Medical robot ward-round navigation method and system

Abstract

The invention provides a medical robot ward-round navigation method and a medical robot ward-round navigation system, which are applied to the technical field of medical robots and are characterized in that a mechanism for identifying and compensating progressive drifting of auxiliary positioning sensing equipment is introduced and is combined with dynamic management of an uncertainty area of the position of a robot, so that when the performance of the sensing equipment is reduced, the positioning accuracy of the robot is improved; and the high-precision and high-safety navigation capability can still be maintained. According to the method, the problems of inaccurate robot positioning, low operation efficiency and increased potential safety hazards caused by progressive drift of sensing equipment in the prior art are effectively solved, and the robustness and reliability of the medical robot in a complex and dynamic medical environment are remarkably improved; and low-efficiency correction and potential physical contact caused by reduced navigation performance due to'silence 'are avoided, so that the overall operation efficiency and safety of the medical robot are improved.

Description

Technical Field

[0001] This application relates to the field of medical robot technology, and in particular to a method and system for ward round navigation of a medical robot. Background Technology

[0002] In modern medical environments, medical robots are playing an increasingly important role in ward round navigation, aiming to improve hospital operational efficiency and reduce the burden on medical staff. These robots are typically equipped with various sensing devices, such as laser rangefinders, optical cameras, inertial measurement units, and wheeled odometers. They intelligently process information from these different sensing devices to achieve precise positioning, map building, and dynamic obstacle avoidance. However, in long-term, high-intensity practical applications, some potential technical problems may gradually emerge, posing a serious challenge to the robot's navigation capabilities.

[0003] As the robot continues its demanding patrol schedule, especially when traversing long corridors, making multiple turns between wards, or passing through areas with repetitive visual features, this small, systematic error from the odometry system begins to accumulate. The robot's internal estimation of its position gradually deviates from its true physical location. While this accumulated positional error remains "tolerable" for general, coarse navigation, it begins to manifest as problems affecting efficiency and safety in practical operation. When the robot attempts to perform precise tasks, such as precisely aligning with a charging dock, accurately stopping in front of a supply cabinet for automated distribution, or navigating narrow doorways, it may require multiple, small, hesitant corrections. These repetitive adjustments consume more time and energy, reducing overall operational efficiency. More critically, when it approaches a ward to deliver items, its internal map may show it perfectly aligned with a door, but its actual physical orientation may be slightly off. This could cause the robot to attempt to open a door that is not centered in its field of vision, or misjudge clearance when passing through narrow doorways, potentially scraping against doorframes, walls, or even unintentionally bumping into a mobile medical cart that it "thinks" has enough clearance. Instead of recording traditional "collisions" to trigger emergency stops, the system records a series of subtle, unexpected physical contacts or inefficient movements, indicating a fundamental and growing misalignment between its perceived internal state and actual physical reality. The primary sensing devices (lasers and cameras) may still be reporting obstacles and features, but their interpretations relative to the robot's "perceived" locations become increasingly inconsistent, leading to hesitant or inefficient movements without triggering a clear "navigation failure" state that would require system reset or human intervention. The processing system struggles to reconcile subtle, persistent drift from odometry with seemingly correct but context-inconsistent data from lasers and cameras, resulting in a silent decline in navigation performance.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] In view of the shortcomings of the prior art, this application provides a medical robot ward round navigation method and system, which aims to solve the problem that in the long-term operation of medical robots, the gradual drift of the auxiliary positioning and sensing device leads to inaccurate position estimation of the robot itself, which in turn affects navigation accuracy, operating efficiency and safety.

[0006] In a first aspect, a method for navigating ward rounds with a medical robot includes the following steps:

[0007] S1: Obtain the uncertain region of the robot's own position;

[0008] S2: Monitor the consistency between the output data of the auxiliary positioning sensing device and the positioning result of the main sensing device to identify progressive drift of the auxiliary positioning sensing device;

[0009] S3: Based on the identified progressive drift, expand the uncertainty region of the robot's own position;

[0010] S4: Based on the expanded uncertainty region, plan the robot's travel path and use the expanded uncertainty region as the robot's virtual operating body;

[0011] S5: Based on the travel path, predict the distance between the virtual manipulator and the obstacle, and adjust the robot's motion parameters when the distance is less than a preset threshold.

[0012] The proposed method for ward round navigation of medical robots can effectively identify the progressive drift of auxiliary positioning sensing devices and dynamically adjust the uncertain area of ​​the robot's own position accordingly. The uncertain area is used as a virtual operating body for path planning and obstacle avoidance. Thus, even when the performance of the sensing device deteriorates, the navigation accuracy and operational safety of the robot can still be ensured, avoiding inefficient or dangerous operations caused by positioning errors.

[0013] Furthermore, step S5 includes:

[0014] S6: Obtain the area of ​​the expanded uncertainty region;

[0015] S7: Compare the area with a preset area threshold. When the area is greater than or equal to the preset area threshold, adopt a conservative adjustment strategy to adjust the robot's motion parameters.

[0016] The conservative adjustment strategy includes at least the following:

[0017] S71: Perform a preset small movement and update the uncertain area of ​​the robot's own position after the movement;

[0018] S72: Based on the updated uncertainty region, obtain the deviation between the robot's current position and the target alignment point;

[0019] S73: Based on the deviation, determine whether another small movement is needed;

[0020] S74: If necessary, repeat small-amplitude movements, updates of uncertain regions, acquisition of deviations, and judgment of deviations until the robot and the target alignment point reach the preset alignment accuracy requirement.

[0021] The proposed medical robot ward round navigation method can proactively adopt a conservative adjustment strategy when the uncertainty area of ​​the robot's own position is too large. By moving in small increments and iteratively calibrating, the positioning deviation can be gradually reduced. Thus, even under conditions of high uncertainty, high-precision alignment and positioning can still be achieved, further improving the robot's operational accuracy and reliability.

[0022] Furthermore, step S72 includes:

[0023] S721: Calculate the overlap probability between the updated uncertainty region and the target alignment point;

[0024] S722: Compare the overlap probability with a preset probability threshold range. If the overlap probability is greater than the upper limit of the preset probability threshold range, then determine that the deviation corresponds to a level one deviation. The deviation value corresponding to the level one deviation is 0.01, and the unit is meters.

[0025] S723: If the overlap probability is less than or equal to the upper limit of the set probability threshold range and greater than or equal to the lower limit of the set probability threshold range, then the level corresponding to the deviation is determined to be a level two deviation, and the deviation value corresponding to the level two deviation is 0.05, in meters;

[0026] S724: If the overlap probability is less than the lower limit of the set probability threshold range, then the level corresponding to the deviation is determined to be a level three deviation, and the deviation value of the level three deviation is 0.1, in meters.

[0027] The proposed medical robot ward round navigation method can quantify the level of positioning deviation based on the overlap probability between the uncertain area and the target alignment point, and assign different deviation values, thereby providing a more refined and targeted basis for subsequent motion adjustments, enabling the robot to perform more intelligent calibration based on the actual deviation.

[0028] Furthermore, step S1 includes:

[0029] S11: Obtain the robot's real-time perception data;

[0030] S12: Determine the robot's current position information and its surrounding environment information based on the real-time sensing data;

[0031] S13: Based on the current location information, a preliminary uncertain region is preset with the current location information as the center point;

[0032] S14: Match the environmental information with a preset map. When the matching degree exceeds a preset matching degree threshold, reduce the initial uncertain area by a preset ratio to obtain the uncertain area.

[0033] The proposed method for ward round navigation of medical robots can dynamically adjust the initially uncertain area by combining real-time perception data and map matching results. When the environmental matching degree is high, the uncertainty is reduced, thereby more accurately reflecting the actual uncertainty of the robot's own position and providing more reliable initial data for subsequent navigation and obstacle avoidance.

[0034] Furthermore, step S2 includes:

[0035] S21: Continuously collect the output data of the auxiliary positioning sensing device and the positioning results of the main sensing device;

[0036] S22: Calculate the positional difference between the output data and the positioning result;

[0037] S23: Perform a trend analysis on the positional differences and identify the cumulative trend of the differences based on the trend analysis results, thereby identifying the progressive drift of the auxiliary positioning sensing device.

[0038] Furthermore, step S22 includes:

[0039] S221: Time-align the output data with the positioning result;

[0040] S222: Convert the time-aligned output data and the positioning result to a unified reference coordinate system;

[0041] S223: Calculate the geometric distance and angle difference between the pose of the output data and the pose of the positioning result, and use the geometric distance and angle difference as the position difference.

[0042] Furthermore, step S3 includes:

[0043] S31: Obtain the cumulative amount or rate of change of the gradual drift;

[0044] S32: Adjust the size of the uncertainty region of the robot's own position according to the accumulated amount or rate of change, so that there is a positive correlation between the size of the uncertainty region and the accumulated amount or rate of change.

[0045] Furthermore, step S4 includes:

[0046] S41: Perform geometric expansion on the expanded uncertainty region to obtain a safe region;

[0047] S42: Based on the safe area, a graph search algorithm is used to plan the robot's travel path. The planning process uses the expanded uncertainty area as the robot's virtual operating body.

[0048] Furthermore, step S5 includes:

[0049] S51: On the travel path, sample multiple future pose points at a preset time step;

[0050] S52: For each of the plurality of future pose points, based on the uncertainty region, simulate the spatial range that the virtual manipulator may occupy at that future pose point;

[0051] S53: Based on the spatial range, calculate the closest distance between the virtual operator and the obstacle;

[0052] S54: If the nearest distance is less than the preset threshold, then adjust the robot's speed, acceleration or turning angular velocity according to the distance difference between the nearest distance and the preset threshold, and generate motion adjustment instructions.

[0053] Secondly, a medical robot ward round navigation system is provided for implementing any of the methods described above, the system comprising:

[0054] Acquisition module: Acquires the uncertain region of the robot's own position;

[0055] Monitoring module: Monitors the consistency between the output data of the auxiliary positioning sensing device and the positioning results of the main sensing device to identify progressive drift of the auxiliary positioning sensing device;

[0056] Extension module: Based on the identified progressive drift, expand the uncertainty region of the robot's own position;

[0057] Planning module: Based on the expanded uncertainty region, the module plans the robot's travel path and uses the expanded uncertainty region as the robot's virtual operating body;

[0058] Adjustment module: Based on the travel path, predict the distance between the virtual manipulator and the obstacle, and adjust the robot's motion parameters when the distance is less than a preset threshold.

[0059] Beneficial Effects: The medical robot ward round navigation method and system proposed in this application introduces a mechanism for identifying and compensating for the progressive drift of auxiliary positioning sensing devices, and combines this with dynamic management of the robot's own positional uncertainty area. This achieves high-precision and high-safety navigation capabilities even when the performance of the sensing devices deteriorates. This method effectively solves the problems of inaccurate robot positioning, reduced operational efficiency, and increased safety hazards caused by progressive drift of sensing devices in existing technologies. It significantly improves the robustness and reliability of medical robots in complex and dynamic medical environments, avoids inefficient corrections and potential physical contact caused by "silent" decline in navigation performance, thereby improving the overall operational efficiency and safety of medical robots. Attached Figure Description

[0060] Figure 1 This is a flowchart of a medical robot ward round navigation method proposed in this application.

[0061] Figure 2 This is a structural diagram of a medical robot ward round navigation system proposed in this application.

[0062] Figure 3 This is a simplified schematic diagram of a medical robot ward round navigation system proposed in this application. Detailed Implementation

[0063] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and marked in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0064] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0065] Please refer to Figure 1A method for ward round navigation using a medical robot, the method comprising the following steps:

[0066] S1: Obtain the uncertain region of the robot's own position;

[0067] S2: Monitor the consistency between the output data of the auxiliary positioning sensing device and the positioning result of the main sensing device to identify progressive drift of the auxiliary positioning sensing device;

[0068] S3: Based on the identified progressive drift, expand the uncertainty area of ​​the robot's own position;

[0069] S4: Based on the expanded uncertainty region, plan the robot's travel path and use the expanded uncertainty region as the robot's virtual operating body;

[0070] S5: Based on the travel path, predict the distance between the virtual manipulator and obstacles, and adjust the robot's motion parameters when the distance is less than a preset threshold.

[0071] The "uncertainty region" mentioned in this application refers to the range of errors in a robot's estimation of its own position, typically represented by a probability distribution or a geometric region (e.g., an ellipse or polygon). The larger the region, the lower the certainty of the robot's own position. "Auxiliary positioning sensing devices" can refer to wheeled odometry, inertial measurement units (IMUs), etc., which provide relative positioning information but are susceptible to accumulated errors. "Primary sensing devices" typically refer to lidar, cameras, etc., which provide absolute or semi-absolute positioning information and have strong perception capabilities of environmental features. "Progressive drift" refers to the phenomenon where the output error of the auxiliary positioning sensing device slowly accumulates over time; this drift is usually difficult to detect by traditional fault detection mechanisms. A "virtual manipulator" is a combination of the robot's physical dimensions and the uncertainty region, used to more conservatively represent the space the robot may occupy in path planning and obstacle avoidance.

[0072] In step S1, obtaining the uncertainty region of the robot's own position is fundamental to the navigation process. One implementation is that the robot can initially determine an uncertainty region based on its initial position and a preset error model, such as a Gaussian distribution. As the robot moves, this region expands according to the motion model and sensor noise. For example, when the robot starts, its initial uncertainty region can be set to a small circular area, indicating high initial positioning accuracy. Another implementation is that the robot can periodically obtain its own position through manual calibration or matching with known landmarks, and determine the uncertainty region based on the calibration results or matching errors. For example, before each ward round task begins, the operator can manually move the robot to a designated location for calibration and set the initial uncertainty region based on the calibration error.

[0073] In step S2, the consistency between the output data of the auxiliary positioning sensing device and the positioning result of the primary sensing device is monitored to identify progressive drift of the auxiliary positioning sensing device. One implementation is that the system can simply compare the displacement reported by the auxiliary positioning sensing device (e.g., odometry) with the displacement estimated by the primary sensing device (e.g., LiDAR SLAM) and calculate the difference between them. If this difference continues to increase, it may indicate progressive drift. For example, if the odometry reports that the robot has moved 1 meter, while LiDAR SLAM estimates that the robot has only moved 0.98 meters, and this 0.02-meter difference continues to accumulate, it may be identified as drift. Another implementation is that the system can perform a simple difference calculation between the output data of the auxiliary positioning sensing device and the positioning result of the primary sensing device and set a fixed threshold. When the difference exceeds this threshold, drift is considered to exist. For example, if the difference between the odometry and LiDAR positioning results exceeds 5 centimeters for 10 consecutive seconds, the system will issue a drift warning.

[0074] In step S3, the uncertainty region of the robot's own position is expanded based on the identified progressive drift. One implementation is that when progressive drift is detected, the system can simply increase the size of the uncertainty region by a fixed scaling factor. For example, if drift is detected, the radius or semi-axis length of the uncertainty region is directly increased by 10%. Another implementation is that the system can linearly increase the size of the uncertainty region based on the duration or accumulation of the drift. For example, for every minute the drift lasts, the area of ​​the uncertainty region increases by 5 square centimeters.

[0075] In step S4, the robot's path is planned based on the expanded uncertainty region, and this expanded uncertainty region is used as the robot's virtual manipulator. One implementation is that, during path planning, the system can simply treat the expanded uncertainty region as a fixed-size circular or square area and superimpose it onto the robot's physical dimensions to form a larger virtual manipulator. Path planning algorithms (such as A* or Dijkstra's algorithm) will prevent this virtual manipulator from colliding with obstacles. For example, if the robot's physical radius is 0.3 meters and the radius of the uncertainty region is 0.1 meters, then the radius of the virtual manipulator will be considered as 0.4 meters for path planning. Another implementation is that, during path planning, the system can use the geometric center of the expanded uncertainty region as the robot's reference point and ensure that this reference point maintains a safe distance from obstacles. The size of this safe distance is determined by the size of the uncertainty region. For example, if the distance between the robot's center and the obstacle is less than the radius of the uncertainty region, a collision risk is considered to exist.

[0076] In step S5, based on the travel path, the distance between the virtual manipulator and obstacles is predicted, and the robot's motion parameters are adjusted if the distance is less than a preset threshold. One implementation is that the system can simply calculate the Euclidean distance between the virtual manipulator and the nearest obstacle at each sampling point on the path. If this distance is less than the preset threshold, the robot stops immediately. For example, if the distance between the virtual manipulator and a wall is less than 0.1 meters, the robot will stop immediately. Another implementation is that the system can predict all the space the virtual manipulator might occupy within a future time step and calculate the minimum distance between that space and obstacles. If this minimum distance is less than a preset threshold, the robot's speed is reduced. For example, if it is predicted that the virtual manipulator may come into contact with an obstacle within the next 2 seconds, the robot's speed is reduced by 50%.

[0077] Compared with existing technologies, the core innovation of this application lies in its sensitive identification and active compensation mechanism for the progressive drift of the auxiliary positioning sensing device. Traditional navigation methods often focus on dealing with random noise or sudden failures, but lack effective means to deal with such slowly accumulating systematic errors. For example, in existing technologies, when a wheeled odometer produces a small but continuous error due to wear, the system may misinterpret it as normal operating noise, causing the robot's internal position estimation to gradually deviate from the true position, ultimately resulting in hesitation, inefficiency, or even minor collisions when performing precise tasks. This application, by introducing monitoring of the consistency between the output data of the auxiliary positioning sensing device and the positioning results of the main sensing device, can promptly detect and quantify this progressive drift. By expanding the uncertainty region according to the degree of drift and using it as a virtual manipulator for path planning and obstacle avoidance, this application can provide the robot with a more conservative and safer navigation strategy. This method not only improves the robot's positioning accuracy and navigation robustness in long-term operation, but also significantly reduces the collision risk and task execution efficiency decline caused by accumulated errors, thus demonstrating significant technological progress and practical value in real-world applications.

[0078] In some embodiments described above in this application, the robot's motion parameters are adjusted when the distance between the virtual manipulator of the medical robot and an obstacle is less than a preset threshold. However, in practical applications, if the uncertainty range of the robot's own position is too large, simply adjusting the motion parameters may not be sufficient to ensure the robot's safety and navigation accuracy in complex or narrow environments, especially in scenarios requiring precise alignment or obstacle avoidance, where there may be potential risks due to excessive uncertainty.

[0079] Furthermore, step S5 includes:

[0080] S6: Obtain the area of ​​the expanded uncertainty region;

[0081] S7: Compare the area with the preset area threshold. When the area is greater than or equal to the preset area threshold, adopt a conservative adjustment strategy to adjust the robot's motion parameters.

[0082] A conservative adjustment strategy should include at least the following:

[0083] S71: Perform a preset small movement and update the uncertain area of ​​the robot's own position after the movement;

[0084] S72: Based on the updated uncertainty region, obtain the deviation between the robot's current position and the target alignment point;

[0085] S73: Based on the deviation, determine whether another small movement is needed;

[0086] S74: If necessary, repeat small-amplitude movements, updates of uncertain areas, acquisition of deviations, and judgment of deviations until the robot and the target alignment point reach the preset alignment accuracy requirements.

[0087] Specifically, obtaining the area of ​​the expanded uncertainty region in step S6 refers to quantifying the uncertainty of the robot's current position by calculating the geometric size of this region in two-dimensional or three-dimensional space. This area can serve as an important indicator for evaluating the robot's positioning accuracy. The preset area threshold is a reference value set based on the robot's positioning accuracy requirements in the actual application scenario. For example, in open areas such as hospital corridors, this threshold can be set relatively leniently; while in ward entrances, narrow passages, or areas requiring precise stopping, the threshold will be set more strictly. When the obtained area of ​​the uncertainty region is greater than or equal to the preset area threshold, it indicates that the robot's current positioning accuracy is low and there is significant uncertainty. In this case, a conservative adjustment strategy is necessary. The conservative adjustment strategy aims to gradually reduce uncertainty and precisely adjust the robot's position through a series of refined operations.

[0088] In the conservative adjustment strategy, step S71 involves performing a preset small-amplitude movement, which means the robot moves a short distance with extremely low speed and displacement, for example, 0.01 meters or 0.05 meters each time. The purpose of this small-amplitude movement is to update the uncertainty region of its own position using new sensor data without significantly changing the robot's overall path, thereby gradually reducing the uncertainty range. The process of updating the uncertainty region can utilize Kalman filtering, particle filtering, or other localization algorithms, combined with the sensor data after the movement, for re-estimation. In step S72, based on the updated uncertainty region, the deviation between the robot's current position and the target alignment point is obtained. The target alignment point refers to a preset position that the robot needs to accurately reach or stop at, such as a ward door, a charging station location, or a work area designated by a doctor. Obtaining the deviation aims to quantify the degree of mismatch between the robot's current position and the target alignment point. Step S73, based on the deviation, determines whether another small-amplitude movement is needed. This determination logic can be based on the magnitude, direction, or preset alignment accuracy requirements of the deviation. For example, if the deviation is still large, or the preset alignment accuracy requirements have not yet been met, it is determined that another small-amplitude movement is needed. In step S74, if necessary, the small-amplitude movement, updating of the uncertainty region, acquisition of deviation, and determination of deviation are repeated until the robot and the target alignment point reach the preset alignment accuracy requirement. The preset alignment accuracy requirement refers to the maximum permissible error between the robot's final position and the target alignment point, for example, it can be set to 0.01 meters or 0.02 meters. Through this iterative loop, the robot can gradually converge its position in a high-uncertainty environment, ultimately achieving high-precision positioning and alignment.

[0089] Furthermore, step S72 includes:

[0090] S721: Calculate the overlap probability between the updated uncertainty region and the target alignment point;

[0091] S722: Compare the overlap probability with a preset probability threshold range. If the overlap probability is greater than the upper limit of the preset probability threshold range, the deviation is judged to be a level 1 deviation. The deviation value corresponding to level 1 deviation is 0.01, and the unit is meters.

[0092] S723: If the overlap probability is less than or equal to the upper limit of the set probability threshold range and greater than or equal to the lower limit of the set probability threshold range, then the level corresponding to the deviation is a second-level deviation, and the deviation value corresponding to the second-level deviation is 0.05, in meters.

[0093] S724: If the overlap probability is less than the lower limit of the set probability threshold range, the judgment deviation is classified as a level three deviation, with a deviation value of 0.1 in meters.

[0094] Specifically, overlap probability refers to the likelihood that, considering the uncertainty of the robot's own position, the robot's actual position will fall within a predetermined area near the target alignment point. The calculation of overlap probability involves comparing the uncertainty area of ​​the robot's current position with a "preset area" around the target alignment point. This preset area can be understood as an acceptable error range around the target alignment point; if the robot's actual position falls within this preset area, the robot is considered to have successfully aligned or reached the target. Overlap probability calculates the amount of overlap between the uncertainty area and this preset area of ​​the target alignment point, and the probability of such overlap. Specifically, if the uncertainty area of ​​the robot's own position completely overlaps with the preset area near the target alignment point, or if most of the uncertainty area falls within the preset area, the overlap probability will be high, indicating that the robot's current position is very close to the target alignment point and the positioning accuracy is high. Conversely, if the overlap between the two areas is small or even nonexistent, the overlap probability will be low, indicating that the robot's current position deviates significantly from the target alignment point and the positioning accuracy is low.

[0095] The higher the probability value, the closer the robot is to the target alignment point. A preset probability threshold range is used to define the boundaries of different deviation levels, typically including an upper limit and a lower limit. Level 1, Level 2, and Level 3 deviations represent different degrees of positioning deviation and are assigned corresponding deviation values, such as 0.01 meters, 0.05 meters, and 0.1 meters, respectively. These deviation levels and corresponding deviation values ​​provide a quantitative assessment of the distance between the robot's current position and the target alignment point. Step S721 aims to quantify the spatial relationship between the robot's uncertainty area and the target alignment point through probability calculation. Steps S722 to S724, based on the calculated overlap probability, compare it with the preset probability threshold range to classify the deviation into different levels and assign specific deviation values, thereby achieving refined management of the deviation.

[0096] The proposed solution quantifies the degree of matching between the robot's current position and the target alignment point by calculating the overlap probability between the updated uncertainty region and the target alignment point. By introducing a probability threshold range to classify the overlap probability, deviations can be meticulously divided into different levels, such as first-level, second-level, and third-level deviations, with a specific deviation value assigned to each level. This classification mechanism allows the system to adopt more refined and appropriate subsequent adjustment strategies based on the severity of the deviation, rather than simply judging whether alignment has occurred, thereby improving the accuracy and efficiency of robot localization and alignment.

[0097] Through the aforementioned technical solution, the robot can obtain more refined positioning deviation information. This tiered deviation acquisition method allows the system to select a more appropriate adjustment range or strategy based on the deviation level when implementing a conservative adjustment strategy. For example, a minor adjustment can be made for a level-one deviation, while a larger adjustment may be required for a level-three deviation. This not only improves the accuracy and robustness of robot alignment but also optimizes the efficiency of the adjustment process, avoiding unnecessary over-adjustment or under-adjustment, thereby enhancing the overall performance and safety of medical robot ward round navigation.

[0098] Furthermore, step S1 includes:

[0099] S11: Acquire real-time perception data of the robot;

[0100] S12: Determine the robot's current position and environment information based on real-time sensing data;

[0101] S13: Preset a preliminary uncertain area centered on the current location information based on the current location information;

[0102] S14: Match the environmental information with the preset map. When the matching degree exceeds the preset matching degree threshold, reduce the initial uncertain area by a preset ratio to obtain the uncertain area.

[0103] Specifically, acquiring the robot's real-time perception data S11 refers to the raw data about its own motion state and surrounding environment collected by the medical robot at the current moment through its various onboard sensors, such as LiDAR, depth cameras, inertial measurement units (IMUs), or odometry. This data is the foundation for the robot's localization and environmental understanding.

[0104] Specifically, determining the robot's current position information and environmental information based on the real-time perception data S12 refers to processing and fusing the real-time perception data using a localization algorithm (such as Simultaneous Localization and Mapping SLAM algorithm) to calculate the robot's precise pose (including position and attitude) in the global or local coordinate system, and to identify local features of the robot's surrounding environment, such as the distribution of obstacles, the structure of walls, or key feature points.

[0105] In practical applications, the preliminary uncertain region S13, centered on the current location information, refers to the preliminary estimation of the possible range of the robot's current position based on the error model of the robot's own localization algorithm before any map matching or external correction. This preliminary uncertain region can typically be represented as an area with a specific size and shape (e.g., circular, elliptical, or polygonal) centered on the current location information, and its size reflects the confidence level of the initial localization.

[0106] Furthermore, the environmental information is matched with a preset map. When the matching degree exceeds a preset matching degree threshold, the initial uncertain region is reduced by a preset ratio to obtain the uncertain region S14. This refers to comparing the local environmental information obtained through real-time sensing data with a global map pre-stored by the robot. The preset map may contain the geometric structure and feature information of the environment. The matching degree is an indicator that measures the degree of agreement between the real-time environmental information and the preset map. When the matching degree reaches or exceeds the preset matching degree threshold, it indicates that the robot has a high degree of confidence in its own position. At this time, the range of the initial uncertain region can be reduced by a preset ratio to obtain a smaller and more accurate uncertain region of the robot's own position.

[0107] The proposed solution, through the aforementioned step-by-step refinement, first acquires the robot's real-time perception data, which forms the foundation for all localization and environmental understanding. Subsequently, based on this real-time data, the robot's current position and local information about its surrounding environment are determined. After the initial position determination, a preliminary uncertainty region is preset, reflecting the initial positioning error range. To improve positioning accuracy and reduce uncertainty, the real-time acquired environmental information is matched with a preset global map. When the matching result reaches a preset reliability standard (i.e., the matching degree exceeds a preset matching degree threshold), it indicates that the robot's judgment of its own position is more accurate. At this point, the preliminary uncertainty region can be reduced according to a preset ratio, resulting in a smaller and more accurate uncertainty region for the robot's own position. This process effectively utilizes map information to correct and optimize real-time positioning, enabling the range of the uncertainty region to dynamically reflect the confidence level of the positioning.

[0108] Furthermore, step S2 includes:

[0109] S21: Continuously collect the output data of the auxiliary positioning sensing device and the positioning results of the main sensing device;

[0110] S22: Calculate the positional difference between the output data and the positioning results;

[0111] S23: Perform a trend analysis on the sequence of positional differences, and based on the trend analysis results, identify the cumulative trend of the differences, thereby identifying the progressive drift of the auxiliary positioning sensing device.

[0112] In step S21, continuous collection refers to the uninterrupted acquisition of positioning data from auxiliary positioning sensing devices (such as inertial measurement units, odometry, etc.) and primary sensing devices (such as lidar SLAM, visual SLAM, etc.) during the ward round navigation of the medical robot. The output data of the auxiliary positioning sensing devices typically provides high-frequency relative pose information that may have accumulated errors, while the positioning results of the primary sensing devices provide relatively accurate but potentially low-frequency global pose information.

[0113] Further, in step S22, calculating the positional difference between the two devices involves comparing the output data of the auxiliary positioning sensing device with the positioning result of the main sensing device to quantify their spatial inconsistencies. This positional difference may include, but is not limited to, geometric distance and angular deviation. For example, the pose data of the two devices can be converted to a unified coordinate system, and then their translational differences along the X, Y, and Z axes, as well as their rotational differences along the pitch, yaw, and roll angles, can be calculated.

[0114] Specifically, in step S23, performing a trend analysis on the positional differences refers to analyzing the continuously calculated positional difference data over time to determine whether there is a systematic, non-random accumulation of bias. For example, a moving average method can be used to monitor the mean, variance, or rate of change of the positional differences. By identifying the accumulation trend of the differences, such as a continuous increase or a shift in a certain direction, it is possible to accurately determine whether the auxiliary positioning sensing device is experiencing progressive drift. Progressive drift refers to the phenomenon where positioning errors gradually accumulate over time due to factors such as sensor errors, environmental changes, or model inaccuracies. Its characteristic is that the error does not occur suddenly but increases slowly and continuously.

[0115] The solution in this application effectively identifies progressive drift in the auxiliary positioning sensing device by continuously collecting positioning data from both the auxiliary and primary sensing devices and systematically analyzing their positional differences. Specifically, step S21 ensures sufficient data flow for subsequent analysis; step S22 quantifies the inconsistencies between the two positioning sources; and step S23 distinguishes random noise from systematic drift by performing sequence trend analysis on these differences, thereby accurately capturing the slow, cumulative errors that may occur in the auxiliary positioning sensing device. This step-by-step and refined processing approach makes the identification of progressive drift more reliable and timely.

[0116] Furthermore, step S22 includes:

[0117] S221: Time-align the output data with the positioning results;

[0118] S222: Convert the time-aligned output data and positioning results to a unified reference coordinate system;

[0119] S223: Calculate the geometric distance and angle difference between the pose of the output data and the pose of the localization result, and use the geometric distance and angle difference as the position difference.

[0120] Specifically, in step S221, since the auxiliary positioning sensing device and the main sensing device may generate data at different sampling frequencies or asynchronously, it is necessary to time-align their respective output data and positioning results to ensure the accuracy of subsequent comparisons. This can be achieved through techniques such as timestamp synchronization, so that the pose information corresponding to the two types of devices can be obtained at the same point in time.

[0121] In step S222, the time-aligned output data and positioning results may still be in different local coordinate systems. To enable direct and effective comparison and difference calculation, these data need to be transformed into a unified reference coordinate system. For example, all data can be transformed into the robot's global navigation coordinate system. This transformation process typically involves the application of coordinate transformation matrices to eliminate coordinate system differences between different sensors or positioning systems.

[0122] In practical applications, in step S223, after the data undergoes time alignment and coordinate system transformation, the specific difference between the output pose of the auxiliary positioning sensing device and the positioning pose of the main sensing device can be calculated. The pose typically includes position (e.g., three-dimensional coordinates X, Y, Z) and orientation (e.g., Euler angles, quaternions, or rotation matrices). Geometric distance can be understood as the Euclidean distance between two position points, used to quantify spatial deviations. Angular differences are used to quantify rotational deviations between two orientations, for example, by calculating the angular difference between two orientation representations. Combining these geometric distances and angular differences yields the positional difference, the purpose of which is to comprehensively reflect the inconsistency between the two positioning information sets.

[0123] This application's solution first addresses the comparison challenge caused by data asynchrony by temporally aligning the output data of the auxiliary positioning sensing device with the positioning results of the main sensing device. Subsequently, by transforming the aligned data to a unified reference coordinate system, the inconsistency between different sensor coordinate systems is eliminated, ensuring spatial comparability of the data. Finally, by calculating the geometric distance and angular differences in pose, the degree of inconsistency between the two positioning information sets can be comprehensively and accurately quantified. It is precisely the synergistic effect of these steps that makes the calculation of positional differences more accurate and reliable, providing a solid data foundation for subsequent identification of the progressive drift of the auxiliary positioning sensing device.

[0124] The above technical solutions ensure the accuracy and reliability of calculating the differences between the output data of the auxiliary positioning sensing device and the positioning results of the main sensing device. Specifically, time alignment avoids misjudgments caused by asynchronous data sampling, unified coordinate system transformation eliminates systematic errors introduced by differences in sensor installation positions or internal coordinate systems, and the comprehensive calculation of geometric distance and angle differences provides a comprehensive quantification of pose differences. Therefore, it is possible to more accurately capture subtle, gradual drifts that may exist in the auxiliary positioning sensing device, improving the sensitivity and accuracy of drift recognition, and thus enhancing the overall robustness and safety of the medical robot navigation system.

[0125] Furthermore, step S3 includes:

[0126] S31: Obtain the cumulative amount or rate of change of the gradual drift;

[0127] S32: Adjust the size of the uncertainty region of the robot's own position according to the cumulative amount or rate of change, so that there is a positive correlation between the size of the uncertainty region and the cumulative amount or rate of change.

[0128] Specifically, in step S31, obtaining the cumulative amount or rate of change of the progressive drift refers to quantifying the degree of drift through long-term monitoring and analysis of the difference between the output data of the auxiliary positioning sensing device and the positioning result of the main sensing device. The cumulative amount can be understood as the sum or average of the positional differences between the auxiliary positioning sensing device and the main sensing device over a period of time, reflecting the overall trend and magnitude of the drift. The rate of change refers to the change in the amount of drift per unit time, reflecting the dynamics and urgency of the drift. These quantitative indicators can be calculated from historical data using statistical methods (such as moving average, Kalman filtering, etc.).

[0129] In step S32, adjusting the size of the uncertainty region of the robot's own position according to the accumulated amount or rate of change means that the expansion of the uncertainty region is no longer fixed, but dynamically adjusted according to the actual drift situation. For example, when the accumulated amount or rate of change of drift is large, the size of the uncertainty region will be increased accordingly to more reliably cover the robot's true position; conversely, when the drift is small or stable, the size of the uncertainty region can remain small to improve the accuracy and efficiency of navigation.

[0130] In practical applications, a positive correlation is established between the size of the uncertainty region and the cumulative amount or rate of change. The aim is to ensure that the expansion of the uncertainty region accurately reflects the severity of the drift. This means that the larger the drift, the larger the uncertainty region; and the smaller the drift, the smaller the uncertainty region. This positive correlation can be achieved through a predefined functional relationship, a lookup table, or an adaptive algorithm. For example, a linear or nonlinear mapping function can be defined to map the cumulative amount or rate of change of the drift to the expansion coefficient of the uncertainty region.

[0131] This application's solution effectively addresses the limitations of traditional solutions that lack fine-grained management of uncertainty region expansion by introducing the acquisition of the cumulative amount or rate of change of progressive drift and dynamically adjusting the size of the uncertainty region in the robot's own position based on this. It is precisely because the degree of drift is quantified that the expansion of the uncertainty region can be matched with the actual positioning error risk. When drift intensifies, the uncertainty region is moderately enlarged, thus providing a greater safety margin for subsequent path planning and obstacle avoidance, ensuring the robot can still operate safely even with increased positioning uncertainty. Conversely, when drift is small, the size of the uncertainty region is controlled, avoiding unnecessary conservative planning, thereby improving the robot's operational efficiency and flexibility.

[0132] Furthermore, step S4 includes:

[0133] S41: Perform geometric expansion on the expanded uncertainty region to obtain a safe region;

[0134] S42: Based on the safe area, a graph search algorithm is used to plan the robot's travel path. The planning process uses the expanded uncertain area as the robot's virtual operating body.

[0135] Geometric expansion of the expanded uncertainty region refers to further enlarging or extending the boundaries of the expanded uncertainty region after the robot's own position uncertainty region has been expanded, in order to further ensure the safety of the robot during movement. This results in a safety region larger than the actual uncertainty region, which provides the robot with additional buffer space during path planning to cope with uncertainties such as positioning errors, sensor noise, or dynamic environmental changes.

[0136] Furthermore, based on the aforementioned safe area, a graph search algorithm is used to plan the robot's path. A graph search algorithm is an algorithm that finds a path from a starting node to a target node in a graph structure, such as the A* algorithm, Dijkstra's algorithm, or RRT (Rapid Random Tree) algorithm. During path planning, the expanded uncertainty area is used as a virtual manipulator for the robot. This means that when calculating the path, the robot is no longer considered a point, but rather an entity with a certain size and shape, the size of which is determined by the expanded uncertainty area. By using the uncertainty area as a virtual manipulator, the path planning algorithm can more accurately consider the space the robot may occupy during actual movement, thereby planning a safer and more reliable path.

[0137] The proposed solution generates a safe region by geometrically expanding the expanded uncertainty region. This safe region provides the robot with an additional safety margin during path planning. Therefore, even when there is some uncertainty in the robot's actual position, it can be ensured that the robot will not collide with obstacles while moving along the planned path. Subsequently, based on this safe region, a graph search algorithm is used for path planning, with the expanded uncertainty region serving as the robot's virtual manipulator. This approach allows the path planning algorithm to take into account the robot's size and positional uncertainties, thereby effectively avoiding potential conflicts between the virtual manipulator and obstacles during path planning.

[0138] Through the above technical solution, this application can more accurately handle the positional uncertainty of the robot, and significantly improve the robustness and safety of path planning by introducing a safe zone and treating the uncertain zone as a virtual operating body. Therefore, during the navigation of medical robots during ward rounds, even in the face of complex dynamic environments and positioning errors, the risk of collisions can be effectively reduced, ensuring that the robot completes the ward round task safely and efficiently.

[0139] Furthermore, step S5 includes:

[0140] S51: Sample multiple future pose points along the travel path at a preset time step;

[0141] S52: For each of the multiple future pose points, based on the uncertainty region, simulate the spatial range that the virtual manipulator may occupy at that future pose point;

[0142] S53: Based on the spatial range, calculate the closest distance between the virtual manipulator and the obstacle;

[0143] S54: If the nearest distance is less than the preset threshold, adjust the robot's speed, acceleration or turning angular velocity according to the distance difference between the nearest distance and the preset threshold, and generate motion adjustment instructions.

[0144] Specifically, in step S51, the travel path is the expected motion trajectory planned by the robot. To predict the robot's future motion state, sampling needs to be performed along this travel path at a preset time step to obtain a series of discrete future pose points. The preset time step can be set according to factors such as the robot's maximum speed, sensor refresh rate, and required prediction accuracy; for example, it can be set to 0.1 seconds, 0.5 seconds, or 1 second. Each future pose point contains the robot's position and attitude information at that moment.

[0145] In step S52, for each sampled future pose point, the uncertainty region of the robot's own position needs to be considered. This uncertainty region reflects the error range of the robot's current position estimation. Based on this uncertainty region, the spatial range that the virtual manipulator might occupy at that future pose point can be simulated. The virtual manipulator is the expanded uncertainty region, representing the maximum space the robot might actually occupy. Simulating this spatial range can be achieved by translating and rotating the uncertainty region at each future pose point, taking into account its shape and size. For example, if the uncertainty region is modeled as an ellipse or polygon, then at each future pose point, the ellipse or polygon will be transformed accordingly to represent the actual physical space the robot might occupy.

[0146] In step S53, once the spatial range that the virtual manipulator may occupy at each future pose point is determined, the nearest distance between this spatial range and known obstacles in the environment can be calculated. The obstacles can be static walls or furniture, or dynamic people or other mobile devices. The calculation of the nearest distance can be achieved using geometric algorithms; for example, for polygonal or elliptical virtual manipulators and obstacles, the minimum Euclidean distance between points on their boundaries can be calculated.

[0147] In step S54, the calculated nearest distance is compared with a preset threshold. The preset threshold is a safety distance used to define the minimum acceptable safe distance between the robot and an obstacle. If the nearest distance is less than the preset threshold, it indicates a potential collision risk. In this case, the robot's motion parameters need to be adjusted based on the distance difference between the nearest distance and the preset threshold. The larger the distance difference, the more urgent the collision risk, and the greater the required adjustment. The motion parameters include, but are not limited to, the robot's speed, acceleration, or turning angular velocity. For example, when the distance difference is small, the speed can be moderately reduced or the turning angular velocity adjusted; when the distance difference is large, it may be necessary to decelerate sharply or even stop, and make a large turn. The adjusted motion parameters are used to generate motion adjustment commands, which are then sent to the robot's motion controller to change the robot's actual motion in real time.

[0148] This application's solution simulates the potential spatial range of a virtual manipulator by sampling future pose points along the travel path and combining this with the uncertainty region of the robot's own position. This allows for a more accurate prediction of potential collision risks between the robot and obstacles. By calculating the closest distance between the virtual manipulator and obstacles and comparing it to a preset threshold, collision risks can be detected and quantified in a timely manner. Furthermore, the robot's speed, acceleration, or turning angular velocity are dynamically adjusted based on the difference between the closest distance and the preset threshold, enabling the robot to effectively avoid collisions with obstacles while maintaining navigation efficiency. This prediction and adjustment mechanism based on uncertainty regions allows the robot to perform room-checking navigation in a safer and more robust manner in complex environments.

[0149] Please refer to Figure 2 , Figure 3 A medical robot ward round navigation system for implementing any of the above methods, the system comprising:

[0150] Acquisition Module 201: Acquires the uncertain region of the robot's own position;

[0151] Monitoring module 202: Monitors the consistency between the output data of the auxiliary positioning sensing device and the positioning result of the main sensing device to identify the gradual drift of the auxiliary positioning sensing device;

[0152] Extension module 203: Based on the identified progressive drift, expands the uncertainty area of ​​the robot's own position;

[0153] Planning Module 204: Based on the expanded uncertainty region, plan the robot's travel path and use the expanded uncertainty region as the robot's virtual operating body;

[0154] Adjustment module 205: Based on the travel path, predict the distance between the virtual manipulator and obstacles, and adjust the robot's motion parameters when the distance is less than a preset threshold.

[0155] The medical robot ward round navigation system of this application aims to solve the problems of decreased positioning accuracy and insufficient navigation robustness caused by the gradual drift of sensing devices in existing medical robots during long-term operation. Specifically, the acquisition module 201 is used to acquire the uncertainty region of the robot's own position. The specific method for acquiring the uncertainty region of the robot's own position has been described in the above embodiments and will not be repeated here. It should be emphasized that the acquisition module 201 can be an independent software component running on the robot's main control unit, responsible for receiving the initial position estimate and its corresponding covariance matrix from the robot's positioning system (e.g., SLAM system) and converting it into a geometric uncertainty region (e.g., ellipse or polygon) for representation. As a preferred embodiment, the acquisition module 201 can also be a hardware-software combined unit, for example, containing a dedicated processor for real-time processing of sensor data and calculation of the uncertainty region, or for obtaining more accurate initial positioning information by communicating with an external positioning base station and generating the uncertainty region accordingly.

[0156] The monitoring module 202 is used to monitor the consistency between the output data of the auxiliary positioning sensing device and the positioning result of the main sensing device, in order to identify gradual drift of the auxiliary positioning sensing device. The specific method for monitoring the consistency between the output data of the auxiliary positioning sensing device and the positioning result of the main sensing device has been described in the above embodiments and will not be repeated here. It should be emphasized that the monitoring module 202 can be configured as a data fusion and analysis unit, which receives raw output data from the auxiliary positioning sensing device (e.g., wheeled odometer, inertial measurement unit) and positioning results from the main sensing device (e.g., lidar, camera). This module can use statistical analysis methods, such as residual analysis using Kalman filtering or particle filtering, to continuously evaluate the consistency between the two data sources. For example, the monitoring module can periodically calculate the root mean square error (RMSE) between the two positioning results and compare it with historical data to detect whether there is a continuously increasing trend.

[0157] In practical applications, the expansion module 203 is used to expand the uncertainty region of the robot's own position based on the identified progressive drift. The specific method for expanding the uncertainty region of the robot's own position based on the identified progressive drift has already been described in the above embodiments and will not be repeated here. It should be emphasized that the expansion module 203 can be a decision logic unit. When the monitoring module reports the existence of progressive drift, this module dynamically adjusts the parameters of the uncertainty region according to the degree of drift (e.g., the cumulative amount or rate of change of the drift). For example, if the degree of drift is small, the expansion module 203 may increase the semi-axis length of the uncertainty region by a small fixed value; if the degree of drift is large, it may increase the area or volume of the uncertainty region proportionally. As a specific implementation, the expansion module 203 can also determine the expansion amount of the uncertainty region by consulting a preset drift-expansion mapping table, which can be trained and optimized based on the robot's actual operating data.

[0158] Furthermore, the planning module 204 is used to plan the robot's path based on the expanded uncertainty region, and to treat the expanded uncertainty region as the robot's virtual manipulator. The specific method for planning the robot's path based on the expanded uncertainty region has already been described in the above embodiments, and will not be repeated here. It should be emphasized that the planning module 204 can be an implementation of a path planning algorithm, such as the A* algorithm based on a grid map, the Dijkstra algorithm, or the RRT (Fast Random Tree) algorithm based on sampling points. When performing path search, this module treats the robot as a "virtual manipulator" defined by its physical size and the expanded uncertainty region, thus occupying a larger safe space on the map. For example, when calculating path cost, the planning module 204 considers the minimum distance between the virtual manipulator and obstacles, and ensures that the virtual manipulator does not collide with obstacles throughout the entire path.

[0159] Furthermore, the adjustment module 205 is used to predict the distance between the virtual manipulator and obstacles based on the travel path, and adjust the robot's motion parameters when the distance is less than a preset threshold. The specific method for predicting the distance between the virtual manipulator and obstacles and adjusting the robot's motion parameters based on the travel path has been described in the above embodiments, and will not be repeated here. It should be emphasized that the adjustment module 205 can be a motion controller that receives the travel path from the planning module 204 and real-time obstacle information from sensors (e.g., LiDAR, ultrasonic sensors). This module continuously predicts the trajectory of the virtual manipulator over a future period and calculates the potential collision risk between it and obstacles in the environment. For example, the adjustment module can employ model predictive control (MPC) to dynamically adjust the robot's speed, acceleration, or turning angular velocity based on the predicted collision risk in each control cycle to ensure that the virtual manipulator always maintains a safe distance from obstacles. In some embodiments, the adjustment module can also generate an emergency stop command to immediately stop the robot when an imminent collision risk is detected.

[0160] Through the above technical solution, the medical robot ward round navigation system of this application demonstrates significant technological advancements compared to existing technologies. Traditional medical robot navigation systems often struggle to effectively identify and compensate for the gradual drift caused by the long-term operation of auxiliary positioning sensing devices, leading to a gradual decrease in robot positioning accuracy and resulting in hesitation, inefficiency, or even minor collisions when performing precise tasks. The system of this application, by introducing a dedicated monitoring module 202 and an expansion module 203, can sensitively identify this imperceptible gradual drift and dynamically adjust the robot's position within uncertain areas accordingly. This proactive compensation mechanism allows the planning module 204 to treat the robot as a more conservative "virtual operating body" during path planning, thus reserving a greater safety margin. Finally, the adjustment module 205 can adjust motion parameters in real time based on this more accurate and conservative position estimate, effectively avoiding potential collision risks. Therefore, the system of this application not only improves the navigation robustness and safety of medical robots during long-term operation but also significantly enhances their task execution efficiency and reliability in complex medical environments, providing a solid technical guarantee for the widespread application of medical robots.

[0161] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0162] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for ward round navigation of a medical robot, characterized in that, The method includes the following steps: S1: Obtain the uncertain region of the robot's own position; S2: Monitor the consistency between the output data of the auxiliary positioning sensing device and the positioning result of the main sensing device to identify progressive drift of the auxiliary positioning sensing device; S3: Based on the identified progressive drift, expand the uncertainty region of the robot's own position; S4: Based on the expanded uncertainty region, plan the robot's travel path and use the expanded uncertainty region as the robot's virtual operating body; S5: Based on the travel path, predict the distance between the virtual manipulator and the obstacle, and adjust the robot's motion parameters when the distance is less than a preset threshold.

2. The medical robot ward round navigation method according to claim 1, characterized in that, Step S5 and the following steps include: S6: Obtain the area of ​​the expanded uncertainty region; S7: Compare the area with a preset area threshold. When the area is greater than or equal to the preset area threshold, adopt a conservative adjustment strategy to adjust the robot's motion parameters. The conservative adjustment strategy includes at least the following: S71: Perform a preset small movement and update the uncertain area of ​​the robot's own position after the movement; S72: Based on the updated uncertainty region, obtain the deviation between the robot's current position and the target alignment point; S73: Based on the deviation, determine whether another small movement is needed; S74: If necessary, repeat small-amplitude movements, updates of uncertain regions, acquisition of deviations, and judgment of deviations until the robot and the target alignment point reach the preset alignment accuracy requirement.

3. The medical robot ward round navigation method according to claim 2, characterized in that, Step S72 includes: S721: Calculate the overlap probability between the updated uncertainty region and the target alignment point; S722: Compare the overlap probability with a preset probability threshold range. If the overlap probability is greater than the upper limit of the preset probability threshold range, then determine that the deviation corresponds to a level one deviation. The deviation value corresponding to the level one deviation is 0.01, and the unit is meters. S723: If the overlap probability is less than or equal to the upper limit of the set probability threshold range and greater than or equal to the lower limit of the set probability threshold range, then the level corresponding to the deviation is determined to be a level two deviation, and the deviation value corresponding to the level two deviation is 0.05, in meters; S724: If the overlap probability is less than the lower limit of the set probability threshold range, then the level corresponding to the deviation is determined to be a level three deviation, and the deviation value of the level three deviation is 0.1, in meters.

4. The medical robot ward round navigation method according to claim 1, characterized in that, Step S1 includes: S11: Obtain the robot's real-time perception data; S12: Determine the robot's current position information and its surrounding environment information based on the real-time sensing data; S13: Based on the current location information, a preliminary uncertain region is preset with the current location information as the center point; S14: Match the environmental information with a preset map. When the matching degree exceeds a preset matching degree threshold, reduce the initial uncertain area by a preset ratio to obtain the uncertain area.

5. The medical robot ward round navigation method according to claim 1, characterized in that, Step S2 includes: S21: Continuously collect the output data of the auxiliary positioning sensing device and the positioning results of the main sensing device; S22: Calculate the positional difference between the output data and the positioning result; S23: Perform a trend analysis on the positional differences and identify the cumulative trend of the differences based on the trend analysis results, thereby identifying the progressive drift of the auxiliary positioning sensing device.

6. The medical robot ward round navigation method according to claim 5, characterized in that, Step S22 includes: S221: Time-align the output data with the positioning result; S222: Convert the time-aligned output data and the positioning result to a unified reference coordinate system; S223: Calculate the geometric distance and angle difference between the pose of the output data and the pose of the positioning result, and use the geometric distance and angle difference as the position difference.

7. The medical robot ward round navigation method according to claim 1, characterized in that, Step S3 includes: S31: Obtain the cumulative amount or rate of change of the gradual drift; S32: Adjust the size of the uncertainty region of the robot's own position according to the accumulated amount or rate of change, so that there is a positive correlation between the size of the uncertainty region and the accumulated amount or rate of change.

8. The medical robot ward round navigation method according to claim 1, characterized in that, Step S4 includes: S41: Perform geometric expansion on the expanded uncertainty region to obtain a safe region; S42: Based on the safe area, a graph search algorithm is used to plan the robot's travel path. The planning process uses the expanded uncertainty area as the robot's virtual operating body.

9. A method for ward round navigation of a medical robot according to claim 1, characterized in that, Step S5 includes: S51: On the travel path, sample multiple future pose points at a preset time step; S52: For each of the plurality of future pose points, based on the uncertainty region, simulate the spatial range that the virtual manipulator may occupy at that future pose point; S53: Based on the spatial range, calculate the closest distance between the virtual operator and the obstacle; S54: If the nearest distance is less than the preset threshold, then adjust the robot's speed, acceleration or turning angular velocity according to the distance difference between the nearest distance and the preset threshold, and generate motion adjustment instructions.

10. A medical robot ward round navigation system, characterized in that, The system for implementing the method according to any one of claims 1-9 comprises: Acquisition module: Acquires the uncertain region of the robot's own position; Monitoring module: Monitors the consistency between the output data of the auxiliary positioning sensing device and the positioning results of the main sensing device to identify progressive drift of the auxiliary positioning sensing device; Extension module: Based on the identified progressive drift, expand the uncertainty region of the robot's own position; Planning module: Based on the expanded uncertainty region, the module plans the robot's travel path and uses the expanded uncertainty region as the robot's virtual operating body; Adjustment module: Based on the travel path, predict the distance between the virtual manipulator and the obstacle, and adjust the robot's motion parameters when the distance is less than a preset threshold.

Images