Intelligent path planning method and system of intelligent medicine distribution robot based on UWB and laser radar

Abstract

The invention discloses an intelligent path planning method for an intelligent medicine delivery robot based on UWB and laser radar, and the method comprises the steps: carrying out the tight coupling fusion of UWB absolute coordinates, laser radar scanning matching data and odometer / IMU track plotting data through a filter, and obtaining a real-time high-precision pose without an accumulative error; and then constructing and maintaining a semantic grid map aligned with the UWB coordinate system. In the path planning stage, a hierarchical strategy is adopted, a task target is analyzed based on a semantic map to plan a static optimal path, and a TEB algorithm is adopted to carry out space-time joint optimization to realize dynamic obstacle avoidance; and performing global re-planning when the path is seriously deviated or the prediction is continuously blocked. According to the invention, the absolute positioning robustness of UWB and the fine environment perception capability of the laser radar are fused, and the hierarchical planning architecture of global guidance and local optimization is combined, so that the drug delivery robot can efficiently, reliably and safely complete autonomous navigation and drug delivery tasks in a complex and dynamic indoor environment.

Description

Technical Field

[0001] This invention belongs to the field of intelligent delivery technology, specifically relating to an intelligent path planning method and system for an intelligent medicine delivery robot based on UWB and LiDAR. Background Technology

[0002] In elderly care facilities, hospital wards, and even smart home environments, automated medicine delivery robots are crucial tools for alleviating human resource shortages and improving service efficiency and accuracy. One of the core capabilities of these robots is autonomous navigation and path planning, whose performance directly determines whether medicine delivery tasks can be completed efficiently and reliably. Existing mobile robot navigation solutions primarily rely on the following technologies: LiDAR-based SLAM (Simultaneous Localization and Mapping) is currently the most mainstream technology. Robots use LiDAR to scan their surroundings, acquiring high-precision 2D or 3D point cloud data, building an environmental map in real time, and locating themselves within the map. Based on this map, path planning algorithms (such as A, D, and RRT) can calculate the path from the starting point to the target point. However, this approach has inherent limitations: Robots rely on wheeled odometry to estimate incremental changes in their pose, but odometry can accumulate errors due to wheel slippage and uneven ground, leading to positioning drift. Although LiDAR can be corrected through scan matching (such as the ICP algorithm), matching is prone to failure in long corridors, symmetrical or feature-sparse environments, and the accumulated error will continue to increase, eventually causing the robot to "get lost".

[0003] LiDAR can detect suddenly appearing dynamic obstacles (such as pedestrians or moving chairs), but traditional global path planners can usually only perform local, reactive obstacle avoidance (such as artificial potential field method) when encountering dynamic obstacles. They lack the ability to reconsider the global path, which can easily lead to the robot getting stuck in local deadlock (such as swinging back and forth in a region) or planning a non-optimal detour path.

[0004] Using a camera as the primary sensor is relatively inexpensive. However, it is greatly affected by lighting conditions, with performance dropping sharply in low light or overexposed conditions. Furthermore, calculating depth and geometric information accurately from images is computationally intensive and susceptible to environmental texture variations, making it less reliable than LiDAR.

[0005] UWB technology can achieve centimeter-level high-precision indoor absolute positioning by measuring the time-of-flight of radio signals. It provides robots with a global coordinate system without accumulated errors. However, it cannot perceive environmental details or identify the shape and size of obstacles, thus it cannot independently perform obstacle avoidance and fine navigation tasks. Using UWB alone, a robot is like moving in a world with only coordinate points but no map, making it extremely prone to collisions.

[0006] Therefore, existing single technical solutions are insufficient to meet the stringent requirements of intelligent drug delivery robots for long-term reliable positioning and efficient dynamic path planning in complex and dynamic indoor environments. Summary of the Invention

[0007] To address the challenges of long-term reliable positioning and efficient dynamic obstacle avoidance for robots in indoor environments, this invention provides an intelligent path planning method for a smart drug delivery robot based on UWB and LiDAR. This method integrates the advantages of multiple technologies and compensates for their respective disadvantages, enabling the robot to function like an experienced human, possessing both a "global map and compass" (absolute sense of position) and "keen eyes" (environmental perception), thus planning and executing the optimal path in various situations.

[0008] According to one aspect of the present invention, an intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR is provided, comprising: By fusing UWB absolute coordinates, LiDAR scan matching data, and odometry and IMU trajectory calculation data through filters, the robot's real-time high-precision pose with no cumulative error is obtained. Based on LiDAR data, a grid map aligned with the UWB coordinate system is constructed and maintained, and key semantic information of room doors and beds is marked to form a semantic map; when a medicine delivery task is received, the semantic location coordinates of the target are parsed, and a global path planning algorithm is used to calculate the static global optimal path from the current pose to the target point; The robot moves along a global path and detects dynamic obstacles in real time using LiDAR. It uses a local path planning algorithm as a local planner to generate a time-parameterized initial trajectory with the global path as a reference. It takes real-time obstacles as constraints and performs spatiotemporal joint optimization with the goals of shortest time, shortest path, away from obstacles, and smooth motion to generate a local obstacle avoidance trajectory in real time. When local planning deviates significantly or for an extended period due to environmental complexity, the robot's current precise pose is used as a new starting point to recalculate the global path to the target and continue local planning and obstacle avoidance.

[0009] As a further technical solution, the method also includes the following system construction steps: At least three UWB base stations are deployed in the working environment, and the robot is equipped with UWB tags, LiDAR, wheeled odometers, and inertial measurement units.

[0010] As a further technical solution, by fusing UWB absolute coordinates, LiDAR scan matching data, and odometry and IMU trajectory calculation data through filters, a real-time high-precision pose of the robot without accumulated errors is obtained, including: First, pose prediction is performed based on odometry and IMU. Then, LiDAR scanning and matching data are used for observation updates and initial correction. Finally, UWB absolute coordinates are introduced for correction to eliminate cumulative drift.

[0011] As a further technical solution, the triggering condition for introducing UWB absolute coordinates for correction is: The real-time confidence level of the LiDAR scan matching is lower than a preset threshold, or the area where the robot is located is pre-marked as a feature degradation area of ​​a long straight corridor or symmetrical space.

[0012] As a further technical solution, in the filter fusion step, UWB absolute coordinates are periodically introduced at set intervals for correction.

[0013] As a further technical solution, during spatiotemporal joint optimization, the weight coefficient of the optimization objective of moving away from obstacles is dynamically adjusted according to the type of obstacle identified by the lidar in real time; when the obstacle is identified as a pedestrian, the weight coefficient is increased to increase the safe distance; when the obstacle is identified as a static temporary obstacle, the weight coefficient is decreased to prioritize ensuring trajectory smoothness and traffic efficiency.

[0014] As a further technical solution, the triggering condition for recalculating the global path to the target also includes: making a predictive judgment based on the density and movement trend of dynamic obstacles within the perception range set in front of the robot; and triggering global replanning in advance when the prediction indicates that passage based on the current global path will continue to be blocked.

[0015] As a further technical solution, in the semantic map, specific key semantic points are associated with preset behavioral strategies; when the robot navigates to the nurse station or room door, the local planner will automatically switch to a behavior mode of low-speed slow movement and enhanced safety distance check.

[0016] According to one aspect of the present invention, an intelligent drug delivery robot is provided, including a UWB tag, a lidar, a wheeled odometer, an IMU, a processor, and a memory. The memory stores a computer program, and when the computer program is executed by the processor, it implements the intelligent path planning method of the UWB and lidar-based intelligent drug delivery robot.

[0017] According to one aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the intelligent path planning method for the intelligent drug delivery robot based on UWB and lidar.

[0018] Compared with existing technologies, this invention, through the deep fusion of UWB and LiDAR, combined with a hierarchical intelligent path planning strategy, brings the following significant benefits: 1. This invention uses the absolute pose provided by UWB as an "anchor" to periodically "pull" the robot's position back to the correct coordinates, ensuring that the robot knows its precise global position throughout the entire mission cycle, regardless of distance or the simplicity of environmental features. This solves the fatal weakness of pure LiDAR-SLAM in large-scale scenes and achieves reliable positioning 24 / 7.

[0019] 2. Unlike traditional reactive obstacle avoidance, the TEB-based local planner performs spatiotemporal joint optimization. It plans not a simple geometric path, but a time-parameterized trajectory that takes into account the robot's motion capabilities. This makes the robot's obstacle avoidance actions more forward-looking, smooth, and efficient, avoiding unnecessary pauses and swaying in front of obstacles, resulting in smoother passage and a shorter overall task time.

[0020] 3. This invention, through its layered architecture, ensures that the robot follows a globally optimal path most of the time. When the local environment changes drastically, the global replanning mechanism can provide timely new global guidance, preventing it from falling into local optima traps (such as dead ends). This makes the robot's behavior both "foresightful" and "quick-witted."

[0021] 4. In the corridors or wards of nursing homes with high foot traffic, dynamic obstacles are commonplace. The method of this invention enables the robot to navigate flexibly and politely in dynamic environments, like a skilled driver, ensuring safety and reliably delivering medication to the user's bedside, greatly expanding the boundaries of its application scenarios.

[0022] 5. Accurate and reliable positioning and intelligent route planning mean that medication delivery tasks can be completed efficiently and without errors. Users (nurses or elderly people) can trust the robot to deliver on time, reducing waiting and searching time and improving the overall quality and efficiency of healthcare services. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating the intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR, as provided in an embodiment of the present invention. Detailed Implementation

[0025] First, the key terms used in this invention specification are explained as follows: UWB (Ultra-Wideband): A wireless carrier communication technology that uses nanosecond-level non-sinusoidal narrow pulses for communication. In this invention, it specifically refers to an indoor positioning technology based on the time-of-flight (TOF) or time difference of arrival (TDOA) measurement principle, capable of achieving centimeter-level accuracy.

[0026] LiDAR (Light Detection and Ranging): A radar system that detects target position, velocity, and other characteristics by emitting a laser beam and measuring its reflected signal. In this invention, it refers to a sensor used for environmental scanning, acquiring high-precision point cloud data for mapping, localization, and obstacle detection.

[0027] Tightly coupled fusion refers to the joint processing of raw data or low-level processing results from different sensors within a unified mathematical model (such as a state-space model). In this invention, it specifically refers to inputting the odometry / IMU predictions, the lidar's matched observations, and the UWB's absolute position observations into the state update equation of a filter (such as an EKF) for optimal state estimation.

[0028] Extended Kalman Filter (EKF) and Particle Filter (PF): Both are commonly used state estimation algorithms for nonlinear systems. The EKF approximates the nonlinear system using a first-order Taylor expansion to achieve local linearization; the PF approximates the posterior probability distribution of the system state using a set of random samples (particles) and their weights, making it suitable for strongly nonlinear and non-Gaussian systems. In this invention, both can be used as core algorithms for achieving tightly coupled fusion of multi-source sensors.

[0029] Semantic maps are a form of map representation that includes not only environmental geometric information (such as occupancy grids) but also semantic labels for specific objects or areas within the environment (such as "room door," "bed 101," and "nurse station"). They elevate environmental information from a geometric level to a knowledge level.

[0030] The Timed Elastic Band (TEB) algorithm is an optimization-based local motion planning method. It discretizes the robot's motion trajectory into an "elastic band" consisting of a series of pose points with timestamps, and simultaneously adjusts the position and time interval of these pose points through optimization algorithms to meet the robot's kinematic constraints, dynamic constraints, obstacle avoidance constraints, and various optimization objectives (such as shortest time, path smoothness, etc.).

[0031] Spatiotemporal joint optimization refers to the simultaneous consideration of optimality in both spatial (geometric path) and temporal (velocity, acceleration profile) dimensions during the planning process. The TEB algorithm is a typical spatiotemporal joint optimizer, and its planning result is a time-parameterized trajectory, not just a spatial path.

[0032] Feature degradation regions refer to areas in the environment where geometric features are sparse, repetitive, or highly symmetrical, such as long straight corridors, open halls, or doorways of multiple structurally similar rooms. In these areas, feature-matching-based LiDAR SLAM algorithms are prone to performance degradation or failure due to matching ambiguity.

[0033] Global path replanning refers to the process of completely recalculating a global path from the current position to the target point when the environment undergoes significant changes (such as the original path being blocked for a long time) or the robot deviates significantly from the original path due to obstacle avoidance.

[0034] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. In addition, the technical features of the various embodiments or individual embodiments provided by the present invention can be arbitrarily combined to form new technical solutions. Such combinations are not bound by the order of steps and / or structural composition patterns, but must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0035] like Figure 1 As shown, this embodiment of the invention provides an intelligent path planning method for an intelligent medicine delivery robot based on UWB and LiDAR, including the following steps: Step S100: System construction and environment deployment.

[0036] First, deploy at least three UWB base stations in the target working environment (e.g., nursing home corridors or rooms). These base stations are synchronized via wired or wireless means to form an indoor positioning network covering the working area. Simultaneously, integrate the following key components onto the drug delivery robot: a UWB tag for communicating with the base stations and calculating the robot's absolute coordinates (x, y); a 2D or 3D LiDAR for environmental perception, such as scanning the surrounding environment and acquiring obstacle point clouds; a wheeled odometry and inertial measurement unit (IMU) for measuring its own motion increments, such as providing instantaneous velocity and attitude information; and a processing and control unit (including a processor and memory).

[0037] Step S200: Tightly coupled localization and map building using multi-source sensors.

[0038] After the robot is started, it runs SLAM and localization processes simultaneously.

[0039] State prediction: In each filtering cycle, based on the state estimate of the previous moment and the displacement and rotation increments provided by the odometry and IMU in the current cycle, the robot's pose (including position and orientation) at the current moment is predicted through the system motion model.

[0040] LiDAR Observation Update: Point cloud data of the current environment is acquired using LiDAR. This actual scanned point cloud is matched against the expected scan generated based on the predicted pose and the maintained grid map (e.g., using the Iterative Closest Point (ICP) algorithm or its variants), and the observation error (i.e., the matching residual) is calculated. This observation error is input into an Extended Kalman Filter (EKF) to complete an observation update, thereby performing the initial optimization correction of the predicted pose. This step provides high-frequency, relatively accurate local relative positioning.

[0041] UWB Absolute Observation Update: The UWB localization engine periodically calculates the absolute two-dimensional coordinates (x, y) of the tag. These absolute coordinates are then introduced into the EKF as an additional observation source. Since UWB localization does not have cumulative errors, this observation effectively corrects for cumulative localization deviations caused by odometry drift and LiDAR matching errors in feature degradation areas. Preferably, the UWB update can be triggered in two modes: periodic triggering (e.g., once every 1 second) or conditional triggering. One embodiment of conditional triggering is: real-time calculation of the confidence level of LiDAR scan matching (e.g., matching score, proportion of interior points to total points). When the confidence level falls below a preset threshold (e.g., 0.6), or when the robot determines, based on its current estimated position, that it has entered a pre-marked "long straight corridor" area on the map, a UWB observation update is immediately triggered. This intelligent triggering mechanism optimizes computational and communication resources while ensuring localization accuracy.

[0042] Through the aforementioned tight coupling fusion, the robot can continuously obtain a high-precision, drift-free real-time global pose.

[0043] Simultaneously, the robot uses LiDAR data, with the aforementioned fused pose as a reference, to incrementally construct and maintain a globally consistent grid map. This map is aligned with the UWB network coordinate system from the outset. Furthermore, semantic labels, such as "pharmacy," "nurse station," "elevator," "room 301 door," and "bed 301-1," are added to key points in the map through manual annotation or automatic recognition (combined with visual sensors), forming a semantic map.

[0044] Step S300: Global path planning based on semantic map.

[0045] When the central dispatch system or local interactive interface issues a medicine delivery task (e.g., "deliver to bed 1 in room 301"), the route planning system first parses the instruction and queries the semantic map for the coordinates of "bed 301-1".

[0046] Then, the A* algorithm (a heuristic graph search algorithm) is used as the global path planner. Starting from the robot's current fused localization pose and ending at the target bed coordinates, this algorithm searches within the passable area (non-obstacle grids) of the grid map to calculate a statically global optimal path that avoids all static obstacles (such as walls and fixed cabinets), thus achieving global path planning based on the semantic map. This path consists of a series of continuous grid points or pose points, providing the robot with macroscopic navigation guidance.

[0047] Step S400: Local real-time trajectory optimization and dynamic obstacle avoidance.

[0048] The robot begins tracking and moving along a global path. Its onboard LiDAR continuously scans at high frequencies (e.g., 10Hz) to detect dynamic obstacles in the environment that are not marked on the map, such as walking medical personnel, temporarily placed medical equipment carts, and other mobile robots.

[0049] The core of this invention employs the Timed Elastic Band (TEB) algorithm as a local planner, responsible for generating smooth, safe, and executable motion trajectories. The TEB algorithm not only considers the geometric optimality of the path but also introduces a time dimension, directly optimizing the robot's motion trajectory (including changes in position, velocity, and orientation over time). Its workflow is as follows: Initialization: Using the current global path segment as a reference, initialize an "elastic band" consisting of several robot pose points with timestamps.

[0050] Add constraints and objectives: The position information of all obstacles (static and dynamic) detected in real time by the LiDAR is transformed into obstacle avoidance distance constraints (hard constraints) that must be satisfied on the trajectory points. At the same time, an optimization objective function is set, which is a weighted sum of multiple sub-objectives, including: ① Minimize the total execution time of the trajectory; ② Minimize the total length of the trajectory path; ③ Maximize the average distance between each point on the trajectory and the nearest obstacle (i.e., move away from the obstacle); ④ The acceleration and angular acceleration of the trajectory change gradually (smooth motion).

[0051] Spatiotemporal joint optimization: Based on the above constraints and objectives, the TEB optimizer iteratively optimizes the spatial coordinates and time intervals of all pose points on the elastic band, and realizes spatiotemporal joint optimization through this nonlinear optimization process.

[0052] Output and Execution: After optimization convergence, the system outputs an optimal time-parameterized motion trajectory (including position, velocity, and acceleration information) from the current state to the next few seconds. The underlying motion controller tracks this trajectory to drive the robot's movement.

[0053] Real-time re-optimization: TEB performs optimization calculations at a frequency similar to that of LiDAR (e.g., 10Hz). Once the environment changes (such as the appearance of new dynamic obstacles), it can re-optimize a new local trajectory within tens of milliseconds, achieving agile and smooth dynamic obstacle avoidance.

[0054] In a preferred embodiment of the present invention, the weight coefficient of the "avoiding obstacles" target in the TEB algorithm can be dynamically adjusted according to the type of obstacle. For example, through point cloud clustering and simple classification, when the obstacle is identified as a pedestrian, the weight is increased to ensure that the planned trajectory maintains a greater safety margin with respect to the pedestrian; when the obstacle is identified as a stationary temporary cart, the weight is appropriately reduced to allow the robot to bypass it at a shorter distance and with a smoother trajectory, prioritizing traffic efficiency.

[0055] Step S500: Global path replanning decision and execution.

[0056] In complex and dynamic environments, local obstacle avoidance can cause the robot to deviate significantly from its original global path. This invention establishes an intelligent global replanning trigger mechanism. Triggering conditions include, but are not limited to: The robot's current position deviates from the global path by more than a threshold D (e.g., 2 meters); The robot continues to deviate from the global path for more than a threshold T (e.g., 30 seconds). Alternatively, it can be triggered predictively, based on the density and movement trends of dynamic obstacles in the sensing area ahead (such as crowds gathering and moving slowly), predicting that progress along the current global path will continue to be hindered for a period of time. When either condition is met, the system immediately triggers a global replanning.

[0057] During replanning, the robot takes the precise fused pose obtained through step S200 as the new starting point and the original task objective as the endpoint, and re-executes step S300 to calculate a new globally optimal path. Subsequently, the robot immediately switches to tracking this new path and continues to perform local planning and obstacle avoidance in step S400.

[0058] Step S600: Semantic point association behavior strategy.

[0059] To further enhance safety and compliance in specific scenarios, pre-defined behavioral strategies can be associated with specific key semantic points in the semantic map. For example, when the semantic point of "nurse station" or "room door" enters the robot's local planning field of view, the Local Planner (TEB) will automatically invoke the associated strategy: lowering the maximum speed limit and enabling stricter enhanced safety distance checks (e.g., increasing obstacle avoidance distance constraints). This causes the robot to automatically slow down when passing through these sensitive areas, behaving more cautiously and courteously.

[0060] In summary, this invention solves the long-term reliable positioning problem through the tight coupling of UWB and LiDAR, solves the efficient dynamic obstacle avoidance problem through the hierarchical strategy of "global A* + local TEB + global replanning", and realizes scene adaptive behavior by combining semantic information. Together, they constitute a robust, efficient and intelligent complete solution for path planning of drug delivery robots.

[0061] Based on the same inventive concept as the foregoing embodiments, this embodiment of the invention also provides an intelligent medicine delivery robot, including a UWB tag, a lidar, a wheeled odometer, an IMU, a processor, and a memory. The memory stores a computer program, and when the computer program is executed by the processor, it implements the intelligent path planning method for the intelligent medicine delivery robot based on UWB and lidar.

[0062] Based on the same inventive concept as the foregoing embodiments, this embodiment of the invention also provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the intelligent path planning method for the intelligent drug delivery robot based on UWB and LiDAR, including the following steps: By fusing UWB absolute coordinates, LiDAR scan matching data, and odometry and IMU trajectory calculation data through filters, the robot's real-time high-precision pose with no cumulative error is obtained. Based on LiDAR data, a grid map aligned with the UWB coordinate system is constructed and maintained, and key semantic information of room doors and beds is marked to form a semantic map; when a medicine delivery task is received, the semantic location coordinates of the target are parsed, and a global path planning algorithm is used to calculate the static global optimal path from the current pose to the target point; The robot moves along a global path and detects dynamic obstacles in real time using LiDAR. It uses a local path planning algorithm as a local planner to generate a time-parameterized initial trajectory with the global path as a reference. It takes real-time obstacles as constraints and performs spatiotemporal joint optimization with the goals of shortest time, shortest path, away from obstacles, and smooth motion to generate a local obstacle avoidance trajectory in real time. When local planning deviates significantly or for an extended period due to environmental complexity, the robot's current precise pose is used as a new starting point to recalculate the global path to the target and continue local planning and obstacle avoidance.

[0063] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0064] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0065] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0066] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1The steps of the function specified in one or more boxes.

[0067] In summary, this invention discloses an intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR. This method first uses an extended Kalman filter or particle filter to tightly couple and fuse UWB absolute coordinates, LiDAR scan matching data, and odometry / IMU trajectory estimation data to obtain real-time high-precision pose without accumulated errors, fundamentally solving the problem of long-term reliable positioning in feature-degraded environments. Based on this, a semantic grid map aligned with the UWB coordinate system is constructed and maintained. In the path planning stage, a hierarchical strategy is adopted: at the global level, the task objective is analyzed based on the semantic map, and a static optimal path is planned using algorithms such as A*; at the local level, the Timed Elastic Band algorithm is used for spatiotemporal joint optimization, achieving efficient and smooth dynamic obstacle avoidance with a forward-looking time-parameterized trajectory; simultaneously, an intelligent triggering mechanism is established to perform global replanning when the path deviates significantly or is predicted to be continuously obstructed. This invention integrates the absolute positioning robustness of UWB with the fine environmental perception capability of LiDAR, and combines a hierarchical planning architecture of global guidance and local optimization to ensure that the drug delivery robot can efficiently, reliably and safely complete autonomous navigation and drug delivery tasks in complex and dynamic indoor environments.

[0068] The terms “comprising” and “having”, and any variations thereof, in the specification, claims, and accompanying drawings of this invention are intended to cover a non-exclusive inclusion, such as a process, method, system, product, or apparatus that includes a series of steps or units, not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0069] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR, characterized in that, include: By fusing UWB absolute coordinates, LiDAR scan matching data, and odometry and IMU trajectory calculation data through filters, the robot's real-time high-precision pose with no cumulative error is obtained. Based on LiDAR data, a grid map aligned with the UWB coordinate system is constructed and maintained, and key semantic information of room doors and beds is marked to form a semantic map; when a medicine delivery task is received, the semantic location coordinates of the target are parsed, and a global path planning algorithm is used to calculate the static global optimal path from the current pose to the target point; The robot moves along a global path, detecting dynamic obstacles in real time using lidar. A local path planning algorithm is used as a local planner. The initial trajectory is generated with time parameterization with reference to the global path. Real-time obstacles are used as constraints. Spatiotemporal joint optimization is performed with the goals of shortest time, shortest path, away from obstacles and smooth motion, and local obstacle avoidance trajectory is generated in real time. When local planning deviates significantly or for an extended period due to environmental complexity, the robot's current precise pose is used as a new starting point to recalculate the global path to the target and continue local planning and obstacle avoidance.

2. The intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR according to claim 1, characterized in that, The method also includes the following system construction steps: At least three UWB base stations are deployed in the working environment, and the robot is equipped with UWB tags, LiDAR, wheeled odometers, and inertial measurement units.

3. The intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR according to claim 1, characterized in that, By fusing UWB absolute coordinates, LiDAR scan matching data, and odometry and IMU trajectory estimation data through filters, a real-time high-precision pose of the robot without accumulated errors is obtained, including: First, pose prediction is performed based on odometry and IMU. Then, LiDAR scanning and matching data are used for observation updates and initial correction. Finally, UWB absolute coordinates are introduced for correction to eliminate cumulative drift.

4. The intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR according to claim 3, characterized in that, The triggering condition for introducing UWB absolute coordinates for correction is: The real-time confidence level of the LiDAR scan matching is lower than a preset threshold, or the area where the robot is located is pre-marked as a feature degradation area of ​​a long straight corridor or symmetrical space.

5. The intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR according to claim 3, characterized in that, In the filter fusion step, UWB absolute coordinates are periodically introduced at set intervals for correction.

6. The intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR according to claim 1, characterized in that, During spatiotemporal joint optimization, the weight coefficient of the optimization objective of moving away from obstacles is dynamically adjusted according to the type of obstacle identified by the LiDAR in real time. When the obstacle is identified as a pedestrian, the weight coefficient is increased to increase the safe distance. When the obstacle is identified as a static temporary obstacle, the weight coefficient is decreased to prioritize ensuring trajectory smoothness and traffic efficiency.

7. The intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR according to claim 1, characterized in that, The triggering conditions for recalculating the global path to the target also include: making a predictive judgment based on the density and movement trend of dynamic obstacles within the sensory range in front of the robot; and triggering global replanning in advance when the prediction indicates that passage based on the current global path will continue to be blocked.

8. The intelligent path planning method for an intelligent drug delivery robot based on UWB and LiDAR according to claim 1, characterized in that, In the semantic map, specific key semantic points are associated with preset behavioral strategies; when the robot navigates to the nurse station or room door, the local planner will automatically switch to a behavior mode of slow movement and enhanced safety distance check.

9. An intelligent medicine delivery robot, characterized in that, It includes a UWB tag, a lidar, a wheeled odometer, an IMU, a processor, and a memory. The memory stores a computer program, which, when executed by the processor, implements the intelligent path planning method for an intelligent drug delivery robot based on UWB and lidar as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the intelligent path planning method for the intelligent drug delivery robot based on UWB and LiDAR as described in any one of claims 1 to 8.

Images