Robot following method, device and mobile robot based on pedestrian re-identification

Abstract

The present invention discloses a robot following method, device and mobile robot based on pedestrian re-identification, including the following steps: collecting real-time images at a preset frequency during the following process; identifying the target pedestrian in the real-time image and obtaining the position information of the target pedestrian; predicting the estimated position of the target pedestrian at a preset time based on the position information and based on a second-order approximation algorithm; calculating the target rotation angle and target moving distance of the mobile robot based on the estimated position; and generating control instructions. The present invention extracts the deep feature information of the target pedestrian, such as body shape, clothing, etc., based on a deep neural network. Then, in a scene where people are passing by, the robot can find the most likely target pedestrian from the crowd to prevent the switching of the following target; at the same time, in the case of a target pedestrian being blocked by a large flow of people or the target pedestrian moving too fast, the robot can maintain continuous tracking capability and quickly find the target pedestrian, which greatly enhances the robustness of the product.

Description

Technical Field

[0001] The present invention relates to the field of robots, and in particular to a robot following method, device and mobile robot based on pedestrian re-identification. Background Art

[0002] With the rapid development of the robotics industry, various service robots are emerging one after another, and robots are being used more and more widely in our lives and work. In service industry buildings such as hotels and banks, welcoming robots can assist in guest identity registration, help guests store their backpacks and other items, and follow guests until the end of the service. In actual scenarios, especially in scenarios with large traffic and people walking between each other, robots are prone to misidentifying target guests or missing target guests when they turn. Therefore, it is necessary to improve the existing robot following method to ensure the robot's following effect in various complex environments. Summary of the invention

[0003] The present invention provides a robot following method, device and mobile robot based on pedestrian re-identification, which solves the technical problem in the prior art that the robot easily loses the target guest.

[0004] The technical solution of the present invention to solve the above technical problem is as follows: A robot following method based on pedestrian re-identification comprises the following steps:

[0005] During the following process, real-time images are collected at a preset frequency;

[0006] Identify the target pedestrian in the real-time image, and obtain the location information of the target pedestrian at the first historical moment, the second historical moment, and the current moment respectively;

[0007] Predicting the estimated position of the target pedestrian at a preset time according to the position information and based on a second-order approximation algorithm;

[0008] Calculating a target rotation angle and a target moving distance of the mobile robot according to the estimated position of the target pedestrian;

[0009] A control instruction including the target rotation angle and the target moving distance is generated to drive the mobile robot to follow the target pedestrian.

[0010] In a preferred embodiment, the identifying of the target pedestrian in the real-time image and obtaining the position information of the target pedestrian at the first historical moment, the second historical moment and the current moment respectively specifically includes:

[0011] Collecting a target pedestrian image, and extracting a target feature vector of the target pedestrian from the target pedestrian image;

[0012] Acquire real-time images corresponding to the first historical moment, the second historical moment, and the current moment, identify all pedestrians in the real-time images, and generate location information and corresponding confidence level of each pedestrian;

[0013] Extracting a pedestrian feature vector of each pedestrian in the real-time image, and comparing the target feature vector with the pedestrian feature vector of each pedestrian respectively, to generate a most similar pedestrian sequence;

[0014] The most similar pedestrian with the highest confidence in the most similar pedestrian sequence is taken as the target pedestrian, and the position information of the target pedestrian corresponding to the first historical moment, the second historical moment and the current moment is obtained.

[0015] In a preferred embodiment, the target feature vector updating step is further included, specifically:

[0016] Obtain the pedestrian feature vector of each pedestrian in the real-time image corresponding to the current moment, and compare the current target feature vector with the pedestrian feature vector of each pedestrian respectively to generate the most similar feature vector and the corresponding maximum similarity;

[0017] When the maximum similarity is greater than a preset threshold, the current target feature vector of the target pedestrian is updated using the most similar feature vector and the maximum confidence corresponding to the most similar pedestrian.

[0018] In a preferred embodiment, the specific formula for updating the current target feature vector of the target pedestrian using the most similar feature vector and the maximum confidence corresponding to the most similar pedestrian is:

[0019]

[0020] X update =α*X pre +(1-α)*X obj

[0021]

[0022] Among them, alpha is the adjustment factor, Similarity is the maximum similarity, and X pre is the most similar eigenvector, X obj is the current target feature vector, X update is the updated target feature vector, S pre is the maximum confidence level.

[0023] In a preferred embodiment, the specific algorithm for predicting the estimated position of the target pedestrian at a preset time is as follows:

[0024]

[0025]

[0026] Considering the situation that the detected values ​​have abnormal values, the constraint on (Lx4, Ly4) is as follows:

[0027] Θ=arctan(||L x4 ||,||L y4 ||)<=45°,

[0028] Among them, Lx1, Ly1 are the location information of the target pedestrian at the first historical moment, Lx2, Ly2 are the location information of the target pedestrian at the second historical moment, Lx3, Ly3 are the location information of the target pedestrian at the current moment, and Lx4, Ly4 are the estimated locations of the target pedestrian at the preset moment.

[0029] In a preferred embodiment, the calculation formula for calculating the target rotation angle and target moving distance of the mobile robot according to the estimated position of the target pedestrian is as follows:

[0030] Θ=arctan(L x4 , L y4 )

[0031]

[0032] Where θ represents the target rotation angle of the mobile robot, and Δd represents the target moving distance of the mobile robot.

[0033] In a preferred embodiment, a registration step is further included, and the registration step specifically comprises: obtaining a facial image of a target pedestrian, matching the facial image with registered guest information, and automatically following the target pedestrian after a successful match.

[0034] A second aspect of an embodiment of the present invention provides a robot following device based on pedestrian re-identification, comprising a collection module, a position acquisition module, a position prediction module, a calculation module and a control module.

[0035] The acquisition module is used to acquire real-time images at a preset frequency during the following process;

[0036] The position acquisition module is used to identify the target pedestrian in the real-time image and obtain the position information of the target pedestrian corresponding to the first historical moment, the second historical moment and the current moment respectively;

[0037] The position prediction module is used to predict the estimated position of the target pedestrian at a preset time according to the position information and based on a second-order approximation algorithm;

[0038] The calculation module is used to calculate the target rotation angle and target moving distance of the mobile robot according to the estimated position of the target pedestrian;

[0039] The control module is used to generate a control instruction including the target rotation angle and the target moving distance to drive the mobile robot to follow the target pedestrian.

[0040] A third aspect of an embodiment of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the robot following method based on pedestrian re-identification described above.

[0041] A fourth aspect of the present invention provides a mobile robot, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned robot following method based on pedestrian re-identification when executing the computer program.

[0042] The present invention provides a robot following method, device and mobile robot based on pedestrian re-identification. Pedestrian re-identification is based on a deep neural network to extract deep feature information of target pedestrians, such as body shape, clothing, etc. Then, in a scene with a flow of people, the robot can find the most likely target pedestrian from the flow of people to prevent the switching of the following target. At the same time, in the event of a target pedestrian disappearing temporarily, such as when the target pedestrian is blocked by a large flow of people or when the target pedestrian moves too fast, the robot can maintain continuous tracking capability and quickly find the target pedestrian, thereby greatly enhancing the robustness of the product.

[0043] In order to make the above-mentioned objects, features and advantages of the invention more obvious and easy to understand, the preferred embodiments of the present invention are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0045] Figure 1 is a flow chart of a robot following method based on pedestrian re-identification provided in Example 1;

[0046] Figure 2 is a structural schematic diagram of a robot following device based on pedestrian re-identification provided in Example 2;

[0047] Figure 3This is a schematic diagram of the structure of a mobile robot provided in Example 3. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0049] It should be noted that, if there is no conflict, the various features in the embodiments of the present invention can be combined with each other, all within the scope of protection of the present invention. In addition, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flow chart, in some cases, the steps shown or described can be performed in a different order from the module division in the device or the flow chart. Furthermore, the words "first", "second", "third", etc. used in the present invention do not limit the data and execution order, but only distinguish the same items or similar items with basically the same functions and effects.

[0050] See also Figure 1 , which is a flowchart of a robot following method based on pedestrian re-identification provided by Embodiment 1 of the present invention, and is applied to a mobile robot, such as Figure 1 As shown, the method comprises the following steps:

[0051] S1, collect real-time images at a preset frequency during the following process. When the target pedestrian starts to move, the mobile robot starts the automatic following function and collects real-time images at a preset frequency through the camera during the following process. The real-time images include the target pedestrian in the field of view of the mobile robot and other pedestrian information.

[0052] S2, identifying the target pedestrian in the real-time image, and obtaining the position information of the target pedestrian at the first historical moment, the second historical moment and the current moment. In a preferred embodiment, the steps are specifically as follows:

[0053] S201, collect a target pedestrian image, and extract a target feature vector of the target pedestrian from the target pedestrian image. The preferred embodiment can extract feature information such as clothing, hairstyle, height, body shape, etc. of the target pedestrian according to a deep neural network method, and form a multi-dimensional feature vector.

[0054] S202, obtaining real-time images corresponding to the first historical moment, the second historical moment, and the current moment, identifying all pedestrians in the real-time images, and generating location information and corresponding confidence of each pedestrian. In a preferred embodiment, a target detection method based on a deep neural network can be used to identify all pedestrians in the real-time image, and obtain the location information Locationi(Lxi,Lyi) and corresponding confidence Score_i(S_i) of each pedestrian.

[0055] S203, extracting a pedestrian feature vector of each pedestrian in the real-time image, and comparing the target feature vector with the pedestrian feature vector of each pedestrian respectively, to generate a most similar pedestrian sequence Index;

[0056] S204: The most similar pedestrian with the highest confidence in the most similar pedestrian sequence is taken as a target pedestrian, and position information of the target pedestrian corresponding to the first historical moment, the second historical moment and the current moment is obtained.

[0057] In actual application scenarios, since the guest is close to the robot camera, the image of the guest seen by the robot is incomplete, resulting in errors in the target feature vector extracted in step S201. Therefore, in subsequent pedestrian retrieval, the target feature vector needs to be updated in real time. Specifically, the following steps are included:

[0058] S205, obtaining a pedestrian feature vector of each pedestrian in the real-time image corresponding to the current moment, and comparing the current target feature vector with the pedestrian feature vector of each pedestrian respectively, to generate a most similar feature vector and a corresponding maximum similarity;

[0059] S206, when the maximum similarity is greater than a preset threshold, the current target feature vector of the target pedestrian is updated using the most similar feature vector and the maximum confidence corresponding to the most similar pedestrian. The specific formula is as follows:

[0060]

[0061] X update =α*X pre +(1-α)*X obj

[0062]

[0063] Among them, alpha is the adjustment factor, Similarity is the maximum similarity, and X pre is the most similar eigenvector, X obj is the current target feature vector, X updat . is the updated target feature vector, S preis the maximum confidence. The maximum confidence is positively correlated with the accuracy of the pedestrian detection box. The accuracy of the box detection directly affects the accuracy of the pedestrian features. Therefore, the alpha factor is greatly affected by the maximum confidence. In a preferred embodiment, the preset threshold, i.e., threshold, is usually 0.7, that is, the target feature vector is updated when the maximum similarity is greater than 0.7.

[0064] Usually, when the target pedestrian is walking, the mobile robot captures images through the camera, and there is a 100ms delay in the process of processing and calculating the positioning target. Therefore, in order to enhance the tracking ability of the robot, the present invention adopts a second-order approximation estimation method and predicts the possible future position of the target pedestrian. That is, S3 is executed, and the estimated position of the target pedestrian at the preset time is predicted based on the second-order approximation algorithm according to the position information corresponding to the target pedestrian at the first historical moment, the second historical moment and the current moment. In a preferred embodiment, the specific algorithm is as follows:

[0065]

[0066]

[0067] Considering the situation that the detected values ​​have abnormal values, the constraint on (Lx4, Ly4) is as follows:

[0068] Θ=arctan(||L x4 ||,||L y4 ||)<=45°,

[0069] Among them, Lx1, Ly1 are the location information of the target pedestrian at the first historical moment, Lx2, Ly2 are the location information of the target pedestrian at the second historical moment, Lx3, Ly3 are the location information of the target pedestrian at the current moment, and Lx4, Ly4 are the estimated locations of the target pedestrian at the preset moment.

[0070] Then, S4 is executed to calculate the target rotation angle and target moving distance of the mobile robot according to the estimated position of the target pedestrian. In a preferred embodiment, the calculation formula used is as follows:

[0071] Θ=arctan(L x4 , L y4 )

[0072]

[0073] Where θ represents the target rotation angle of the mobile robot, Δd represents the target moving distance of the mobile robot, and d const Usually 2m is taken.

[0074] Finally, S5 is executed to generate a control instruction including the target rotation angle and the target moving distance to drive the mobile robot to follow the target pedestrian.

[0075] In a more preferred embodiment, the robot following method based on pedestrian re-identification also includes a registration step, which specifically includes: obtaining a facial image of a target pedestrian, matching the facial image with registered guest information, and if the match is successful, automatically following the target pedestrian according to the control instructions of the target pedestrian; if the match fails, re-obtaining the facial image of the target pedestrian and repeating the matching process.

[0076] The present invention provides a robot following method based on pedestrian re-identification. Pedestrian re-identification is based on deep neural networks to extract deep feature information of target pedestrians, such as body shape, clothing, etc. Then, in a scene with a flow of people, the robot can find the most likely target pedestrian from the flow of people to prevent the switching of the following target. At the same time, in the event of a target pedestrian disappearing temporarily, such as when the target pedestrian is blocked by a large flow of people or when the target pedestrian moves too fast, the robot can maintain continuous tracking capability and quickly find the target pedestrian, thereby greatly enhancing the robustness of the product.

[0077] It should be noted that, in each of the above embodiments, there is not necessarily a certain order between the above steps. A person skilled in the art can understand, based on the description of the embodiments of the present invention, that in different embodiments, the above steps may have different execution orders, that is, they may be executed in parallel, may be executed interchangeably, and so on.

[0078] As another aspect of the embodiment of the present invention, the embodiment of the present invention further provides a robot following device based on pedestrian re-identification. The robot following device based on pedestrian re-identification can be a software module, which includes several instructions stored in a memory, and the processor can access the memory and call the instructions for execution to complete the robot following method based on pedestrian re-identification described in the above embodiments.

[0079] In some embodiments, the robot following device based on pedestrian re-identification can also be constructed by hardware devices. For example, the robot following device based on pedestrian re-identification can be constructed by one or more chips, and each chip can work in coordination with each other to complete the robot following method based on pedestrian re-identification described in the above embodiments. For another example, the robot following device based on pedestrian re-identification can also be constructed by various logic devices, such as a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a single-chip microcomputer, an ARM (Acorn RISC Machine) or other programmable logic devices, discrete gates or transistor logic, discrete hardware components, or any combination of these components.

[0080] Figure 2 1 is a schematic diagram of a structure of a robot following device based on pedestrian re-identification provided by Embodiment 2 of the present invention. The robot following device based on pedestrian re-identification includes a collection module 100, a position acquisition module 200, a position prediction module 300, a calculation module 400 and a control module 500.

[0081] The acquisition module 100 is used to acquire real-time images at a preset frequency during the following process;

[0082] The position acquisition module 200 is used to identify the target pedestrian in the real-time image and obtain the position information of the target pedestrian at the first historical moment, the second historical moment and the current moment respectively;

[0083] The position prediction module 300 is used to predict the estimated position of the target pedestrian at a preset time according to the position information and based on a second-order approximation algorithm;

[0084] The calculation module 400 is used to calculate the target rotation angle and target moving distance of the mobile robot according to the estimated position of the target pedestrian;

[0085] The control module 500 is used to generate a control instruction including the target rotation angle and the target moving distance to drive the mobile robot to follow the target pedestrian.

[0086] It should be noted that the above-mentioned robot following device based on pedestrian re-identification can execute the robot following method based on pedestrian re-identification provided in the embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not fully described in the embodiment of the robot following device based on pedestrian re-identification, please refer to the robot following method based on pedestrian re-identification provided in the embodiment of the present invention.

[0087] Figure 3Schematic diagram of the structure of a mobile robot provided by an embodiment of the present invention. Figure 3 As shown, the mobile robot 600 includes one or more processors 61 and a memory 62. Figure 3 A processor 61 is taken as an example.

[0088] The processor 61 and the memory 62 may be connected via a bus or other means. Figure 3 The example of connecting through bus is taken in the following.

[0089] The memory 62 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the robot following method based on pedestrian re-identification in the embodiment of the present invention. The processor 61 executes various functional applications and data processing of the robot following device based on pedestrian re-identification by running the non-volatile software programs, instructions and modules stored in the memory 62, that is, realizes the robot following method based on pedestrian re-identification provided by the above method embodiment and the functions of each module or unit of the above device embodiment.

[0090] The memory 62 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some embodiments, the memory 62 may optionally include a memory remotely arranged relative to the processor 61, and these remote memories may be connected to the processor 61 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0091] The program instructions / modules are stored in the memory 62 , and when executed by the one or more processors 61 , the robot following method based on pedestrian re-identification in any of the above method embodiments is executed.

[0092] An embodiment of the present invention further provides a non-volatile computer storage medium, wherein the computer storage medium stores computer executable instructions, and the computer executable instructions are executed by one or more processors, such as Figure 3 A processor 61 in the method may enable the one or more processors to execute the robot following method based on pedestrian re-identification in any of the above method embodiments.

[0093] An embodiment of the present invention also provides a computer program product, which includes a computer program stored on a non-volatile computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by an electronic device, the electronic device executes any one of the robot following methods based on pedestrian re-identification.

[0094] The above described device or equipment embodiments are merely illustrative, wherein the unit modules described as separate components may or may not be physically separated, and the components displayed as module units may or may not be physical units, that is, they may be located in one place, or may be distributed on multiple network module units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.

[0095] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0096] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may also be combined, the steps may be implemented in any order, and there are many other changes in different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they may still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A robot following method based on pedestrian re-identification, applied to a mobile robot, characterized in that: The following steps are involved: During the following process, real-time images are collected at a preset frequency; Identify the target pedestrian in the real-time image, and obtain the location information of the target pedestrian at the first historical moment, the second historical moment, and the current moment respectively; Predicting the estimated position of the target pedestrian at a preset time according to the position information and based on a second-order approximation algorithm; Calculating a target rotation angle and a target moving distance of the mobile robot according to the estimated position of the target pedestrian; Generate a control instruction including the target rotation angle and the target moving distance to drive the mobile robot to follow the target pedestrian. The specific algorithm for predicting the estimated position of the target pedestrian at a preset time is as follows: Considering the situation that the detected values ​​have abnormal values, the constraint on (Lx4,Ly4) is as follows: Θ=arctan(||Lx4 ||||Ly4||)<45°, Where Lx1, Ly1 is the position information of the target pedestrian at the first historical moment, Lx2, Ly2 is the position information of the target pedestrian at the second historical moment, Lx3, Ly3 is the position information of the target pedestrian at the current moment, Lx4, Ly4 is the estimated position of the target pedestrian at the preset moment, and the calculation formula for calculating the target rotation angle and target moving distance of the mobile robot based on the estimated position of the target pedestrian is as follows: Θ=arctan(Lx4,Ly4) Among them, θ represents the target rotation angle of the mobile robot, Δd represents the target moving distance of the mobile robot, dconst takes a value of 2, and also includes a registration step, which specifically includes: obtaining a facial image of the target pedestrian, matching the facial image with the registered guest information, and automatically following the target pedestrian after the match is successful.

2. The robot following method based on pedestrian re-identification according to claim 1, characterized in that: Identify the target pedestrian in the real-time image and obtain the location information of the target pedestrian at the first historical moment, the second historical moment and the current moment, including: Collecting a target pedestrian image, and extracting a target feature vector of the target pedestrian from the target pedestrian image; Acquire real-time images corresponding to the first historical moment, the second historical moment, and the current moment, identify all pedestrians in the real-time images, and generate location information and corresponding confidence level of each pedestrian; Extracting a pedestrian feature vector of each pedestrian in the real-time image, and comparing the target feature vector with the pedestrian feature vector of each pedestrian respectively, to generate a most similar pedestrian sequence; The most similar pedestrian with the highest confidence in the most similar pedestrian sequence is taken as the target pedestrian, and the position information of the target pedestrian corresponding to the first historical moment, the second historical moment and the current moment is obtained.

3. The robot following method based on pedestrian re-identification according to claim 2, characterized in that: It also includes the target feature vector update step, specifically: Obtain the pedestrian feature vector of each pedestrian in the real-time image corresponding to the current moment, and compare the current target feature vector with the pedestrian feature vector of each pedestrian respectively to generate the most similar feature vector and the corresponding maximum similarity; When the maximum similarity is greater than a preset threshold, the current target feature vector of the target pedestrian is updated using the most similar feature vector and the maximum confidence corresponding to the most similar pedestrian.

4. A robot following device based on pedestrian re-identification, which can implement the steps of the robot following method based on pedestrian re-identification according to claim 1, characterized in that: It includes acquisition module, location acquisition module, location prediction module, calculation module and control module. The acquisition module is used to acquire real-time images at a preset frequency during the following process; The position acquisition module is used to identify the target pedestrian in the real-time image and obtain the position information of the target pedestrian corresponding to the first historical moment, the second historical moment and the current moment respectively; The position prediction module is used to predict the estimated position of the target pedestrian at a preset time according to the position information and based on a second-order approximation algorithm; The calculation module is used to calculate the target rotation angle and target moving distance of the mobile robot according to the estimated position of the target pedestrian; The control module is used to generate a control instruction including the target rotation angle and the target moving distance to drive the mobile robot to follow the target pedestrian.

5. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the robot following method based on pedestrian re-identification according to any one of claims 1 to 3 are implemented.

6. A mobile robot, characterized in that: It comprises the computer-readable storage medium and a processor as described in claim 5, and when the processor executes the computer program on the computer-readable storage medium, it implements the steps of the robot following method based on pedestrian re-identification as described in any one of claims 1-3.

Images