Intra-hospital intelligent logistics robot autonomous obstacle avoidance system and method

Abstract

The invention discloses an intelligent logistics robot autonomous obstacle avoidance system and method in a hospital, relates to the technical field of logistics robot autonomous obstacle avoidance, and solves the problems that in the prior art, a moving path of a dynamic obstacle cannot be predicted in advance, so that an obstacle avoidance decision is lack of perspectiveness, frequent sudden brake or repeated path adjustment is easily caused, and the obstacle avoidance efficiency is high. And the operation efficiency is reduced. The method comprises the following steps: constructing a three-dimensional model of a hospital; acquiring obstacle data near the robot in real time; acquiring a plurality of hypothetical paths of dynamic obstacle movement based on the obstacle data; and acquiring an obstacle avoidance strategy of the robot based on the obstacle type and a plurality of hypothetical paths, namely, selecting an optimal strategy by analyzing a time window of the robot and the dynamic obstacle at a potential coincidence point and combining the time consumption cost of waiting for avoidance and bypassing avoidance, thereby solving the technical problem.

Description

Technical Field

[0001] The present invention belongs to the field of autonomous obstacle avoidance for logistics robots, and specifically relates to an autonomous obstacle avoidance system and method for intelligent logistics robots in hospitals. Background Art

[0002] With the advancement of the construction of smart hospitals, logistics robots have become an important carrier for transporting medical supplies. However, their autonomous obstacle avoidance in complex nosocomial infection environments still faces severe challenges. Traditional solutions mostly use two-dimensional grid maps or simple topological maps, lacking fine three-dimensional modeling of the hospital (such as semantic information of building areas like rooms and elevators), resulting in insufficient path planning accuracy.

[0003] The existing intelligent logistics robot obstacle avoidance systems in hospitals usually rely on real-time sensor data and simple path planning algorithms (such as A* or Dijkstra algorithms). In complex environments, they mainly bypass static obstacles, while the obstacle avoidance strategies for dynamic obstacles (such as pedestrians and carts) are relatively passive. Often, they only perform emergency avoidance or simple bypass through a fixed safety distance threshold. This method cannot predict the moving paths of dynamic obstacles in advance, resulting in a lack of foresight in obstacle avoidance decisions, prone to frequent sudden braking or repeated path adjustments, and reducing the operating efficiency.

[0004] Therefore, the present invention proposes an autonomous obstacle avoidance system and method for intelligent logistics robots in hospitals to solve the above problems. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art. For this purpose, the present invention proposes an autonomous obstacle avoidance system and method for intelligent logistics robots in hospitals to solve the technical problem that the prior art cannot predict the moving paths of dynamic obstacles in advance, resulting in a lack of foresight in obstacle avoidance decisions, prone to frequent sudden braking or repeated path adjustments, and reducing the operating efficiency.

[0006] To achieve the above object, the first aspect of the present invention provides an autonomous obstacle avoidance system for intelligent logistics robots in hospitals, including: a data acquisition module, a data analysis module, and a path planning module;

[0007] A three-dimensional model construction module: used to construct a three-dimensional model of the hospital;

[0008] A data acquisition module: used to obtain real-time obstacle data near the robot;

[0009] A data analysis module: based on the obstacle data, obtain several hypothetical paths for the movement of dynamic obstacles;

[0010] A path planning module: based on the obstacle type and several hypothetical paths, obtain the obstacle avoidance strategy of the robot.

[0011] Preferably, the construction of the three-dimensional model of the hospital includes:

[0012] Construct a three-dimensional model of the hospital through three-dimensional model construction technology, and establish a space coordinate system with the center point of the three-dimensional model of the hospital as the origin; among them, the three-dimensional model construction technology includes: BIM or CAD.

[0013] Preferably, the acquisition of obstacle data near the robot includes:

[0014] Real-time collect obstacle images in the 360° direction of the robot through an image acquisition device;

[0015] Train an artificial intelligence model with standard training data to obtain an obstacle recognition model; among them, the artificial intelligence model includes: YOLO, Faster R-CNN;

[0016] The standard training data includes training input data and training output data; among them, the training input data is several frames of obstacle images; the training output data is obstacle data; the obstacle data includes: the type of obstacle and the moving speed and moving direction of the dynamic obstacle; the type of obstacle includes: static obstacle and dynamic obstacle; the moving speed includes: minimum moving speed and maximum moving speed.

[0017] It should be noted that the number of obstacle images input for each training is greater than 2;

[0018] The moving direction is represented by a displacement vector in the space coordinate system;

[0019] The static obstacle refers to an object with a fixed position and no movement in the robot's operating environment, such as a hospital bed, fixed equipment, etc.;

[0020] The dynamic obstacle refers to an object or person whose position changes with time in the robot's operating environment, such as patients, doctors, trolleys, stretchers, etc.

[0021] Preferably, the obtaining of several hypothetical paths for the movement of the dynamic obstacle based on the obstacle data includes:

[0022] A1: Call the three-dimensional model of the hospital and obtain the floor plan layout where the dynamic obstacle is located;

[0023] A2: Take the building areas in the floor plan as nodes, divide the passages between the building areas into several sub-passages based on a preset passage width, and take the several sub-passages as edges to obtain the graph structure of the floor where the dynamic obstacle is located; among them, the building areas include rooms, elevators, etc.;

[0024] A3: Obtain the position of the dynamic obstacle in the graph structure and take the position where the dynamic obstacle is located as the starting node;

[0025] A4: Obtain several nodes connected to the starting node through edges based on the moving direction of the dynamic obstacle;

[0026] A5: Extract the paths between the starting node and several nodes to obtain several hypothetical paths.

[0027] It should be noted that dividing the passage between building areas into several sub-passages based on a preset passage width means dividing the passage into several sub-passages that can be passed by pedestrians or hospital beds, etc.;

[0028] In a complex environment, the movable area of the obstacle may be very large. According to the moving direction, the attention can be concentrated on the area where the obstacle may go, reducing the calculation amount and improving the path planning efficiency.

[0029] Preferably, the obstacle avoidance strategy of the robot obtained based on the obstacle type and several hypothetical paths includes:

[0030] For static obstacles: Obtain the optimal path for the robot to bypass the static obstacle through a path planning algorithm; among them, the path planning algorithm includes: A* algorithm or Dijkstra algorithm;

[0031] For dynamic obstacles: Perform obstacle avoidance based on the position of the dynamic obstacle relative to the robot.

[0032] Preferably, the obstacle avoidance based on the position of the dynamic obstacle relative to the robot includes:

[0033] B1: Extract the sub-passage where the robot is located;

[0034] B2: Determine whether the dynamic obstacle is in the same sub-passage as the robot; if so, perform obstacle avoidance based on a preset safety distance threshold; if not, extract several hypothetical paths of the dynamic obstacle and extract the optimal path of the robot;

[0035] B3: Obtain the intersection points between several hypothetical paths and the optimal path to obtain several coincidence points;

[0036] B4: Sort the distances between the coincidence points and the robot in ascending order to obtain a coincidence point sequence;

[0037] B5: Analyze and judge the coincidence point sequence to obtain the obstacle avoidance strategy of the robot.

[0038] Preferably, the obstacle avoidance based on a preset safety distance threshold includes:

[0039] Judge whether the distance between the dynamic obstacle and the robot is greater than the preset safety distance threshold; if so, the robot can pass normally; if not, obtain the optimal avoidance path through a path planning algorithm for avoidance.

[0040] It should be noted that the avoidance path refers to the optimal path obtained by the path planning algorithm to bypass dynamic obstacles, starting from the robot and ending at the destination of the robot.

[0041] Preferably, obtaining the obstacle avoidance strategy of the robot by analyzing and judging the coincidence point sequence includes:

[0042] C1: Based on the minimum moving speed of the dynamic obstacle, obtain the time from the dynamic obstacle to each coincidence point in the coincidence point sequence in turn, and obtain the maximum time sequence;

[0043] C2: Based on the maximum moving speed of the dynamic obstacle, obtain the time from the dynamic obstacle to each coincidence point in the coincidence point sequence in turn, and obtain the minimum time sequence;

[0044] C3: Integrate the minimum time sequence and the maximum time sequence to obtain the time range sequence; where each time range in the time range sequence corresponds to the time from a dynamic obstacle to a coincidence point;

[0045] C4: Obtain the time required for the robot to reach each coincidence point in the coincidence point sequence in turn, and obtain the target time sequence;

[0046] C5: One-to-one correspond the time range sequence and the target time sequence, and judge in turn whether the target time is within the time range; if so, it is determined that the robot and the dynamic obstacle will meet, and the avoidance method of the robot is selected based on the avoidance time data of the robot; if not, the robot passes normally; where the avoidance methods include: the robot waits for avoidance and the robot bypasses for avoidance; the avoidance time data includes: waiting time and bypassing time.

[0047] Preferably, selecting the avoidance method of the robot based on the avoidance time data of the robot includes:

[0048] Calculate the average moving speed of the dynamic obstacle based on the minimum moving speed and the maximum moving speed of the dynamic obstacle;

[0049] If the robot waits for avoidance, obtain the time required for the dynamic obstacle to pass by the robot based on the average moving speed of the dynamic obstacle, and obtain the waiting time;

[0050] Calculate the total waiting time by adding the waiting time and the initial path time of the robot; where the initial path time refers to the time required for the robot to reach the destination before detecting the dynamic obstacle;

[0051] If the robot bypasses for avoidance, obtain the optimal avoidance path through the path planning algorithm for avoidance, and obtain the time required for the optimal avoidance path based on the moving speed of the robot, and obtain the bypassing time;

[0052] Select the avoidance method of the robot by comparing the target time and the detour time;

[0053] If the target time is greater than the detour time, select the robot to detour and avoid;

[0054] If the target time is equal to the detour time, select the robot to wait and avoid or the robot to detour and avoid;

[0055] If the target time is less than the detour time, select the robot to wait and avoid.

[0056] Based on the minimum and maximum moving speeds of dynamic obstacles, the present invention calculates the time range, combines the time consumption of the robot's detour path and the total waiting avoidance time, and through multi-dimensional time series comparison, enables the robot to select the optimal avoidance method (waiting or detouring) according to the real-time situation. This not only avoids the frequent emergency braking or repeated path adjustment caused by the existing method's simple reliance on distance thresholds, but also reduces energy consumption through accurate calculation of time costs, while ensuring the global optimality of obstacle avoidance decisions, and significantly improves the efficiency and safety of obstacle avoidance in complex environments.

[0057] The second aspect of the present invention provides an autonomous obstacle avoidance method for an intelligent logistics robot in a hospital, including:

[0058] Step 1: Construct a three-dimensional model of the hospital;

[0059] Step 2: Obtain obstacle data near the robot;

[0060] Step 3: Obtain several assumed paths for the movement of dynamic obstacles based on the obstacle data;

[0061] Step 4: Obtain the obstacle avoidance strategy of the robot based on the obstacle type and several assumed paths.

[0062] Compared with the prior art, the beneficial effects of the present invention are:

[0063] 1. The prior art cannot predict the movement path of dynamic obstacles in advance, resulting in the lack of foresight in obstacle avoidance decisions, which easily leads to frequent emergency braking or repeated path adjustments, reducing the operation efficiency. The present invention establishes a three-dimensional model and a spatial coordinate system including building areas and channels through three-dimensional construction technology, converts the plane layout into a graph structure, takes the building areas as nodes, and divides the channels into sub-channels as edges, providing global semantic information for path planning. The artificial intelligence model is used to identify the type of obstacles and the moving speed and direction of dynamic obstacles in real time, and based on its moving direction, possible hypothetical paths are generated in the graph structure, turning the obstacle avoidance decision from passive response to active prediction. By analyzing the time window of the robot and the dynamic obstacle at the potential coincidence point, and combining the time-consuming costs of waiting for avoidance and detouring avoidance to select the optimal strategy, it not only avoids the frequent emergency braking caused by the lack of path prediction in the existing method, but also reduces the computational complexity in complex scenarios through sub-channel division and graph structure optimization, realizing forward-looking, global, and low-energy consumption obstacle avoidance path planning.

[0064] 2. The present invention divides sub-channels based on the channel width and generates hypothetical paths, which not only compresses the possible movement range of obstacles to reduce redundant calculations, but also realizes the probabilistic assessment of the risk of path crossing through spatio-temporal conflict analysis (calculating the matching degree between the time window corresponding to the speed range of dynamic obstacles and the time series of the robot reaching the coincidence point). Hierarchical decision-making is carried out according to whether the dynamic obstacle and the robot are in the same sub-channel (direct safety distance obstacle avoidance or spatio-temporal path conflict prediction), which not only ensures real-time response in narrow channel scenarios, but also dynamically selects the global optimal strategy by quantifying the time cost difference between waiting for avoidance and detouring avoidance, so as to balance the obstacle avoidance safety and path planning efficiency in complex dynamic environments, and avoid the problems of frequent start-stop or long detour paths caused by misjudgment of path conflicts in the existing method. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0066] Figure 1 It is a schematic diagram of the system module of the embodiment of the present invention;

[0067] Figure 2 It is a schematic diagram of the method for obtaining hypothetical paths of the embodiment of the present invention;

[0068] Figure 3 It is a schematic diagram of the method steps of the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0069] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative work belong to the scope of protection of the present invention.

[0070] Please refer to Figure 1 , an embodiment of the first aspect of the present invention provides an autonomous obstacle avoidance system for intelligent logistics robots in a hospital, including: a data acquisition module, a data analysis module, and a path planning module;

[0071] Three-dimensional model construction module: used to construct a three-dimensional model of the hospital;

[0072] Data acquisition module: used to obtain real-time obstacle data near the robot;

[0073] Data analysis module: obtain several hypothetical paths for the movement of dynamic obstacles based on the obstacle data;

[0074] Path planning module: obtain the obstacle avoidance strategy of the robot based on the obstacle type and several hypothetical paths.

[0075] Constructing a three-dimensional model of the hospital includes:

[0076] Construct a three-dimensional model of the hospital through three-dimensional model construction technology, and establish a space coordinate system with the center point of the three-dimensional model of the hospital as the origin; wherein, the three-dimensional model construction technology includes: BIM or CAD.

[0077] Obtaining obstacle data near the robot includes:

[0078] Real-time collect obstacle images in the 360° direction of the robot through an image acquisition device;

[0079] Train an artificial intelligence model with standard training data to obtain an obstacle recognition model; wherein, the artificial intelligence model includes: YOLO, Faster R-CNN;

[0080] The standard training data includes training input data and training output data; wherein, the training input data is several frames of obstacle images; the training output data is obstacle data; the obstacle data includes: the type of the obstacle and the moving speed and moving direction of the dynamic obstacle; the type of the obstacle includes: static obstacle and dynamic obstacle; the moving speed includes: minimum moving speed and maximum moving speed.

[0081] Refer to Figure 2 , obtaining several hypothetical paths for the movement of dynamic obstacles based on the obstacle data includes:

[0082] A1: Invoke the 3D model of the hospital and obtain the floor plan of the floor where the dynamic obstacle is located;

[0083] A2: Use the building areas in the floor plan as nodes, divide the passages between the building areas into several sub-passages based on a preset passage width, and use the several sub-passages as edges to obtain the graph structure of the floor where the dynamic obstacle is located; wherein, the building areas include rooms, elevators, etc.;

[0084] A3: Obtain the position of the dynamic obstacle in the graph structure and use the position where the dynamic obstacle is located as the starting node;

[0085] A4: Based on the moving direction of the dynamic obstacle, obtain several nodes connected to the starting node through edges;

[0086] A5: Extract the paths between the starting node and the several nodes to obtain several hypothetical paths.

[0087] Obtain the obstacle avoidance strategy of the robot based on the obstacle type and the several hypothetical paths, including:

[0088] For static obstacles: Obtain the optimal path for the robot to bypass the static obstacle through a path planning algorithm; wherein, the path planning algorithm includes: A* algorithm or Dijkstra algorithm;

[0089] For dynamic obstacles: Perform obstacle avoidance based on the position of the dynamic obstacle relative to the robot.

[0090] Performing obstacle avoidance based on the position of the dynamic obstacle relative to the robot includes:

[0091] B1: Extract the sub-passage where the robot is located;

[0092] B2: Determine whether the dynamic obstacle is in the same sub-passage as the robot; if so, perform obstacle avoidance based on a preset safety distance threshold; if not, extract several hypothetical paths of the dynamic obstacle and extract the optimal path of the robot;

[0093] B3: Obtain the intersection points between the several hypothetical paths and the optimal path to obtain several coincidence points;

[0094] B4: Sort the coincidence points in ascending order according to the distance between the coincidence points and the robot to obtain a coincidence point sequence;

[0095] B5: Obtain the obstacle avoidance strategy of the robot by analyzing and judging the coincidence point sequence.

[0096] Performing obstacle avoidance based on a preset safety distance threshold includes:

[0097] Determine whether the distance between the dynamic obstacle and the robot is greater than the preset safe distance threshold; if so, the robot can move forward normally; if not, obtain the optimal avoidance path through the path planning algorithm for avoidance.

[0098] Obtain the obstacle avoidance strategy of the robot by analyzing and judging the coincidence point sequence, including:

[0099] C1: Based on the minimum moving speed of the dynamic obstacle, obtain the time from the dynamic obstacle to each coincidence point in the coincidence point sequence in turn, and get the maximum time sequence;

[0100] C2: Based on the maximum moving speed of the dynamic obstacle, obtain the time from the dynamic obstacle to each coincidence point in the coincidence point sequence in turn, and get the minimum time sequence;

[0101] C3: Integrate the minimum time sequence and the maximum time sequence to obtain the time range sequence; where each time range in the time range sequence corresponds to the time from a dynamic obstacle to a coincidence point;

[0102] C4: Obtain the time required for the robot to reach each coincidence point in the coincidence point sequence in turn, and get the target time sequence;

[0103] C5: One-to-one correspond the time range sequence and the target time sequence, and judge in turn whether the target time is within the time range; if so, it is determined that the robot and the dynamic obstacle will meet, and the avoidance method of the robot is selected based on the avoidance time data of the robot; if not, the robot moves forward normally; where the avoidance methods include: the robot waits for avoidance and the robot bypasses for avoidance; the avoidance time data includes: waiting time and bypassing time.

[0104] Select the avoidance method of the robot based on the avoidance time data of the robot, including:

[0105] Calculate the average moving speed of the dynamic obstacle based on the minimum moving speed and the maximum moving speed of the dynamic obstacle;

[0106] If the robot waits for avoidance, obtain the time required for the dynamic obstacle to pass by the robot based on the average moving speed of the dynamic obstacle, and get the waiting time;

[0107] Calculate the total waiting time by adding the waiting time and the initial path time of the robot; where the initial path time refers to the time required for the robot to reach the destination before detecting the dynamic obstacle;

[0108] If the robot bypasses for avoidance, obtain the optimal avoidance path through the path planning algorithm for avoidance, and obtain the time required for the optimal avoidance path based on the moving speed of the robot, and get the bypassing time;

[0109] Select the avoidance method of the robot by comparing the target time and the bypass time;

[0110] If the target time is greater than the bypass time, select the robot to bypass and avoid;

[0111] If the target time is equal to the bypass time, select the robot to wait and avoid or the robot to bypass and avoid;

[0112] If the target time is less than the bypass time, select the robot to wait and avoid.

[0113] For example: Suppose there are three rooms on both sides of the corridor on the third floor of a hospital. The two rooms on the left are Node A and Node B respectively, and the one room on the right is Node C. The logistics robot needs to pass through the current corridor to deliver medicine to other places, such as Node E; divide the current corridor into several sub-corridors based on the preset corridor width. Suppose it is divided into three sub-corridors with a width of 0.8 meters, and the sub-corridors are marked as Sub-corridor 1, Sub-corridor 2, and Sub-corridor 3 in order from left to right;

[0114] Take the current position of the robot (in Sub-corridor 2) as Node M, and its preset optimal path is: M → Sub-corridor 2 →... → E; At this time, a nurse is pushing an equipment cart (dynamic obstacle) from the corridor (in Sub-corridor 1) towards the robot head-on at a speed range of 0.5 - 1.2 m / s. After the robot detects her, it makes an avoidance, specifically as follows:

[0115] 1. Environmental modeling and data collection;

[0116] The three-dimensional model construction module generates a three-dimensional model of the hospital through BIM technology. The corridor width is 2.4 meters, and it is divided into 3 sub-corridors (numbered 1, 2, 3 from left to right) according to the preset sub-corridor width of 0.8 meters.

[0117] The robot captures the image of the cart through a 360° camera, and the YOLO model identifies it as a dynamic obstacle, and outputs its current position (set as Node N), the moving direction vector (along Sub-corridor 1 pointing to the robot), and the speed range (0.5 - 1.2 m / s).

[0118] 2. Prediction of the path of the dynamic obstacle;

[0119] The data analysis module calls the floor plan structure of the third floor, takes Node N as the starting point, and filters the connected nodes (A, B, C) according to the moving direction.

[0120] Generate three hypothetical paths, which are respectively:

[0121] N → Sub-corridor 1A;

[0122] N → Sub-corridor 1 → B;

[0123] N → Sub-channel 1 → Sub-channel 2 → Sub-channel 3 → C.

[0124] 3. Space-time conflict analysis and obstacle avoidance decision-making;

[0125] The path planning module extracts the intersection point of the robot's current path and the assumed path of the cart: There will be an intersection point (coincidence point P) between Sub-channel 2 and the third assumed path.

[0126] Time window calculation:

[0127] Time range for the cart to reach point P: The distance is 10 meters, the longest time = 10 / 0.5 = 20 seconds, and the shortest time = 10 / 1.2 ≈ 8.3 seconds, that is, the time range is [8.3s, 20s].

[0128] Time for the robot to reach point P: The distance is 15 meters and the speed is 1 m / s, so it takes 15 seconds (target time).

[0129] Decision-making judgment: The robot's target time of 15 seconds is within the cart's time window [8.3s, 20s], and it is determined that a conflict may occur.

[0130] Selection of avoidance method:

[0131] If waiting for avoidance: The average speed of the cart is 0.85 m / s, assuming the passing time is 2 seconds, and the robot needs to wait for the cart to pass and then continue to move forward;

[0132] Assuming the initial path time is 20 seconds, the total waiting time is 22 seconds.

[0133] If bypassing for avoidance: Generate a new path to bypass the cart through the path planning algorithm, and the path increases by 5 meters, then the bypassing time is 25 seconds.

[0134] Since 22 seconds (waiting) < 25 seconds (bypassing), the waiting for avoidance is selected.

[0135] Refer to Figure 3 , the second aspect embodiment of the present invention provides an autonomous obstacle avoidance method for an intelligent logistics robot in a hospital, including:

[0136] Step 1: Construct a three-dimensional model of the hospital;

[0137] Step 2: Obtain obstacle data near the robot;

[0138] Step 3: Obtain several assumed paths for the movement of dynamic obstacles based on the obstacle data;

[0139] Step 4: Obtain the obstacle avoidance strategy of the robot based on the obstacle type and several assumed paths.

[0140] Some of the data in the above formula is calculated by removing the dimension and taking its value. The formula is the one closest to the actual situation obtained through software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0141] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An autonomous obstacle avoidance system for intelligent logistics robots in a hospital, characterized in that, Including: A data acquisition module, a data analysis module, and a path planning module; A three-dimensional model construction module: used to construct a three-dimensional model of the hospital; The data acquisition module: used to obtain the obstacle data near the robot in real time; The data analysis module: based on the obstacle data, obtain several hypothetical paths for the movement of dynamic obstacles; The path planning module: based on the obstacle type and several hypothetical paths, obtain the obstacle avoidance strategy of the robot.

2. The autonomous obstacle avoidance system of an intelligent logistics robot in a hospital according to claim 1, wherein The construction of the three-dimensional model of the hospital includes: Construct a three-dimensional model of the hospital through three-dimensional model construction technology, and establish a space coordinate system with the center point of the three-dimensional model of the hospital as the origin; among them, the three-dimensional model construction technology includes: BIM or CAD.

3. An autonomous obstacle avoidance system for an intelligent logistics robot in a hospital according to claim 1, characterized in that, The obtaining of the obstacle data near the robot includes: Real-time collect the obstacle images in the 360° direction of the robot through an image acquisition device; Train the artificial intelligence model with standard training data to obtain an obstacle recognition model; among them, the artificial intelligence model includes: YOLO, Faster R-CNN; The standard training data includes training input data and training output data; among them, the training input data is several frames of obstacle images; the training output data is obstacle data; the obstacle data includes: the type of obstacle and the moving speed and moving direction of the dynamic obstacle; the type of obstacle includes: static obstacle and dynamic obstacle; the moving speed includes: the minimum moving speed and the maximum moving speed.

4. An autonomous obstacle avoidance system for an intelligent logistics robot in a hospital according to claim 1, characterized in that, The obtaining of several hypothetical paths for the movement of dynamic obstacles based on the obstacle data includes: A1: Call the three-dimensional model of the hospital and obtain the floor plan layout of the floor where the dynamic obstacle is located; A2: Take the building areas in the floor plan layout as nodes, divide the channels between the building areas into several sub-channels based on a preset channel width, and take the several sub-channels as edges to obtain the graph structure of the floor where the dynamic obstacle is located; A3: Obtain the position of the dynamic obstacle in the graph structure and take the position where the dynamic obstacle is located as the starting node; A4: Based on the moving direction of the dynamic obstacle, obtain several nodes connected to the starting node through edges; A5: Extract the paths between the starting node and the several nodes to obtain several hypothetical paths.

5. An autonomous obstacle avoidance system for an intelligent logistics robot in a hospital according to claim 1, characterized in that, The obtaining of the obstacle avoidance strategy of the robot based on the obstacle type and several hypothetical paths includes: For static obstacle avoidance: Obtain the optimal path for the robot to bypass the static obstacle through a path planning algorithm; among them, the path planning algorithm includes: A* algorithm or Dijkstra algorithm; For dynamic obstacle avoidance: Avoid obstacles based on the position of the dynamic obstacle relative to the robot.

6. The autonomous obstacle avoidance system of an intelligent logistics robot in a hospital according to claim 5, characterized in that, The avoiding of obstacles based on the position of the dynamic obstacle relative to the robot includes: B1: Extract the sub-channel where the robot is located; B2: Determine whether the dynamic obstacle is in the same sub-channel as the robot; if so, avoid obstacles based on a preset safety distance threshold; if not, extract several hypothetical paths of the dynamic obstacle and extract the optimal path of the robot; B3: Obtain the intersection points between the several hypothetical paths and the optimal path to obtain several coincidence points; B4: Sort the distances between the coincidence points and the robot in ascending order to obtain a coincidence point sequence; B5: Obtain the obstacle avoidance strategy of the robot by analyzing and judging the coincidence point sequence.

7. An autonomous obstacle avoidance system for an intelligent logistics robot in a hospital according to claim 6, characterized in that, The obstacle avoidance based on the preset safety distance threshold includes: Judge whether the distance between the dynamic obstacle and the robot is greater than the preset safety distance threshold; if so, the robot moves forward normally; if not, obtain the optimal avoidance path through the path planning algorithm for avoidance.

8. An autonomous obstacle avoidance system for an intelligent logistics robot in a hospital according to claim 6, characterized in that, The obtaining of the obstacle avoidance strategy of the robot by analyzing and judging the coincidence point sequence includes: C1: Based on the minimum moving speed of the dynamic obstacle, sequentially obtain the time from the dynamic obstacle to each coincidence point in the coincidence point sequence to obtain the maximum time sequence; C2: Based on the maximum moving speed of the dynamic obstacle, sequentially obtain the time from the dynamic obstacle to each coincidence point in the coincidence point sequence to obtain the minimum time sequence; C3: Integrate the minimum time sequence and the maximum time sequence to obtain the time range sequence; wherein, each time range in the time range sequence corresponds to the time from a dynamic obstacle to a coincidence point; C4: Sequentially obtain the time required for the robot to reach each coincidence point in the coincidence point sequence to obtain the target time sequence; C5: One-to-one correspond the time range sequence and the target time sequence, and sequentially judge whether the target time is within the time range; if so, determine that the robot and the dynamic obstacle will meet, and select the avoidance method of the robot based on the avoidance time data of the robot; if not, the robot moves forward normally; wherein, the avoidance methods include: the robot waits for avoidance and the robot bypasses for avoidance; the avoidance time data includes: waiting time and bypassing time.

9. The autonomous obstacle avoidance system of an intelligent logistics robot in a hospital according to claim 8, characterized in that, The selection of the avoidance method of the robot based on the avoidance time data of the robot includes: Calculate the average moving speed of the dynamic obstacle based on the minimum moving speed and the maximum moving speed of the dynamic obstacle; If the robot waits for avoidance, obtain the time required for the dynamic obstacle to pass by the robot based on the average moving speed of the dynamic obstacle to obtain the waiting time; Obtain the total waiting time by calculating the sum of the waiting time and the initial path time of the robot; wherein, the initial path time refers to the time required for the robot to reach the destination before detecting the dynamic obstacle; If the robot bypasses for avoidance, obtain the optimal avoidance path through the path planning algorithm for avoidance, and obtain the time required for the optimal avoidance path based on the moving speed of the robot to obtain the bypassing time; Select the avoidance method of the robot by comparing the target time and the bypassing time; If the target time is greater than the bypassing time, select the robot to bypass for avoidance; If the target time is equal to the bypassing time, select the robot to wait for avoidance or the robot to bypass for avoidance; If the target time is less than the bypassing time, select the robot to wait for avoidance.

10. A method for autonomous obstacle avoidance of an intelligent logistics robot in a hospital, which is applied to an autonomous obstacle avoidance system of an intelligent logistics robot in a hospital according to any one of claims 1-9, and is characterized in that, It includes: Step 1: Construct a three-dimensional model of the hospital; Step 2: Obtain the obstacle data near the robot; Step 3: Based on the obstacle data, obtain several assumed paths for the movement of the dynamic obstacle; Step 4: Based on the obstacle type and several assumed paths, obtain the obstacle avoidance strategy of the robot.

Images