A robot navigation and comfortable following interaction method, system, device and medium

Abstract

The present invention relates to the field of service robots and artificial intelligence, and in particular to a method, system, device and medium for robot navigation and comfortable following interaction, the method comprising: constructing a priori knowledge base of a service robot; receiving natural language instructions from a user for processing, matching the processed natural language instructions with the knowledge base for keywords to extract coordinate information of multiple target landmarks and target person information; after identifying the target person, following the target person under the planned optimal path according to the initial following distance, optimizing the initial following distance in real time, adjusting the following distance of the target person according to the optimized following distance, and updating the optimized following distance as the initial distance for the next following to the target person knowledge base. The present invention can realize multi-coordinate navigation and intelligent following tasks, and update the following distance with the help of the priori knowledge base and the user's feedback information to provide a comfortable and personalized human-computer interaction experience.

Description

Technical Field

[0001] The present invention relates to the fields of service robots and artificial intelligence, and in particular to a method, system, device and medium for robot navigation and comfortable following interaction. Background Art

[0002] In the past few decades, service robot technology has been widely used in daily life and work, providing people with more convenient and intelligent services. As an important part of human-computer interaction, navigation and following tasks are crucial to the function and practicality of mobile service robots.

[0003] In order to better meet user needs, researchers have been working to improve navigation and following algorithms so that robots can navigate more intelligently according to user instructions and achieve a comfortable and natural human-computer interaction experience. As an autonomous mobile platform with artificial intelligence, service robots have been widely used in fields such as home, medical care, catering, and public services. They can not only perform simple tasks, such as navigating to a designated location or delivering items, but also interact with users in natural language and provide corresponding services according to user needs. With the advancement of technology and the development of intelligent algorithms, service robots are gradually becoming an indispensable part of people's daily lives, bringing users a more convenient and efficient service experience.

[0004] However, although service robots have made some progress in navigation and following tasks, they still face some challenges and demands. First, multi-coordinate navigation is one of the important functions of service robots. In home and hospital scenarios, it involves multiple landmarks, such as bedrooms, living rooms, and kitchens. Service robots need to perform navigation tasks in these landmark locations according to user instructions. Secondly, in service robots combined with navigation, when the service robot arrives at the designated location, it is usually necessary for humans to initiate interaction. Users need to approach the service robot and conduct human-computer interaction through dialogue or control panels. In addition, traditional following methods are usually based on fixed rules or interaction modes, which are difficult to meet the personalized needs of different users. Users may have different preferences and comfort requirements for the social distance of robots, and a more intelligent method is needed to achieve a personalized following experience. Summary of the invention

[0005] In order to solve the above technical problems, the present invention also provides a robot navigation and comfortable following interaction method, the method comprising the following steps:

[0006] S1: Based on the prior knowledge required by the service robot, a knowledge base is constructed, wherein the knowledge base includes a landmark point knowledge base and a target person knowledge base;

[0007] S2: receiving a natural language instruction from a user for processing, and matching the processed natural language instruction with the knowledge base for keywords to extract coordinate information of a plurality of target landmark points and information of a target person;

[0008] S3: planning an optimal path including multiple target landmark points according to the extracted coordinate information of multiple target landmark points and target person information;

[0009] S4: after identifying the target person, calculating an initial following distance, following the target person on the optimal path according to the initial following distance, optimizing the initial following distance in real time, adjusting the following distance for the target person according to the optimized following distance, and updating the optimized following distance as the initial distance for the next following into the target person knowledge base;

[0010] The step of obtaining the optimized following distance includes:

[0011] Obtaining feedback information of the target person, wherein the feedback information includes semantic information, expression information, and identity information;

[0012] Based on the feedback information, construct a pedestrian following prediction model;

[0013] The minimum following distance in the following task is obtained through the pedestrian following prediction model, and the minimum following distance is the optimized following distance.

[0014] In one embodiment of the present invention, the method for constructing the landmark point knowledge base includes: storing the names and coordinates of the landmark points that the service robot needs to reach in the annotated prior map in the landmark point knowledge base, wherein each landmark point name corresponds to a coordinate.

[0015] In one embodiment of the present invention, the target person knowledge base includes basic information and social information of the target person, the basic information includes gender, age, identity and facial information, and the social information includes facial expression information shown by the target person at different times and the following distance between the service robot and the target person.

[0016] In one embodiment of the present invention, a method for extracting coordinate information of a plurality of target landmark points and target person information includes:

[0017] Matching the natural language instruction with the landmark points stored in the landmark point knowledge base, extracting multiple target landmark points in the natural language instruction from the landmark point knowledge base, and obtaining the coordinate points of the prior map corresponding to each target landmark point in the landmark point knowledge base;

[0018] The natural language instruction is matched with the target person stored in the target person knowledge base, and the identity information and face information of the target person in the natural language instruction are extracted from the target person knowledge base.

[0019] In one embodiment of the present invention, the calculation method of the initial following distance is: after identifying the target person, the depth camera is used to obtain the depth distances of the four vertices and the center point of the minimum circumscribed detection rectangle of the target person in the pixel image, and the average value of the depth distances of the four vertices and the center point is used as the initial following distance distance:

[0020]

[0021] Among them, d i Indicates the depth distance, i = 1, 2, 3, 4, 5 correspond to the four vertices and the center point respectively, indicator (d i ) represents the indicator function, the depth distance d i When within the preset range, indicator(d i ) is 1, the depth distance d i When it is not within the preset range, the indicator (d i ) has a value of 0.

[0022] In one embodiment of the present invention, the method for following the optimal path according to the initial following distance includes: calculating the linear velocity and angular velocity of the robot according to the initial following distance to ensure the consistency of the initial following distance; the linear velocity is calculated as follows:

[0023] linear_speed = distance * k s +h s ,

[0024] k s =(L max -L min ) / (D max -D min ),

[0025] h s =L min -k s *D min .

[0026] The calculation formula of the angular velocity is:

[0027] rotation_speed=-k 1 *center_x+h i .

[0028] Among them, L max With L min Respectively represent the maximum and minimum values ​​of the robot's linear velocity, D max With D min Respectively represent the maximum and minimum values ​​of the robot’s following distance, k 1 represents the rotation coefficient, h i Indicates the offset for left or right rotation.

[0029] In one embodiment of the present invention, the step of constructing a pedestrian following prediction model includes: through multiple interactions with the service robot, collecting the following distances at different times, and using this as training data, selecting a feedforward neural network as the network architecture of the pedestrian following prediction model, and obtaining the pedestrian following prediction model after multiple iterative training.

[0030] Based on the same inventive concept as the method, the present invention also provides a robot navigation and comfortable following interaction system, comprising:

[0031] A knowledge base construction module is used to construct a knowledge base based on the prior knowledge required by the service robot, wherein the knowledge base includes a landmark point knowledge base and a target person knowledge base;

[0032] A landmark point and target person matching module, used for receiving and processing natural language instructions from a user, and matching the processed natural language instructions with the knowledge base for keywords to extract coordinate information of multiple target landmark points and target person information;

[0033] A navigation path planning module is used to plan an optimal path including multiple target landmark points according to the extracted coordinate information of multiple target landmark points and target person information;

[0034] The following distance optimization module is used to calculate the initial following distance after identifying the target person, follow the optimal path according to the initial following distance, optimize the initial following distance in real time, adjust the following distance of the target person according to the optimized following distance, and update the optimized following distance as the initial distance for the next following into the target person knowledge base.

[0035] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements instructions of the robot navigation and comfortable following interaction method when executing the program.

[0036] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and the computer program, when executed by a processor, implements the robot navigation and comfortable following interaction method.

[0037] The above technical solution of the present invention has the following advantages compared with the prior art:

[0038] The present invention can handle and complete more complex navigation and following tasks by processing natural language instructions, and by building a feedforward neural network and establishing a knowledge base, it can use user feedback information to update the following distance in real time, realize personalized following interaction, and meet the user's comfortable interaction needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to make the content of the present invention more clearly understood, the present invention is further described in detail below according to specific embodiments of the present invention in conjunction with the accompanying drawings, wherein

[0040] Figure 1 is a flow chart of the robot navigation and comfortable following interaction method of the present invention;

[0041] Figure 2 is a specific flow chart of the robot navigation and comfortable following interaction method described in Embodiment 1 of the present invention;

[0042] Figure 3 is a flowchart of a landmark point and a follow-up target in the first embodiment of the present invention;

[0043] Figure 4 is a flowchart of pedestrian following in Embodiment 1 of the present invention;

[0044] Figure 5 is a schematic diagram of a feedforward neural network in Embodiment 1 of the present invention;

[0045] Figure 6 is a simulation environment structure diagram of scene 1 in embodiment 1 of the present invention;

[0046] Figure 7 is a simulation environment structure diagram of scene 2 in embodiment 1 of the present invention;

[0047] Figure 8 It is a simulation environment structure diagram of scene three in embodiment one of the present invention. DETAILED DESCRIPTION

[0048] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0049] Embodiment 1

[0050] Reference Figures 1-2 As shown, the present invention provides a robot navigation and comfortable following interaction method, the method comprising the following steps:

[0051] S1: Based on the prior knowledge required by the service robot, a knowledge base is constructed, wherein the knowledge base includes a landmark point knowledge base and a target person knowledge base;

[0052] S2: receiving a natural language instruction from a user for processing, and matching the processed natural language instruction with the knowledge base for keywords to extract coordinate information of a plurality of target landmark points and information of a target person;

[0053] S3: planning an optimal path including multiple target landmark points according to the extracted coordinate information of multiple target landmark points and target person information;

[0054] S4: After identifying the target person, the initial following distance is calculated, and the target person is followed on the optimal path according to the initial following distance, and the initial following distance is optimized in real time, and the following distance of the target person is adjusted according to the optimized following distance, and the optimized following distance is updated as the initial distance for the next following in the target person knowledge base.

[0055] From the above technology, the present invention uses natural language processing technology to realize the multi-coordinate navigation function of the robot, and utilizes a following algorithm based on a neural network to obtain the semantic information, identity information and expression information of the target person to predict the interaction distance most suitable for the user, aiming to realize multi-coordinate navigation and intelligent following tasks, and further utilize the user's feedback information with the help of the existing knowledge base to update the following distance, so as to provide a more comfortable and personalized human-computer interaction experience.

[0056] The method for constructing the landmark point knowledge base includes: storing the names and coordinates of the landmark points that the service robot needs to reach in the annotated prior map into the landmark point knowledge base, where each landmark point name corresponds to a coordinate. As shown in Table 1:

[0057] Table 1

[0058] Scenario Landmarks coordinate Landmarks coordinate Landmarks coordinate Hospital pharmacy (15,5) Window 1 (5,8) Elevator 1 (3,11) ...... family Master Bedroom (16,9) Second Bedroom (10,5) kitchen (1,2) ...... Office Pantry (20,5) hall (8,2) Office Area (8,6) ......

[0059] The target person knowledge base includes the target person's basic information and social information. The basic information includes gender, age, identity and face information. The social information includes the target person's facial expression information at different times and the following distance between the service robot and the target person. As shown in Table 2:

[0060] Table 2

[0061]

[0062] like Figure 3 As shown, the method for extracting coordinate information of multiple target landmark points and target person information includes:

[0063] Matching the natural language instruction with the landmark points stored in the landmark point knowledge base, extracting multiple target landmark points in the natural language instruction from the landmark point knowledge base, and obtaining the coordinate points of the prior map corresponding to each target landmark point in the landmark point knowledge base;

[0064] The natural language instruction is matched with the target person stored in the target person knowledge base, and the identity information and face information of the target person in the natural language instruction are extracted from the target person knowledge base.

[0065] like Figure 4 As shown, the calculation method of the initial following distance is: after the target person is identified by the yolov5 model, the depth camera is used to obtain the depth distances of the four vertices and the center point of the minimum circumscribed detection rectangle of the target person in the pixel image. By calculating the average of the depth distances of the five points, the noise of the depth information can be reduced and the robustness of the distance estimation can be improved. The average of the depth distances of the four vertices and the center point is used as the initial following distance distance:

[0066]

[0067] Among them, d i represents the depth distance, i=1, 2, 3, 4, 5 correspond to the four vertices and the center point respectively, indicator(x) represents the indicator function, and the depth distance d i When within the preset range (0.4m, 10m), the indicator (d i ) is 1, the depth distance d i When the value is less than 0.4 or greater than 10, the value of indicator(x) is 0.

[0068] The target person is matched with the face information in the target person knowledge base, and after the target person is identified, the method of following the robot under the optimal path according to the initial following distance obtained by the above calculation formula includes: calculating the linear velocity and angular velocity of the robot by the initial following distance, so as to ensure the consistency of the initial following distance; the linear velocity is calculated as follows:

[0069] linear_speed = distance * k s +h s ,

[0070] k s =(L max -L min ) / (D max -D min ),

[0071] hs =L min -k s *D min .

[0072] The calculation formula of the angular velocity is:

[0073] rotation_speed=-k 1 *center_x+h i .

[0074] Among them, L max With L min Respectively represent the maximum and minimum values ​​of the robot’s linear velocity; D max With D min Respectively represent the maximum and minimum values ​​of the robot’s following distance, D min The value of is predicted by the feedforward neural network, D max The value of is set artificially based on experience; k 1 represents the rotation coefficient; h i Indicates the offset for left or right rotation.

[0075] The step of obtaining the optimized following distance includes:

[0076] S41: Obtain feedback information of the target person by interacting with the service robot multiple times, wherein the feedback information includes semantic information, expression information and identity information; the method for obtaining the feedback information of the target person is: obtaining semantic information by having the user send natural language instructions to the service robot, obtaining the user's expression information by using the MTCNN network and the FaceNet network, and obtaining the identity information by querying the target person knowledge base.

[0077] S42: Based on the feedback information, construct a pedestrian following prediction model; the step of constructing the pedestrian following prediction model includes: by interacting with the service robot multiple times, collecting the following distances at different times, and using this as training data, selecting a feedforward neural network as the network architecture of the pedestrian following prediction model, such as Figure 5 As shown, a feedforward neural network is selected as the network architecture of the pedestrian following prediction model. The input layer of the pedestrian following prediction model contains 128 neurons, uses the ReLU activation function, accepts a three-dimensional input feature vector (including identity information, semantic information and expression information), the hidden layer contains 128 neurons, uses the ReLU activation function, the output layer contains one neuron, and uses the mean square error as the loss function. After 200 iterations of training, the pedestrian following prediction model is obtained.

[0078] S43: Obtaining a minimum following distance in a following task through the pedestrian following prediction model, wherein the minimum following distance is an optimized following distance.

[0079] To fully demonstrate the effectiveness of the present invention, the present invention was tested in three simulation scenarios: home, hospital and office. Figures 6 to 8 shown.

[0080] For each experimental environment, a total of 20 experiments were conducted. The experimental indicators include navigation success rate, following success rate, combined navigation and following success rate, and task duration. The experimental data are shown in Table 3. In the three scenarios, the method proposed in the present invention achieves a high success rate in multi-waypoint navigation and maintains the stable completion of the pedestrian following task.

[0081] Table 3

[0082]

[0083]

[0084] In this experiment, 10 volunteers who have never interacted with the Xiaopang robot were recruited as research subjects. The questionnaire shown in Table 4 was used to collect user feedback and the experience of each follower interaction. The effectiveness of using a feedforward neural network (denoted by Ours in the table) to predict the comfort distance of following and using a nonlinear equation (denoted by NE in the table) to predict the distance was compared by scoring. The experimental data are shown in Table 5. The evaluation indicators of the model include M, SD, F, P and η2, where M is the mean; SD is the standard deviation; the F value is used to compare the difference between the sum of squares between groups (SSB) and the sum of squares within groups (SSW). The larger the F value, the more significant the difference between groups is relative to the difference within groups; the P value indicates the probability of the observed F value or more extreme cases, compared with the pre-set significance level (for example: 0.05). If P<0.05, the difference between groups can be considered significant; η2 indicates the degree to which the factor explains the variation of the variable. The value range of η2 is between 0 and 1. The closer the value of η2 is to 1, the higher the degree to which the factor explains the variable.

[0085] By ANOVA, for question Q1 (Table 5, row 2), there was a statistically significant difference between the feedforward neural network model and the nonlinear equation model (F(1, 9) = 15, p < 0.05, η2 = 0.455). Therefore, these results indicate that the performance of the neural network prediction based on feedback information is better than the prediction using the nonlinear equation in terms of comfortable following distance. For question Q2 (Table 5, row 3), the difference between the two models was statistically significant (F(1, 9) = 5.512, p < 0.05, η2 = 0.234). Therefore, these results mean that the proposed method is more effective in following safety. For question Q3 (Table 5, row 4), there was a significant difference between the two models under the pedestrian following task (F(1, 9) = 5.312, p < 0.05, η2 = 0.228). Most participants believed that the feedforward neural network prediction method catered to the user's following preference more effectively than the nonlinear equation.

[0086] Table 4

[0087]

[0088] Table 5

[0089]

[0090] Embodiment 2

[0091] Based on the same inventive concept as the method described in Example 1, the present invention also provides a robot navigation and comfortable following interaction system, comprising:

[0092] A knowledge base construction module is used to construct a knowledge base based on the prior knowledge required by the service robot, wherein the knowledge base includes a landmark point knowledge base and a target person knowledge base;

[0093] A landmark point and target person matching module, used for receiving and processing natural language instructions from a user, and matching the processed natural language instructions with the knowledge base for keywords to extract coordinate information of multiple target landmark points and target person information;

[0094] A navigation path planning module is used to plan an optimal path including multiple target landmark points according to the extracted coordinate information of multiple target landmark points and target person information;

[0095] The following distance optimization module is used to calculate the initial following distance after identifying the target person, follow the optimal path according to the initial following distance, optimize the initial following distance in real time, adjust the following distance of the target person according to the optimized following distance, and update the optimized following distance as the initial distance for the next following into the target person knowledge base.

[0096] Embodiment 3

[0097] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the instructions of the robot navigation and comfortable following interaction method described in Example 1 when executing the program.

[0098] Embodiment 4

[0099] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the robot navigation and comfortable following interaction method described in Embodiment 1 is implemented.

[0100] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0101] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0102] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0103] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0104] Obviously, the above embodiments are merely examples for clear explanation and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from these are still within the protection scope of the invention.

Claims

1. A robot navigation and comfortable following interaction method, It is characterized in that The following steps are involved: S1: Based on the prior knowledge required by the service robot, a knowledge base is constructed, wherein the knowledge base includes a landmark point knowledge base and a target person knowledge base; S2: receiving a natural language instruction from a user for processing, and matching the processed natural language instruction with the knowledge base for keywords to extract coordinate information of a plurality of target landmark points and information of a target person; S3: planning an optimal path including multiple target landmark points according to the extracted coordinate information of multiple target landmark points and target person information; S4: After the target person is identified, the initial following distance is calculated, including: The calculation method of the initial following distance is: after identifying the target person, the depth camera is used to obtain the depth distances of the four vertices and the center point of the minimum circumscribed detection rectangle of the target person in the pixel image, and the average value of the depth distances of the four vertices and the center point is used as the initial following distance distance: Among them, d i Indicates the depth distance, i = 1, 2, 3, 4, 5 correspond to the four vertices and the center point respectively, indicator (d i ) represents the indicator function, the depth distance d i When within the preset range, the indicator (d i ) is 1, the depth distance d i When it is not within the preset range, indicator(d i ) has a value of 0; S5: following the target person on the optimal path according to the initial following distance, optimizing the initial following distance in real time, adjusting the following distance for the target person according to the optimized following distance, and updating the optimized following distance as the initial distance for next following into the target person knowledge base; The step of obtaining the optimized following distance includes: Obtaining feedback information of the target person, wherein the feedback information includes semantic information, expression information, and identity information; Based on the feedback information, construct a pedestrian following prediction model; The minimum following distance in the following task is obtained through the pedestrian following prediction model, and the minimum following distance is the optimized following distance.

2. The robot navigation and comfortable following interaction method according to claim 1, Features: The method for constructing the landmark point knowledge base includes: storing the names and coordinates of the landmark points that the service robot needs to reach in the annotated prior map in the landmark point knowledge base, wherein each landmark point name corresponds to a coordinate.

3. The robot navigation and comfortable following interaction method according to claim 1, Features: The target person knowledge base includes basic information and social information of the target person, wherein the basic information includes gender, age, identity and face information, and the social information includes facial expression information displayed by the target person at different times and the following distance between the service robot and the target person.

4. The robot navigation and comfortable following interaction method according to claim 1, Features: The method for extracting coordinate information of multiple target landmark points and target person information includes: Matching the natural language instruction with the landmark points stored in the landmark point knowledge base, extracting multiple target landmark points in the natural language instruction from the landmark point knowledge base, and obtaining the coordinate points of the prior map corresponding to each target landmark point in the landmark point knowledge base; The natural language instruction is matched with the target person stored in the target person knowledge base, and the identity information and face information of the target person in the natural language instruction are extracted from the target person knowledge base.

5. The robot navigation and comfortable following interaction method according to claim 1, Features: The method for following the optimal path according to the initial following distance includes: calculating the linear velocity and angular velocity of the robot according to the initial following distance, so as to ensure the consistency of the initial following distance; the linear velocity is calculated as follows: linear_speed=distance*k s +h s , k s =(L max -L min ) / (D max -D min ), h s =L min -k s *D min； The calculation formula of the angular velocity is: rotation_speed=-k 1 *center_x+h i； Among them, L max With L min Respectively represent the maximum and minimum values ​​of the robot's linear velocity, D max With D min Respectively represent the maximum and minimum values ​​of the robot’s following distance, k 1 represents the rotation coefficient, h i Indicates the offset for left or right rotation.

6. The robot navigation and comfortable following interaction method according to claim 1, Features: The steps of constructing a pedestrian following prediction model include: through multiple interactions with the service robot, collecting the following distances at different times and using them as training data, selecting a feedforward neural network as the network architecture of the pedestrian following prediction model, and obtaining the pedestrian following prediction model after multiple iterative training.

7. A robot navigation and comfortable following interactive system, It is characterized in that For implementing the robot navigation and comfortable following interaction method according to any one of claims 1 to 6, the system comprises: A knowledge base construction module is used to construct a knowledge base based on the prior knowledge required by the service robot, wherein the knowledge base includes a landmark point knowledge base and a target person knowledge base; A landmark point and target person matching module, used for receiving and processing natural language instructions from a user, and matching the processed natural language instructions with the knowledge base for keywords to extract coordinate information of multiple target landmark points and target person information; A navigation path planning module is used to plan an optimal path including multiple target landmark points according to the extracted coordinate information of multiple target landmark points and target person information; The following distance optimization module is used to calculate the initial following distance after identifying the target person, follow the optimal path according to the initial following distance, optimize the initial following distance in real time, adjust the following distance of the target person according to the optimized following distance, and update the optimized following distance as the initial distance for the next following into the target person knowledge base.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the program, the instructions of the robot navigation and comfortable following interaction method as described in any one of claims 1 to 6 are implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the robot navigation and comfortable following interaction method according to any one of claims 1 to 6 is implemented.

Images