Indoor path planning method for dual-arm transfer and nursing robot

Abstract

The application provides a double-arm transfer nursing robot indoor path planning method, comprising the following steps: constructing an obstacle map, determining a search node according to a current point and an end point, and calculating the total cost of each search node, wherein the total cost is the sum of the cost of the distance from the starting point, the cost of the distance from the end point and the safety cost; the search node with the minimum total cost is taken as a known node and stored in a second list, and the remaining search nodes except the known node are stored in a first list; the search node with the minimum total cost in the first list is selected as the current point; all the known nodes in the second list are connected according to the index to obtain an initial global path, and the redundant known nodes are deleted to obtain an optimized global path; based on each important node in the optimized global path, an improved dynamic window algorithm is used to plan a local path, the local path is planned based on the optimized global path, and the double-arm transfer nursing robot is guided to find a path, so that the nursing personnel can be carried faster and more safely.

Description

Technical Field

[0001] This application relates to the field of robot path planning technology, specifically to an indoor path planning method for a dual-arm transfer nursing robot. Background Technology

[0002] With the rapid development of robot control technology, path planning has gradually become an indispensable part of robot motion control. To meet the diverse needs of user groups and the environmental constraints faced by robots under different working conditions, researchers often use different pathfinding principles as design standards. In the context of today's aging population and the prevalence of diseases, the application of dual-arm transfer nursing robots is becoming more widespread. However, dual-arm transfer nursing robots often replicate the pathfinding principles of other robots to control robot movement.

[0003] In existing technologies, dual-arm transfer nursing robots often use the shortest path principle to guide the robot's movement, so as to ensure that the person being cared for can move from the current location to the destination more quickly. However, since the people being cared for are often in poor physical condition, if the shortest path principle is used as the only path-finding principle, it is very likely that the robot will accidentally hit obstacles during the calculation process, resulting in secondary injury to the person. Summary of the Invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, this application aims to provide an indoor path planning method for a dual-arm transfer nursing robot, including the following steps:

[0005] S1. Construct an obstacle map to simulate an indoor environment, mark the start and end points on the obstacle map, create an index for each node on the map, and create a first list and a second list.

[0006] S2. Based on the current point and the destination, determine several search nodes and calculate the total cost of each search node. The total cost is the sum of the cost of distance from the starting point, the cost of distance from the destination, and the security cost.

[0007] S3. The search node with the lowest total cost is designated as a known node and stored in the second list. The remaining search nodes, excluding the known nodes, are stored in the first list.

[0008] S4. Select the search node with the minimum total cost from the first list as the current point, and repeat S2-S3 until the current point reaches the endpoint;

[0009] S5. Connect all the known nodes in the second list according to their indices to obtain the initial global path;

[0010] S6. In the initial global path, identify and delete several redundant known nodes, and take the remaining known nodes as important nodes. Connect all the important nodes according to the index to obtain the optimized global path.

[0011] S7. Based on the important nodes in the optimized global path, an improved dynamic window algorithm is used to plan a local path to guide the dual-arm transfer nursing robot in finding its way.

[0012] According to the technical solution provided in the embodiments of this application, determining a plurality of search nodes based on the current point and the endpoint includes at least the following steps:

[0013] Calculate the first angle, which is the angle between the line connecting the current point and the endpoint and the horizontal line;

[0014] The angle database is retrieved to obtain the search direction group corresponding to the first angle. The search direction group includes several search directions, and the nodes in each search direction are the search nodes. The angle database includes at least several angle ranges and search direction groups corresponding to each angle range.

[0015] According to the technical solution provided in the embodiments of this application, calculating the total cost of each search node includes at least the following steps:

[0016] Construct a first computational model. The inputs of the first computational model are the cost of distance to the starting point, the cost of distance to the ending point, and the security cost. The output is the total cost.

[0017] Obtain the cost of the search node from the starting point, the cost of the search node from the ending point, and the distance from the obstacle;

[0018] If the distance between the obstacle and the obstacle is greater than or equal to the first preset distance, the safety cost is replaced by the first value; if the distance between the obstacle and the obstacle is less than the first preset distance, the safety cost is replaced by the second value.

[0019] The cost of the search node from the starting point, the cost from the ending point, and the corresponding security cost are input into the first calculation model to obtain the total cost of the search node.

[0020] According to the technical solution provided in the embodiments of this application, the step of determining and deleting a number of redundant nodes in the initial global path includes at least the following steps:

[0021] Determine the location information of obstacles in the obstacle map and obtain the robot radius of the dual-arm transfer nursing robot;

[0022] Each n adjacent important nodes in the optimized global path are grouped into a node group;

[0023] Calculate the reference distance of the node group. If the reference distance is greater than the robot radius, then the important node located in the middle position of the node group is designated as the redundant node in the node group. The reference distance is the vertical distance between the line connecting the first and last points of the node group and the obstacle.

[0024] Traverse all the node groups to obtain the redundant nodes in each node group, and delete all the redundant nodes.

[0025] According to the technical solution provided in the embodiments of this application, the step of planning a local path based on each important node in the optimized global path using an improved dynamic window algorithm includes at least the following steps:

[0026] At the current point where the dual-arm transfer nursing robot is located, determine the direction of movement toward the dual-arm transfer nursing robot, and the position information of the next important node at the current point; the position information includes at least the presence of obstacles and the absence of obstacles;

[0027] If the location information at the next important node indicates that there are no obstacles, the global path between the current point of the dual-arm transfer nursing robot and the next important node is used as the local path of the corresponding segment.

[0028] According to the technical solution provided in the embodiments of this application, after determining the movement direction toward the dual-arm transfer nursing robot and the position information at the next important node of the current point, the method further includes the following steps:

[0029] If the location information at the next important node indicates the presence of an obstacle, obtain several pre-selected nodes for that important node;

[0030] The first distance between the current point of the dual-arm transfer nursing robot and each of the pre-selected nodes is obtained respectively, and the pre-selected node with the smallest first distance is taken as the updated important node.

[0031] The global path between the current point of the dual-arm transfer nursing robot and the next important node is updated using the updated important nodes to obtain the local path of the corresponding segment.

[0032] According to the technical solution provided in the embodiments of this application, during the movement of the dual-arm transfer nursing robot along a planned local path, the method further includes:

[0033] Acquire several sets of first samples, where the first sample is the linear velocity and angular velocity corresponding to the sampling time;

[0034] Based on each of the first samples, the predicted trajectory corresponding to each of the first samples is obtained;

[0035] Calculate the total trajectory evaluation of each predicted trajectory, and take the linear velocity and angular velocity of the predicted trajectory corresponding to the largest total trajectory evaluation as the target linear velocity and target angular velocity, respectively, so as to control the dual-arm transfer nursing robot to move at the target linear velocity and the target angular velocity.

[0036] According to the technical solution provided in the embodiments of this application, obtaining several groups of first samples includes at least the following steps:

[0037] Obtain the equipment parameters of the dual-arm transfer nursing robot;

[0038] Based on the device parameters, the sampling space is calculated; the device parameters include at least linear velocity, angular velocity, linear acceleration, angular acceleration, and safety distance.

[0039] The first sample is sampled uniformly within the sampling space at a first sampling step size to obtain several groups of the first sample.

[0040] According to the technical solution provided in the embodiments of this application, calculating the sampling space based on the device parameters includes at least the following steps:

[0041] Based on the maximum and minimum values ​​of the linear velocity and the maximum and minimum values ​​of the angular velocity, a first space is obtained;

[0042] The second space is obtained based on the linear velocity and maximum linear acceleration of the dual-arm transfer nursing robot at the current moment, as well as the angular velocity and maximum angular acceleration of the dual-arm transfer nursing robot at the current moment;

[0043] The third space is obtained based on the minimum and maximum linear acceleration of the linear velocity, and the minimum and maximum angular acceleration of the angular velocity;

[0044] The intersection of the first space, the second space, and the third space is taken as the sampling space.

[0045] According to the technical solution provided in the embodiments of this application, the calculation of the overall trajectory evaluation of each predicted trajectory includes at least the following steps:

[0046] Calculate the azimuth evaluation, obstacle evaluation, velocity evaluation, and global path evaluation for each of the predicted trajectories;

[0047] Based on the azimuth evaluation, obstacle evaluation, speed evaluation, and global path evaluation of each predicted trajectory, a total trajectory evaluation for each predicted trajectory is obtained.

[0048] Compared with the prior art, the beneficial effects of this application are as follows: This application calculates the total cost of search nodes including safety costs, connects the search nodes with the minimum total cost according to their indices to obtain an initial global path. Since this initial global path is obtained by connecting the search nodes with the minimum total cost, it ensures the safety of personnel when the dual-arm transfer nursing robot moves the person being cared for. Then, redundant nodes in the initial global path are deleted, and the remaining nodes are connected to obtain an optimized global path. Since redundant nodes will increase the length of the movement path, deleting redundant nodes can shorten the movement path. The optimized global path takes into account the principle of shortest path. Therefore, planning local paths with the optimized global path to guide the dual-arm transfer nursing robot in finding its way can achieve faster and safer movement of the person being cared for. Attached Figure Description

[0049] Figure 1 A flowchart illustrating the steps of the indoor path planning method for a dual-arm transfer nursing robot provided in this application embodiment;

[0050] Figure 2 A schematic diagram of the search node provided in an embodiment of this application;

[0051] Figure 3 A schematic diagram of the pre-selected nodes provided in the embodiments of this application;

[0052] Figure 4 A schematic diagram of the initial global path provided in the embodiments of this application;

[0053] Figure 5 A schematic diagram illustrating the removal of redundant known nodes (point B is a redundant known node, and points A and C are important nodes) provided in an embodiment of this application.

[0054] Figure 6 A schematic diagram of the optimized global path provided in the embodiments of this application;

[0055] Figure 7 This is a schematic diagram of the actual path obtained by the robot moving according to the local path guidance provided in the embodiments of this application. Detailed Implementation

[0056] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0057] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0058] As mentioned in the background section, this application proposes an indoor path planning method for a dual-arm transfer nursing robot to address the problems in the existing technology. Please refer to [link / reference]. Figure 1 As shown, it includes the following steps:

[0059] S1. Construct an obstacle map to simulate an indoor environment, mark the start and end points on the obstacle map, create an index for each node on the map, and create a first list and a second list.

[0060] Specifically, an obstacle map is constructed in Python to simulate an indoor environment. An index is created for each node on the map to identify and record node information. A first list is called an open list, and a second list is called a closed list. Both the first and second lists are empty lists used to store node information.

[0061] S2. Based on the current point and the destination, determine several search nodes and calculate the total cost of each search node. The total cost is the sum of the cost of distance from the starting point, the cost of distance from the destination, and the security cost.

[0062] The safety cost can be characterized by the distance between each search node and the obstacle location. Therefore, the greater the distance to the obstacle location, the smaller the safety cost and the safer it is. Conversely, the smaller the distance, the greater the safety cost and the less safe it is. A greater safety cost means that the robot is more likely to bump into the obstacle during its movement, causing secondary injury to personnel. Therefore, it is necessary to find the search node with the minimum total cost, which can ensure a short path and high safety.

[0063] Furthermore, determining several search nodes based on the current point and the destination point includes at least the following steps:

[0064] Calculate the first angle, which is the angle between the line connecting the current point and the endpoint and the horizontal line;

[0065] The angle database is retrieved to obtain the search direction group corresponding to the first angle. The search direction group includes several search directions, and the nodes in each search direction are the search nodes. The angle database includes at least several angle ranges and search direction groups corresponding to each angle range.

[0066] Furthermore, the angle database is obtained empirically. Existing technologies search in all directions to obtain nodes for each direction. Since the movement directions differ, for directions opposite to the movement direction, it's unnecessary to retain nodes in that direction; only nodes facing the movement direction need to be retained. Therefore, this embodiment divides all nodes around the current point into search nodes and delete nodes, retaining only the search nodes and deleting the delete nodes. Please refer to Table 1 and... Figure 2 As shown, Figure 2 The center position is the current point, and the surrounding points are labeled 1, 2...8 to represent all directions. Specifically, in Table 1... The range is defined by angle; the search direction refers to the retained search direction labels, and the non-search direction refers to the deleted non-search direction labels. Nodes along the search direction are the search nodes, such as... Figure 2 As shown, search directions 1, 2, 3, 4, and 8 are retained, meaning that nodes in these search directions are used as search nodes.

[0067] Table 1 Angle Database

[0068]

[0069] Specifically, the initial current point is taken as the starting point. Based on the starting point and the end point, several search nodes of the starting point are determined. The search node with the minimum total cost is stored in the close list. The next current point is found, and the above process is continued to obtain several search nodes of the current point. The search node with the minimum total cost is stored in the close list again until a certain current point is the end point, thus completing the entire search process.

[0070] S3. The search node with the lowest total cost is designated as a known node and stored in the second list. The remaining search nodes, excluding the known nodes, are stored in the first list.

[0071] Specifically, the close list stores the search node with the minimum total cost obtained in each search, that is, the close list stores several known nodes, and the remaining search nodes other than the known nodes obtained in each search are stored in the open list.

[0072] S4. Select the search node with the minimum total cost from the first list as the current point, and repeat S2-S3 until the current point reaches the endpoint;

[0073] Furthermore, this step explains how to find the current point for the next search, that is, to store the other search nodes except the search node with the minimum total cost in the open list after each search. After each search, the open list contains several search nodes, and the search node with the minimum total cost is selected as the current point.

[0074] S5. Connect all the known nodes in the second list according to their indices to obtain the initial global path; the initial global path is as follows: Figure 4 As shown;

[0075] S6. In the initial global path, identify and delete several redundant known nodes, and treat the remaining known nodes as important nodes. Connect all the important nodes according to their indexes to obtain the optimized global path; the optimized global path is as follows: Figure 6 As shown;

[0076] Furthermore, the process of identifying and deleting several redundant nodes in the initial global path includes at least the following steps:

[0077] Determine the location information of obstacles in the obstacle map and obtain the robot radius of the dual-arm transfer nursing robot;

[0078] Specifically, the robot radius R robot =0.5;

[0079] Each n adjacent important nodes in the optimized global path are grouped into a node group;

[0080] Specifically, the number of n can be evenly distributed according to the total number of nodes. Optionally, n is 3, that is, every 3 adjacent important nodes form a node group. The first node group consists of the starting point, the first important node after the starting point, and the second important node after the starting point.

[0081] Calculate the reference distance of the node group. If the reference distance is greater than the robot radius, then the important node located in the middle position of the node group is designated as the redundant node in the node group. The reference distance is the vertical distance between the line connecting the first and last points of the node group and the obstacle.

[0082] Traverse all the node groups to obtain the redundant nodes in each node group, and delete all the redundant nodes.

[0083] Specifically, the reference distance for the first node group is calculated. This reference distance is the perpendicular distance between the line connecting the starting point and the second important node and the obstacle. It is then determined whether this perpendicular distance is greater than the robot's radius, R. robot =0.5, please refer to Figure 5 As shown, if so, delete the first important node, connect the first and last nodes to generate a new path, and form a new group with the next node; otherwise, proceed to the next node to form a new group. Repeat the same steps until all redundant nodes in all node groups are deleted. The remaining nodes are the important nodes. Calculation stops when the important nodes at the endpoint have been calculated and no new nodes can form a new group. The result is the optimized global path.

[0084] S7. Based on the important nodes in the optimized global path, an improved dynamic window algorithm is used to plan a local path to guide the dual-arm transfer nursing robot in finding its way.

[0085] Specifically, the local path moves the robot by treating each important node in the global path as the destination for each movement. The local path is a short segment of the path during the movement. The global path provides important nodes of the global map, which can guide the robot to move along the local path. The local path is used to follow the global path and avoid obstacles.

[0086] Furthermore, based on the known optimized global path, an improved DWA (Dynamic Window) local path planning algorithm is used for path following and dynamic obstacle avoidance. Please refer to [link / reference needed]. Figure 7 As shown, the robot ultimately moves along a path that takes into account safety, comfort, and the shortest path.

[0087] In a preferred embodiment, calculating the total cost for each of the search nodes includes at least the following steps:

[0088] Construct a first computational model. The inputs of the first computational model are the cost of distance from the starting point, the cost of distance from the ending point, and the security cost, and the output is the total cost.

[0089] Specifically, the first computational model in the Astar algorithm is:

[0090] f(n) = g(n) + h(n) + c(n)

[0091] Where: f(n) represents the total cost of searching node n, g(n) represents the cost of searching node n from the starting point, which is usually the length of the known path between the search node and the starting point, h(n) represents the expected cost of searching node n from the destination, also called the heuristic function of the Astar algorithm, which is the Euclidean distance between the two points, and c(n) represents the safety cost, which is used to introduce safety considerations.

[0092] Obtain the cost of the search node from the starting point, the cost of the search node from the ending point, and the distance from the obstacle;

[0093] If the distance between the obstacle and the obstacle is greater than or equal to the first preset distance, the safety cost is replaced by the first value; if the distance between the obstacle and the obstacle is less than the first preset distance, the safety cost is replaced by the second value.

[0094] Specifically, the first preset distance is 1 meter, the first value is 0, and the second value is 1.

[0095] The cost of the search node from the starting point, the cost from the ending point, and the corresponding security cost are input into the first calculation model to obtain the total cost of the search node.

[0096] In a preferred embodiment, the step of planning a local path using an improved dynamic window algorithm based on each of the important nodes in the optimized global path includes at least the following steps:

[0097] At the current point where the dual-arm transfer nursing robot is located, determine the direction of movement toward the dual-arm transfer nursing robot, and the position information of the next important node at the current point; the position information includes at least the presence of obstacles and the absence of obstacles;

[0098] If the location information at the next important node indicates that there are no obstacles, the global path between the current point of the dual-arm transfer nursing robot and the next important node is used as the local path of the corresponding segment.

[0099] In a preferred embodiment, after determining the movement direction toward the dual-arm transfer nursing robot and the position information at the next important node from the current point, the method further includes the following steps:

[0100] If the location information at the next important node indicates the presence of an obstacle, obtain several pre-selected nodes for that important node;

[0101] The first distance between the current point of the dual-arm transfer nursing robot and each of the pre-selected nodes is obtained respectively, and the pre-selected node with the smallest first distance is taken as the updated important node.

[0102] The global path between the current point of the dual-arm transfer nursing robot and the next important node is updated using the updated important nodes to obtain the local path of the corresponding segment.

[0103] If an obstacle is detected as occupying the next critical node, obstacle identification will be performed in the four directions (left, right, up, and down) of that critical node. After all obstacles are found in the "+" direction, pre-selected nodes will be set at the four endpoints. At this time, the distance between the robot's current position and these four pre-selected nodes will be calculated to obtain the first distance of each pre-selected node. The endpoint with the smallest first distance will be selected as the updated critical node, serving as the endpoint of the next local path to guide the robot forward. Please refer to [reference needed]. Figure 3 As shown, before optimization, the robot rushed towards the obstacle, causing the path to fail. After optimization, the robot bypassed the dynamic obstacle and planned a new route to complete the pathfinding task.

[0104] In a preferred embodiment, during the movement of the dual-arm transfer nursing robot along a planned local path, the method further includes:

[0105] Acquire several sets of first samples, where the first sample is the linear velocity and angular velocity corresponding to the sampling time;

[0106] Specifically, the first sample is (v, ω), where v represents the linear velocity at the sampling time and ω represents the angular velocity at the sampling time.

[0107] Furthermore, obtaining several groups of first samples includes at least the following steps:

[0108] Obtain the equipment parameters of the dual-arm transfer nursing robot;

[0109] Based on the device parameters, the sampling space is calculated; the device parameters include at least linear velocity, angular velocity, linear acceleration, angular acceleration, and safety distance.

[0110] The first sample is sampled uniformly within the sampling space at a first sampling step size to obtain several groups of the first sample.

[0111] Furthermore, calculating the sampling space based on the device parameters includes at least the following steps:

[0112] Based on the maximum and minimum values ​​of the linear velocity and the maximum and minimum values ​​of the angular velocity, a first space is obtained;

[0113] The second space is obtained based on the linear velocity and maximum linear acceleration of the dual-arm transfer nursing robot at the current moment, as well as the angular velocity and maximum angular acceleration of the dual-arm transfer nursing robot at the current moment;

[0114] The third space is obtained based on the minimum and maximum linear acceleration of the linear velocity, and the minimum and maximum angular acceleration of the angular velocity;

[0115] The intersection of the first space, the second space, and the third space is taken as the sampling space.

[0116] Table 2 Equipment Parameters

[0117] <![CDATA[Angular velocity ω / rad·s -1 > 40·π / 180 <![CDATA[Linear acceleration v a / m·s -2 > 0.2 <![CDATA[Angular acceleration w a / rad·s -2 > 40·π / 180 <![CDATA[Safety distance R safe / m]]> 0.2

[0118] Specifically, Table 2 shows the equipment parameters of a dual-arm transfer nursing robot. Substituting the equipment parameters in the table into the following formula, the sampling space V is obtained after calculation. s :

[0119] V m ={(v,ω)|v∈[v min ,v max ],ω∈[ω min ,ω max ]}

[0120]

[0121]

[0122] V s =V m ∩V d ∩V a

[0123] In the formula, V m For the first space, V d For the second space, V a For the third space, v min ,v max ω represents the minimum and maximum linear velocities of the mobile robot. min ,ω max v represents the minimum and maximum angular velocities of the mobile robot. c ,ω c a represents the linear velocity and angular velocity of the mobile robot at the current moment. vmax ,a ωmax This represents the maximum linear acceleration and maximum angular acceleration of the mobile robot, and dist(v,ω) represents the closest distance between the predicted trajectory and the obstacle at the current velocity.

[0124] Specifically, in determining the sampling space V s Then, samples are uniformly taken in the space with the first sampling step size, and the sampling formula is:

[0125] n = [(v high -v low ) / Ε v ]·[(ω high -ω low ) / Ε ω ]

[0126] Where: v high ,v low ω represents the upper and lower limits of the linear velocity space. high ,ω low E represents the upper and lower limits of the angular velocity space. v E ω Indicates the sampling interval. Optionally, E v =0.1,E ω =π / 180.

[0127] Based on each of the first samples, the predicted trajectory corresponding to each of the first samples is obtained;

[0128] Calculate the total trajectory evaluation of each predicted trajectory, and take the linear velocity and angular velocity of the predicted trajectory corresponding to the largest total trajectory evaluation as the target linear velocity and target angular velocity, respectively, so as to control the dual-arm transfer nursing robot to move at the target linear velocity and the target angular velocity.

[0129] After sampling several sets of the first sample (v, ω), trajectory prediction is performed using the kinematic model of the mobile robot to obtain the predicted trajectory. The kinematic model is as follows:

[0130] a n =v·ω≤a n0

[0131] x k =x k-1 +v·cos(θ k-1 )Δt

[0132] y k =y k+1 +v·sin(θ k-1 )Δt

[0133] θ k =θ k-1 +ωΔt

[0134] Where: a n a represents the centripetal acceleration of the robot. n0 Let (x, y, θ) represent the centripetal acceleration threshold, (x, y, θ) represent the robot's pose, k represent the sampling time, and Δt represent the sampling interval. Optionally, a n0 =0.25, Δt =0.1.

[0135] After obtaining the predicted trajectory based on the kinematic model, it is put into the evaluation function for overall trajectory evaluation. The first sample (v,ω) corresponding to the highest evaluated trajectory is recorded and used as the robot's current speed for movement.

[0136] In a preferred embodiment, calculating the overall trajectory evaluation of each of the predicted trajectories includes at least the following steps:

[0137] Calculate the azimuth evaluation, obstacle evaluation, velocity evaluation, and global path evaluation for each of the predicted trajectories;

[0138] Based on the azimuth evaluation, obstacle evaluation, speed evaluation, and global path evaluation of each predicted trajectory, a total trajectory evaluation for each predicted trajectory is obtained.

[0139] Specifically, the evaluation function is as follows:

[0140] G(v,ω)=σ[α·h(v,ω)+β·d(v,ω)+γ·v(v,ω)+δ·A(v,ω)]

[0141] In the formula: G(v,ω) represents the overall trajectory evaluation, generally a higher overall trajectory evaluation indicates a better trajectory; h(v,w) represents the azimuth evaluation, used to assess the error Δθ of the angle between the trajectory endpoint position direction and the line connecting the target point at the current sampling speed; d(v,w) represents the obstacle evaluation, usually the closest distance between the simulated trajectory and the obstacle at the current speed. Here, the threshold in d(v,w) is set to the sum of the robot radius and the safety distance, i.e., R = R robot +R safe =0.5+0.2=0.7, and set the value to 10 when there are no obstacles or the nearest distance is greater than the threshold; if there are no obstacles or the nearest distance is greater than the set threshold, then set its value to a large constant value. v(v,w) represents the speed evaluation, which is generally the current speed. It can be directly represented by the current linear velocity. The larger its value, the faster the speed on the planned trajectory and the higher the evaluation score. A(v,ω) represents the global path evaluation. The higher the fit with the global path, the higher the evaluation. Set A(v,ω) to the reciprocal of the distance between the current point and the important node in the global path that is the end point of this local path segment. If a global path node is detected to be occupied by a dynamic obstacle, then the global path node is searched again according to the improved pathfinding principle. α,β,γ,δ represent the coefficients of the evaluation function, set to α=0.15, β=1.0, γ=1.0, δ=2.0, and σ represents normalization processing.

[0142] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are only preferred embodiments of this application. It should be noted that due to the limitations of written expression, while there are objectively infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this invention, and can also combine the above technical features in an appropriate manner. These improvements, modifications, changes, or combinations, or the direct application of the inventive concept and technical solution to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. An indoor path planning method for a dual-arm transfer nursing robot, characterized in that, Includes the following steps: S1. Construct an obstacle map to simulate an indoor environment, mark the start and end points on the obstacle map, create an index for each node on the map, and create a first list and a second list. S2. Based on the current point and the destination, determine several search nodes, including at least the following steps: Calculate the first angle, which is the angle between the line connecting the current point and the endpoint and the horizontal line; The angle database is retrieved to obtain the search direction group corresponding to the first angle. The search direction group includes several search directions, and the nodes in each search direction are the search nodes. The angle database includes at least several angle ranges and search direction groups corresponding to each angle range. And calculate the total cost of each of the search nodes, the total cost being the sum of the cost of distance from the starting point, the cost of distance from the ending point, and the security cost, including at least the following steps: Construct a first computational model. The inputs of the first computational model are the cost of distance to the starting point, the cost of distance to the ending point, and the security cost. The output is the total cost. Obtain the cost of the search node from the starting point, the cost of the search node from the ending point, and the distance from the obstacle; If the distance between the obstacle and the obstacle is greater than or equal to the first preset distance, the safety cost is replaced by the first value; if the distance between the obstacle and the obstacle is less than the first preset distance, the safety cost is replaced by the second value. The cost of the search node from the starting point, the cost from the ending point, and the corresponding security cost are input into the first calculation model to obtain the total cost of the search node. S3. The search node with the lowest total cost is designated as a known node and stored in the second list. The remaining search nodes, excluding the known nodes, are stored in the first list. S4. Select the search node with the minimum total cost from the first list as the current point, and repeat S2-S3 until the current point reaches the endpoint; S5. Connect all the known nodes in the second list according to their indices to obtain the initial global path; S6. In the initial global path, identify and delete several redundant known nodes, and take the remaining known nodes as important nodes. Connect all the important nodes according to the index to obtain the optimized global path. S7. Based on the important nodes in the optimized global path, a local path is planned using an improved dynamic window algorithm to guide the dual-arm transfer nursing robot in pathfinding, including at least the following steps: At the current point where the dual-arm transfer nursing robot is located, determine the direction of movement toward the dual-arm transfer nursing robot, and the position information of the next important node at the current point; the position information includes at least the presence of obstacles and the absence of obstacles; If the location information at the next important node indicates the presence of an obstacle, obtain several pre-selected nodes for that important node; The first distance between the current point of the dual-arm transfer nursing robot and each of the pre-selected nodes is obtained respectively, and the pre-selected node with the smallest first distance is taken as the updated important node. The global path between the current point of the dual-arm transfer nursing robot and the next important node is updated using the updated important nodes to obtain the local path of the corresponding segment; If the location information at the next important node indicates that there are no obstacles, the global path between the current point of the dual-arm transfer nursing robot and the next important node is used as the local path of the corresponding segment.

2. The indoor path planning method for the dual-arm transfer nursing robot according to claim 1, characterized in that: The process of identifying and deleting redundant nodes in the initial global path includes at least the following steps: Determine the location information of obstacles in the obstacle map and obtain the robot radius of the dual-arm transfer nursing robot; Each n adjacent important nodes in the optimized global path are grouped into a node group; Calculate the reference distance of the node group. If the reference distance is greater than the robot radius, then the important node located in the middle position of the node group is designated as the redundant node in the node group. The reference distance is the vertical distance between the line connecting the first and last points of the node group and the obstacle. Traverse all the node groups to obtain the redundant nodes in each node group, and delete all the redundant nodes.

3. The indoor path planning method for a dual-arm transfer nursing robot according to claim 1, characterized in that: During the movement of the dual-arm transfer nursing robot along a planned local path, the method further includes: Acquire several sets of first samples, where the first sample is the linear velocity and angular velocity corresponding to the sampling time; Based on each of the first samples, the predicted trajectory corresponding to each of the first samples is obtained; Calculate the total trajectory evaluation of each predicted trajectory, and take the linear velocity and angular velocity of the predicted trajectory corresponding to the largest total trajectory evaluation as the target linear velocity and target angular velocity, respectively, so as to control the dual-arm transfer nursing robot to move at the target linear velocity and the target angular velocity.

4. The indoor path planning method for a dual-arm transfer nursing robot according to claim 3, characterized in that: The process of obtaining several groups of first samples includes at least the following steps: Obtain the equipment parameters of the dual-arm transfer nursing robot; Based on the device parameters, the sampling space is calculated; the device parameters include at least linear velocity, angular velocity, linear acceleration, angular acceleration, and safety distance. The first sample is sampled uniformly within the sampling space at a first sampling step size to obtain several groups of the first sample.

5. The indoor path planning method for a dual-arm transfer nursing robot according to claim 4, characterized in that: The calculation of the sampling space based on the device parameters includes at least the following steps: Based on the maximum and minimum values ​​of the linear velocity and the maximum and minimum values ​​of the angular velocity, a first space is obtained; The second space is obtained based on the linear velocity and maximum linear acceleration of the dual-arm transfer nursing robot at the current moment, as well as the angular velocity and maximum angular acceleration of the dual-arm transfer nursing robot at the current moment; Based on the minimum and maximum linear accelerations of the linear velocity, and the minimum and maximum angular accelerations of the angular velocity, the third space is obtained; The intersection of the first space, the second space, and the third space is taken as the sampling space.

6. The indoor path planning method for a dual-arm transfer nursing robot according to claim 3, characterized in that: The calculation of the overall trajectory evaluation for each of the predicted trajectories includes at least the following steps: Calculate the azimuth evaluation, obstacle evaluation, velocity evaluation, and global path evaluation for each of the predicted trajectories; Based on the azimuth evaluation, obstacle evaluation, speed evaluation, and global path evaluation of each predicted trajectory, a total trajectory evaluation for each predicted trajectory is obtained.

Images