high" turning risk level and the "high" load risk level, the system determines the overall risk level to be "very high" through table lookup or weighted calculation. Finally, based on the "very high" overall risk level, the positioning system selects a very strict matching quality threshold, for example, raising the matching score requirement from 0.7 to 0.9. This means that the AMR robot's position will only be updated when the matching score between the lidar data and the map reaches 0.9 or higher. If the matching score falls below this threshold, even a slight deviation will suppress the pose update, prompting the system to take further safety measures, such as slowing down or seeking assistance, to avoid potential positioning errors. In this way, the AMR robot can maintain a high level of positioning vigilance in high-risk working conditions to ensure safe operation.

[0053] Optionally, the steps of selecting a matching quality threshold based on the comprehensive risk level as the matching quality threshold for the positioning system to accept location information include: Get the positioning feature quality of the AMR robot's current environment; Obtain the accuracy requirements of the AMR robot's current task; Based on the comprehensive risk level, positioning feature quality and accuracy requirements, the matching quality threshold is calculated as the matching quality threshold for the positioning system to accept location information.

[0054] Specifically, positioning feature quality refers to the abundance, uniformity, and uniqueness of the visual or LiDAR feature points that can be used for positioning in the AMR robot's current environment. This can be assessed by analyzing metrics such as the match between the current LiDAR scan data and the pre-built map, feature point density, uniformity of feature point distribution, and uniqueness. For example, in aisles or on-shelf areas, positioning features are typically abundant and unique, resulting in high quality. However, in open areas or areas with repeated features, positioning feature quality may be lower.

[0055] Accuracy requirements can be understood as the specific positioning accuracy requirements for the task the AMR robot is currently performing. For example, when an AMR robot is performing tasks such as precise docking, charging, or object grasping, its accuracy requirements are higher; while when it is simply patrolling a large area or navigating a path, its accuracy requirements are relatively lower. Accuracy requirements can be pre-configured in the mission planning system or dynamically acquired based on the mission type and current mission phase.

[0056] In practice, the matching quality threshold is calculated using a pre-set algorithm or function model based on the overall risk level, positioning feature quality, and accuracy requirements. For example, these three input parameters can be mapped to a specific matching score threshold using weighted averaging, fuzzy logic reasoning, or machine learning models. This calculation process ensures that the determined matching quality threshold reflects the AMR robot's current risk status while also taking into account the environmental positioning conditions and the actual requirements of the task.

[0057] In some preferred embodiments, a specific example is provided below to illustrate: Assume that an AMR robot is performing a task in a large warehouse.

[0058] Scenario 1: An AMR robot is navigating an area with densely packed shelves and rich features while simultaneously performing a precise pallet-grabbing task. In this scenario, the system assesses a medium overall risk level (for example, the robot is slowly turning), high positioning feature quality, and high precision requirements. According to the solution in this application, the calculation of the matching quality threshold takes all three factors into account. Due to the high precision requirement, even if the overall risk level is not high, the system calculates a relatively high matching quality threshold (for example, 0.85) to ensure extremely high accuracy of pose updates and, therefore, the success rate of the pallet-grabbing task.

[0059] Scenario 2: The AMR robot is patrolling long distances in an open area with sparse features. At this point, the system assesses the overall risk level as high (for example, the robot is traveling at high speed and making large-radius turns), the positioning feature quality is low, and the accuracy requirement is low. In this case, if only based on the high risk level, a very high matching quality threshold may be set, resulting in frequent failures of pose updates. However, according to the solution of the present application, due to the low quality of the positioning features and the low accuracy requirement, the system calculates a moderate matching quality threshold (for example, 0.70). This enables the AMR robot to maintain a certain pose update capability in a feature-sparse environment, avoiding frequent triggering of positioning failure processing due to overly strict thresholds, thereby maintaining its continuous operation capability.

[0060] Optional, combined Figure 4As shown, S4 determines whether to update the posture of the AMR robot based on the adjusted matching quality threshold when receiving the matching score corresponding to the lidar data, and performs a control action according to the judgment result, including the following steps: S41, upon receiving the matching score corresponding to the lidar data, comparing the current matching score with the adjusted matching quality threshold to obtain a matching comparison result; S42, if the matching comparison result indicates that the current matching score is higher than or equal to the matching quality threshold, then updating the position and posture of the AMR robot; S43, if the matching comparison result indicates that the current matching score is lower than the matching quality threshold, the position and posture of the AMR robot are not updated; S44, when the posture of the AMR robot is not updated, evaluating the positioning confidence value of the AMR robot according to the duration of the posture update failure and the uncertainty of the current posture estimation; S45, adjusting the maximum driving speed of the AMR robot according to the positioning confidence value; S46, adjusting the path planning strategy of the AMR robot according to the positioning confidence value to guide the AMR robot to an area rich in positioning features; S47: If the positioning confidence value is lower than the preset threshold and reaches the preset time, the AMR robot's safe parking behavior is triggered or an assistance request is sent to the central dispatch system.

[0061] Specifically, when the AMR robot receives lidar data and calculates the corresponding matching score, the matching score is immediately compared with the currently adjusted matching quality threshold. This matching quality threshold is dynamically set based on the high-risk judgment results to ensure the reliability of positioning information under different working conditions. If the current matching score is higher than or equal to the matching quality threshold, it is considered that the matching quality between the current lidar data and the map is high enough, and the AMR robot's posture will be updated to reflect its latest position and posture. Conversely, if the matching score is lower than the matching quality threshold, it indicates that the current matching quality is insufficient to support reliable posture updates. At this time, the AMR robot's posture will not be updated to avoid introducing inaccurate positioning information.

[0062] Furthermore, when the AMR robot's pose is not updated, the system initiates a series of mechanisms to assess and manage positioning risks. Specifically, the AMR robot's positioning confidence value is evaluated based on the duration of the pose update failure and the uncertainty of the current pose estimate. The duration of the pose update failure can be a counter that records the number of frames or the length of time during which the pose update has failed. The uncertainty of the current pose estimate is usually represented by the covariance matrix or particle diffusion output by the positioning algorithm (such as the Kalman filter or particle filter). The larger the trace or determinant of the covariance matrix, the higher the particle diffusion, indicating greater uncertainty. The positioning confidence value is a comprehensive indicator that reflects the AMR robot's trust in its current position estimate. The higher the value, the more reliable the positioning, and vice versa.

[0063] Based on the assessed positioning confidence value, the AMR robot's behavior will be adjusted accordingly. First, the AMR robot's maximum driving speed will be dynamically adjusted. When the positioning confidence value decreases, the maximum driving speed will be reduced to reduce the risk caused by inaccurate positioning. For example, a preset functional relationship can be used to map the positioning confidence value to the maximum allowable driving speed. Second, the AMR robot's path planning strategy will also be adjusted to guide the AMR robot to areas rich in positioning features. This means that the path planning algorithm will prioritize areas with more identifiable positioning features (such as walls, pillars, fixed obstacles, etc.) in order to improve subsequent matching scores and positioning success rates. For example, a term related to the density of positioning features can be added to the path planning cost function to reduce the cost of feature-rich path segments.

[0064] Furthermore, as a safety measure, if the positioning confidence value remains below a preset threshold for a predetermined period, the system triggers a safe stop for the AMR robot, immediately halting its operation to avoid potential danger. Alternatively, the system sends an assistance request to the central dispatch system, notifying the operator or a higher-level control system to request manual intervention or additional positioning assistance information. The preset threshold and duration can be configured based on the actual application scenario and safety requirements.

[0065] In some preferred embodiments, a specific example is provided below. Consider an AMR robot performing a handling task in a large warehouse. When the robot moves from a feature-rich area (e.g., an aisle with numerous shelves and walls) into an open, feature-sparse area (e.g., a large, empty loading and unloading area), the matching score between its lidar data and the pre-set map may begin to decline.

[0066] Specifically, when the matching score falls below the matching quality threshold adjusted based on the high-risk judgment result, the AMR robot's pose will not be updated. At this point, the system begins to evaluate the positioning confidence value. If the number of pose update failures continues to increase and the trace of the covariance matrix of the current pose estimate gradually increases, indicating that positioning uncertainty is increasing, the positioning confidence value will decrease accordingly.

[0067] For example, when the positioning confidence value drops from a high level (such as 0.9) to a medium level (such as 0.6), the system immediately adjusts the AMR robot's maximum travel speed from a normal travel speed (such as 1.5 meters per second) to a safe speed (such as 0.8 meters per second). Simultaneously, the path planning module adjusts the cost weights of different path segments in the path planning algorithm based on the decreased positioning confidence value, reducing the cost of paths to aisles with more nearby shelves or walls. This guides the AMR robot to prioritize these areas with rich positioning features, aiming to quickly restore high matching quality.

[0068] If the positioning confidence value further decreases, for example, falling below a preset threshold (such as 0.3) for a preset duration (such as 5 seconds), the system immediately triggers the AMR robot's safe stop behavior, causing it to stop in place and simultaneously sending an assistance request to the central dispatch system, notifying the robot of the abnormal positioning status and requesting manual intervention or remote assistance. Through this series of progressive control actions, the AMR robot can proactively take measures to reduce risks when positioning quality deteriorates and perform a safe shutdown when necessary, effectively avoiding potential accidents.

[0069] Optional, combined Figure 5 As shown, when the posture of the AMR robot is not updated, the step of evaluating the positioning confidence value of the AMR robot according to the duration of the posture update failure and the uncertainty of the current posture estimation includes: S441, obtaining the count of failed pose updates; S442, obtaining the covariance matrix of the current pose estimation; S443, obtaining the particle diffusion degree of the positioning system particle filter; S444: Comprehensively evaluate the positioning confidence value of the AMR robot based on the count of failed pose updates, the covariance matrix, and the degree of particle diffusion.

[0070] The pose update failure count refers to the number of times the AMR robot fails to successfully update its pose over a continuous period of time. This count can intuitively reflect the persistent failure of the positioning system in the current environment. A higher count generally indicates more severe positioning issues. The covariance matrix of the current pose estimate is a mathematical tool that describes the uncertainty of the AMR robot's current pose estimate. Its diagonal elements represent the variance of each pose dimension (such as X, Y, and heading angle), while the off-diagonal elements represent the covariance between these dimensions. A larger trace or determinant of the covariance matrix indicates a higher uncertainty in the current pose estimate, which in turn indicates poorer positioning accuracy. The particle diffusion degree of the positioning system's particle filter refers to the degree of dispersion of a large number of randomly sampled particles in the particle filter in the pose space. The greater the particle dispersion, the less certain the positioning system's estimate of the current pose, and vice versa. In practical applications, the positioning confidence value can be defined as a comprehensive indicator that integrates the above three parameters through weighted averaging, fuzzy logic reasoning, or machine learning models. For example, a function can be set up that takes the count of failed pose updates, the trace of the covariance matrix, and the degree of particle diffusion as input and outputs a confidence value between 0 and 1, where 1 indicates complete confidence and 0 indicates complete distrust. The goal is to provide a quantitative, multi-dimensional positioning reliability indicator.

[0071] In some preferred embodiments, a specific example is provided below. Suppose an AMR robot is navigating a long corridor where positioning features are relatively sparse. When the robot passes through an area with fewer feature points, the matching score between its lidar data and the map may consistently fall below the matching quality threshold, resulting in pose update failures. At this point, the system begins accumulating a "pose update failure count." Simultaneously, due to the lack of sufficient positioning information, the uncertainty of the robot's internal pose estimate (e.g., using an extended Kalman filter or particle filter) gradually increases, which is reflected in an increase in the trace value of the "covariance matrix of the current pose estimate." If the positioning system uses a particle filter, as uncertainty increases, the "particle diffusion" also increases, and the particles become more dispersed. The system inputs these three metrics (e.g., pose update failure count reaches 5, covariance matrix trace value exceeds a preset threshold A, and particle diffusion exceeds a preset threshold B) into a pre-set evaluation function, which may calculate a comprehensive positioning confidence value through a weighted summation. For example, confidence value = W1 (1 / count)+W2 (1 / trace value)+W3 (1 / diffusion degree), where W1, W2, and W3 are weights. If the calculated confidence value falls below a preset threshold, the system will determine that the AMR robot's positioning reliability is low and immediately take appropriate risk control measures, such as reducing the driving speed to a safe speed or adjusting the path planning to quickly move away from the feature-sparse area and toward an area with rich positioning features, thereby effectively avoiding safety accidents caused by positioning failure.

[0072] Optionally, the steps of adjusting the maximum driving speed of the AMR robot according to the positioning confidence value include: According to the positioning confidence value, the maximum driving speed of the AMR robot is calculated through a preset function relationship.

[0073] The pre-set function can be understood as a mathematical model or set of rules that maps the AMR robot's positioning confidence value to its maximum travel speed. This function can be designed and configured based on the needs of the actual application scenario. For example, it can be a linear function, a nonlinear function, a piecewise function, a lookup table, or a set of rules based on fuzzy logic. Its purpose is to ensure that the AMR robot can travel at a faster speed when the positioning confidence value is high, and reduce speed to improve safety when the positioning confidence value is low.

[0074] In some preferred embodiments, the preset functional relationship can be configured as a piecewise function. For example, when the positioning confidence value of the AMR robot is higher than 0.8, the maximum driving speed is set to 1.5 m / s; when the positioning confidence value is between 0.5 and 0.8 (including 0.5 but excluding 0.8), the maximum driving speed is set to 1.0 m / s; when the positioning confidence value is lower than or equal to 0.5, the maximum driving speed is set to 0.5 m / s. This piecewise function can be stored in the controller of the AMR robot and called after each evaluation of the positioning confidence value to calculate the new maximum driving speed. As a specific embodiment, the preset functional relationship can also be a continuous nonlinear function, such as an S-shaped curve function, so that the maximum driving speed smoothly transitions from the minimum value to the maximum value as the positioning confidence value increases, thereby providing a smoother speed adjustment experience.

[0075] Optionally, the steps of adjusting the path planning strategy of the AMR robot based on the positioning confidence value to guide the AMR robot to an area rich in positioning features include: Adjust the cost weights of different path segments in the path planning algorithm based on the positioning confidence value; Based on the adjusted cost weights, a path planning strategy is generated for the AMR robot to move to areas with rich positioning features.

[0076] Specifically, adjusting the cost weights of different path segments in the path planning algorithm means dynamically modifying the path planning algorithm (e.g., AMR) based on the current positioning confidence value of the AMR robot. The weighting parameters used in algorithms (such as the CNN algorithm, Dijkstra algorithm, and RRT algorithm) to evaluate the costs of different path segments are used. When the positioning confidence value is low, the cost weights of path segments associated with areas rich in positioning features can be reduced, while the cost weights of path segments associated with areas sparse in positioning features can be increased. For example, a positioning feature density map of the environment map can be pre-built, and initial costs assigned to path segments in different areas based on this density map. As the positioning confidence value decreases, the system further reduces the cost of path segments leading to areas with high positioning feature density, while increasing the cost of path segments leading to areas with low positioning feature density, based on a preset function or lookup table. Based on these adjusted cost weights, a path planning strategy for the AMR robot is generated to guide it to areas rich in positioning features. This means that when calculating the optimal path, the path planning algorithm will prioritize path segments with lower costs, i.e., those leading to areas rich in positioning features, thereby guiding the AMR robot to actively move toward these areas. This strategy can be generated by rerunning the path planning algorithm or making local adjustments to the existing path to ensure that the new path effectively guides the AMR robot to areas with better positioning features.

[0077] In some preferred embodiments, as illustrated below using a specific example, assume that an AMR robot is performing a handling task in a large warehouse, and its positioning system primarily relies on LiDAR for SLAM positioning. When the AMR robot travels to an open area, such as a large, empty storage area, where positioning features such as walls and shelves are sparse, the matching score corresponding to the LiDAR data remains below the matching quality threshold, causing the AMR robot's positioning confidence value to gradually decrease.

[0078] At this time, according to the solution of this application, the system will read the current positioning confidence value. If the confidence value is lower than the preset threshold, the path planning module will dynamically adjust the cost weights of different path segments in its internal map according to the preset function rules. For example, if the map is marked with areas with rich positioning features such as high-density shelf areas and fixed equipment areas, the cost weights of the path segments leading to these areas will be significantly reduced. On the contrary, the cost weights of the path segments leading to open areas or areas with fuzzy features will be increased. Subsequently, the path planning algorithm (such as A The algorithm recalculates the optimal path based on these adjusted cost weights. The new path no longer solely considers the shortest distance or fastest time, but instead prioritizes paths that can guide the AMR robot into high-density shelf areas or fixed equipment areas as quickly as possible. For example, even if it requires a detour, as long as it can reach an area with rich positioning features more quickly, that path will be preferred. Once the AMR robot enters these areas, its lidar will be able to match more feature points, the matching score will rebound, and the positioning confidence value will increase accordingly, allowing normal path planning and driving speed to resume. This proactive path adjustment mechanism effectively prevents the AMR robot from driving blindly when positioning confidence is insufficient, significantly improving its autonomous navigation capabilities and reliability in complex environments.

[0079] Optionally, the step of adjusting the cost weights of different path segments in the path planning algorithm according to the positioning confidence value includes: Read the positioning confidence value; According to the positioning confidence value, the cost weights of different path segments are calculated through preset function rules.

[0080] Specifically, reading the positioning confidence value refers to the system obtaining the positioning confidence value assessed by the AMR robot's positioning system at the current moment. This positioning confidence value can be a quantitative indicator reflecting the reliability of the AMR robot's current pose estimation. For example, it can be derived from a comprehensive assessment based on factors such as the duration of pose update failures, the uncertainty of the current pose estimate (e.g., the covariance matrix), and the particle diffusion of the positioning system's particle filter. Based on the positioning confidence value, the cost weights of different path segments are calculated using a preset function rule. This can be understood as using a predefined mathematical model or logical rule to map the positioning confidence value to the cost weights of each path segment in the path planning algorithm. This preset function rule can be nonlinear. For example, when the positioning confidence value is low, the function rule can be set to significantly increase the cost weights of path segments leading to areas with sparse positioning features, while reducing or maintaining the cost weights of path segments leading to areas with abundant positioning features. Conversely, when the positioning confidence value is high, this difference can be reduced. Its purpose is to guide the AMR robot to prioritize moving to areas with rich positioning features and high positioning stability by increasing the travel cost of certain areas (such as areas with sparse positioning features) when positioning confidence is insufficient, thereby actively avoiding positioning risks.

[0081] In some preferred embodiments, a specific example is provided below. Assume that an AMR robot is performing material handling in a large warehouse. When the AMR robot enters an area with dense shelves, high positioning feature similarity, and poor lighting conditions, its positioning system may be unable to update its position because the LiDAR data matching score continues to fall below the matching quality threshold, causing the positioning confidence value to gradually decrease.

[0082] At this time, according to the solution of this application, the system will read the current positioning confidence value. For example, if the positioning confidence value drops from 0.9 to 0.5, the preset function rule can be designed as: When the positioning confidence value is higher than 0.7, the cost weight of the path segment leading to the area rich in known positioning features is 1.0, and the cost weight of the path segment leading to the area sparse in positioning features is 1.2.

[0083] When the localization confidence value is between 0.5 and 0.7, the cost weight of the path segments leading to the region rich in known localization features remains 1.0, while the cost weight of the path segments leading to the region sparse in localization features is increased to 1.5.

[0084] When the localization confidence value is lower than 0.5, the cost weight of the path segments leading to the region rich in known localization features remains at 1.0, while the cost weight of the path segments leading to the region sparse in localization features increases dramatically to 2.0 or higher, or can even be set to infinity to avoid it completely.

[0085] An AMR robot path control system is used to perform AMR robot path control, combined with Figure 6 As shown, the AMR robot path control system 1 includes: The state information acquisition module 11 is used to obtain the motion state information of the AMR robot and the current consumption information of the drive motor; The state risk judgment module 12 is used to calculate the turning radius of the AMR robot based on the motion state information, and determine whether the AMR robot is in a high-risk positioning condition based on the turning radius and the current consumption information of the drive motor, thereby obtaining a high-risk judgment result; A matching threshold adjustment module 13 is used to adjust the matching quality threshold of the location information received by the positioning system according to the high-risk judgment result; The posture update control module 14 is used to determine whether to update the posture of the AMR robot based on the adjusted matching quality threshold when receiving the matching score corresponding to the lidar data, and perform control actions according to the judgment result; The matching threshold recovery module 15 is used to restore the matching quality threshold of the positioning system after the high-risk positioning condition is resolved.

[0086] To facilitate a clearer and easier understanding of the technical solutions of this application, some key terms are explained below. An AMR robot is an autonomous mobile robot, a mobile platform capable of autonomously navigating, avoiding obstacles, and performing tasks in complex environments. Motion state information refers to real-time data such as the speed, acceleration, angular velocity, and heading of an AMR robot during operation. Drive motor current consumption information refers to the actual current consumed by the AMR robot's drive motor during operation. This data reflects the impact of factors such as motor load, ground friction, and wheel wear on motor output. Turning radius refers to the radius of the circular arc trajectory followed by the AMR robot's center of mass during a turn, reflecting the abruptness of the turn. A high-risk positioning condition refers to a risky state in which, under specific operating conditions, the AMR robot's positioning system is prone to feature mismatching, resulting in inaccurate pose estimation. The positioning system refers to the system used by the AMR robot to determine its precise position and pose in the environment. It typically integrates data from multiple sensors, such as lidar, an inertial measurement unit (IMU), and an odometry meter. The matching quality threshold refers to the minimum acceptable score used by the positioning system to assess the degree of match between current sensor data and map features when processing sensor data. LiDAR data refers to the three-dimensional point cloud data generated by the LiDAR sensor after scanning the surrounding environment, which contains information about the geometric features of the environment. The matching score is a quantitative indicator used by the positioning algorithm to evaluate the degree of fit after matching the current LiDAR data with pre-built map features. Pose refers to a comprehensive description of the position and posture of the AMR robot in three-dimensional space. Control action refers to the corresponding behavior taken by the AMR robot based on the positioning results and risk assessment. This system is commonly used for AMR robots in scenarios such as smart warehousing and factory automation. These scenarios may include mixed floors, narrow aisles, and high-density storage areas, which place high demands on the path control accuracy and safety of the AMR robot.

[0087] The above are merely examples of the present application and are not intended to limit the scope of protection of the present application. Those skilled in the art will appreciate that various modifications and variations are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A path control method for an AMR robot, characterized in that: include: Obtain the motion status information of the AMR robot and the current consumption information of the drive motor; Calculating a turning radius of the AMR robot according to the motion state information, and determining whether the AMR robot is in a high-risk positioning condition according to the turning radius and the current consumption information of the drive motor, thereby obtaining a high-risk determination result; Adjusting the matching quality threshold of the positioning system for accepting location information based on the high-risk judgment result; Based on the adjusted matching quality threshold, when receiving the matching score corresponding to the lidar data, it determines whether to update the AMR robot's posture and performs control actions based on the judgment result; After the high-risk positioning condition is resolved, the matching quality threshold of the positioning system is restored.

2. The AMR robot path control method according to claim 1, characterized in that: The step of calculating the turning radius of the AMR robot according to the motion state information, and determining whether the AMR robot is in a high-risk positioning condition according to the turning radius and the current consumption information of the drive motor, and obtaining a high-risk determination result includes: Calculate the turning radius of the AMR robot based on the motion state information; Get the AMR robot's travel speed and / or robot load; Adjusting and determining a turning radius threshold according to the driving speed and / or robot load; Obtain the friction characteristics of the floor where the AMR robot is located; adjusting and determining a current consumption threshold according to the driving speed and / or friction characteristics; According to the turning radius, the driving motor current consumption information, the turning radius threshold and the current consumption threshold, it is determined whether the AMR robot is in a high-risk positioning condition to obtain a high-risk judgment result.

3. The AMR robot path control method according to claim 1, characterized in that: The step of adjusting the matching quality threshold of the positioning system for accepting location information according to the high-risk judgment result includes: Obtain the turning radius and drive motor current consumption information of the AMR robot; determining a turning risk level based on the turning radius; determining a load risk level according to the driving motor current consumption information; Determining a comprehensive risk level of the high-risk judgment result according to the turning risk level and the load risk level; According to the comprehensive risk level, a matching quality threshold is selected as the matching quality threshold for the positioning system to accept position information.

4. The AMR robot path control method according to claim 3, characterized in that: The step of selecting a matching quality threshold according to the comprehensive risk level as the matching quality threshold for the positioning system to accept location information includes: Get the positioning feature quality of the AMR robot's current environment; Obtain the accuracy requirements of the AMR robot's current task; A matching quality threshold is calculated based on the comprehensive risk level, positioning feature quality, and accuracy requirements, and serves as the matching quality threshold for the positioning system to accept position information.

5. The AMR robot path control method according to claim 1, characterized in that: The steps of determining whether to update the posture of the AMR robot based on the adjusted matching quality threshold upon receiving the matching score corresponding to the lidar data, and executing a control action according to the determination result include: When receiving the matching score corresponding to the lidar data, the current matching score is compared with the adjusted matching quality threshold to obtain the matching comparison result; If the matching comparison result indicates that the current matching score is higher than or equal to the matching quality threshold, updating the posture of the AMR robot; If the matching comparison result indicates that the current matching score is lower than the matching quality threshold, the position and posture of the AMR robot are not updated; When the AMR robot's posture is not updated, the positioning confidence value of the AMR robot is evaluated based on the duration of posture update failure and the uncertainty of the current posture estimate; Adjusting the maximum travel speed of the AMR robot according to the positioning confidence value; Adjusting the path planning strategy of the AMR robot according to the positioning confidence value to guide the AMR robot to an area rich in positioning features; If the positioning confidence value is lower than a preset threshold and reaches a preset time, the AMR robot's safe parking behavior is triggered or an assistance request is sent to the central dispatch system.

6. The AMR robot path control method according to claim 5, characterized in that: The step of evaluating the positioning confidence value of the AMR robot according to the duration of the pose update failure and the uncertainty of the current pose estimation when the pose of the AMR robot is not updated includes: Get the count of failed pose updates; Get the covariance matrix of the current pose estimate; Get the particle diffusion degree of the positioning system particle filter; The positioning confidence value of the AMR robot is comprehensively evaluated based on the count of pose update failures, covariance matrix, and particle diffusion degree.

7. The AMR robot path control method according to claim 5, characterized in that: The step of adjusting the maximum driving speed of the AMR robot according to the positioning confidence value includes: According to the positioning confidence value, the maximum driving speed of the AMR robot is calculated through a preset functional relationship.

8. The AMR robot path control method according to claim 5, characterized in that: The step of adjusting the path planning strategy of the AMR robot according to the positioning confidence value to guide the AMR robot to an area rich in positioning features includes: Adjusting the cost weights of different path segments in the path planning algorithm based on the positioning confidence value; Based on the adjusted cost weights, a path planning strategy is generated for the AMR robot to move to areas with rich positioning features.

9. The AMR robot path control method according to claim 8, characterized in that: The step of adjusting the cost weights of different path segments in the path planning algorithm according to the positioning confidence value includes: Reading the positioning confidence value; According to the positioning confidence value, the cost weights of different path segments are calculated using a preset function rule.

10. An AMR robot path control system for performing AMR robot path control, characterized in that: include: The state information acquisition module is used to obtain the motion state information of the AMR robot and the current consumption information of the drive motor; a state risk judgment module, configured to calculate a turning radius of the AMR robot based on the motion state information, and determine whether the AMR robot is in a high-risk positioning condition based on the turning radius and the current consumption information of the drive motor, thereby obtaining a high-risk judgment result; A matching threshold adjustment module, configured to adjust a matching quality threshold for receiving location information by the positioning system according to the high-risk judgment result; The pose update control module is used to determine whether to update the pose of the AMR robot based on the adjusted matching quality threshold when receiving the matching score corresponding to the lidar data, and perform control actions based on the judgment result; The matching threshold recovery module is used to restore the matching quality threshold of the positioning system after the high-risk positioning condition is resolved.

Images