A human-machine trust model based on fuzzy control systems

Abstract

The embodiments of this application relate to the field of human-computer interaction technology, and particularly to a human-computer trust model based on a fuzzy control system. The model includes: constructing circular motion tracking experiments and rectangular linear tracking experiments, where an operator controls the robot's movement, and the robot's end-effector pose and joint angles are recorded in real time; calculating safety indicators, robot singularity indicators, smoothness indicators, and tracking indicators at each time step during the human-computer interaction process; comprehensively evaluating the performance of these indicators at each time step to obtain a comprehensive evaluation index for each time step; establishing a dynamic trust model based on the comprehensive evaluation index for each time step using a weighted moving average algorithm; classifying the trust values ​​of the dynamic trust model based on fuzzy control, and adjusting the operator's control permissions over the robot according to the classification results. This model effectively improves the accuracy and general applicability of trust estimation.

Description

Technical Field

[0001] The embodiments of this application relate to the fields of human-computer interaction and human-computer trust technology, and in particular to a human-computer trust model based on a fuzzy control system. Background Technology

[0002] Human-computer interaction (HCI) is an interdisciplinary field that integrates computer science, mechanics, design, biology, psychology, and sociology. It focuses on how to enable robots to better serve human needs. HCI not only emphasizes technological optimization but also prioritizes the harmonious coexistence of technology and human activities. Its ultimate goal is to create an efficient, convenient, comfortable, and human-friendly interactive environment.

[0003] In the process of human-computer interaction, trust is the key to achieving high-quality human-computer interaction. If the operator's level of trust in the robot is inappropriate, it may lead to improper use or over-reliance on the robot. Currently, some researchers and scholars have conducted some research on human-computer trust models, but these studies mainly focus on human trust in robots. These models often involve human subjective emotions and require subjective questionnaires to determine model parameters, which are difficult to quantify with accurate data. Summary of the Invention

[0004] In view of this, embodiments of this application propose a human-machine trust model based on a fuzzy control system, which focuses on the robot's trust in humans, objectively evaluates the technical level of human operators, constructs a dynamic trust estimation mechanism, effectively improves the accuracy of trust estimation, and ensures that the human-machine trust model can adapt to diverse human-machine interaction scenarios.

[0005] In a first aspect, embodiments of this application propose a human-machine trust model based on a fuzzy control system, comprising: constructing and completing circular motion tracking experiments and rectangular straight-line tracking experiments, with an operator observing a reference path and controlling the robot to move, and recording the robot's end-effector pose and joint angles in real time during the movement; based on the real-time recorded end-effector pose and joint angles of the robot during the movement, calculating the safety index, robot singularity index, smoothness index, and tracking index for each time step in the human-machine interaction process; comprehensively evaluating the performance of the safety index, robot singularity index, smoothness index, and tracking index for each time step to obtain a comprehensive evaluation index for each time step; establishing a dynamic trust model based on the comprehensive evaluation index for each time step using a weighted moving average algorithm; classifying the trust value of the dynamic trust model based on fuzzy control, and adjusting the operator's control authority over the robot according to the classification results, with a higher trust value granting the operator higher control authority over the robot.

[0006] Optionally, the safety index is calculated using the following formula:

[0007] p S [n] = f(a[n], x - ,x + ,y - ,y + )

[0008]

[0009] Where, x - and x + These represent the lower and upper critical values ​​of the robot's acceleration when safety indicators change, respectively. - and y + These are the lower and upper critical values ​​of the safety index, respectively. a[n] represents the magnitude of the constant acceleration required for the robot to decelerate from its current position to the obstacle and stop. v[n] represents the robot's velocity at the time step when it collides with the obstacle. v0[n] represents the robot's velocity at the current time step. Δs[n] represents the distance between the robot's end effector and the obstacle at the current time step. n represents the current time step. f(·) is a fifth-order polynomial. f(·) in x - and x + The first and second derivatives at p are both 0. S [n] represents the safety index at the current time step.

[0010] Optionally, the robot singularity index is calculated through the following steps:

[0011] Based on the linear and angular velocities of the robot's end effector at the current time step, as well as the robot's joint angles, the Jacobian matrix for the current time step is constructed.

[0012] Perform singular value decomposition on the Jacobian matrix at the current time step to obtain the minimum singular value at the current time step;

[0013] Based on the minimum singular value at the current time step, the robot singularity index is calculated using the following formula:

[0014]

[0015] Where, σ min [n] represents the minimum singular value at the current time step, σ min [n-1] represents the minimum singular value of the previous time step, σ - and σ + These represent the lower and upper critical values ​​for the degree of singularity of the robot as its singularity index changes. and These are the lower high threshold and the upper high threshold, respectively. and These are the lower and upper critical values, respectively; β is a preset adjustment parameter; s(·) is a fifth-order polynomial; s(·) in σ - and σ + The first and second derivatives at p are both 0. RS [n] represents the robot singularity index at the current time step.

[0016] Alternatively, the smoothness index can be calculated using the following formula:

[0017] p SM [n] = g(S|S) - ,S + ,1,0)

[0018]

[0019] Where S represents the smoothness of the robot's motion, S - and S + These represent the lower and upper critical values ​​of smoothness for the robot as the smoothness index changes. Let g(·) represent the velocity vector of the robot's end effector in Cartesian space at the current time step, and g(·) be a fifth-order polynomial. - and S + The first and second derivatives at p are both 0. SM [n] represents the smoothness index of the current time step.

[0020] Optionally, the tracking performance index is calculated using the following formula:

[0021] p T [n] = h(T|T) - ,T + ,1,0)

[0022]

[0023] Where T represents the goodness of fit of the robot's motion, that is, the goodness of fit between the robot's end effector trajectory and the reference trajectory. - and T + Let x and y represent the lower and upper critical values ​​of the goodness of fit of the robot as the tracking performance index changes, respectively, and let X represent the maximum Cartesian space of the robot's end effector during motion. e and x d Let f(t) represent the actual position and the desired position of the robot's end effector, respectively. h(·) is a fifth-order polynomial, and h(·) in T - and T + The first and second derivatives at p are both 0. T [n] represents the tracking metric for the current time step.

[0024] Optionally, a comprehensive performance evaluation is performed on the safety index, robot singularity index, smoothness index, and tracking index at each time step to obtain a comprehensive evaluation index for each time step, which is achieved through the following formula:

[0025] p[n]=p S [n]·p RS [n]·{C+(1-C)·(γ SM p SM [n]+γ T p T [n])}

[0026] γ SM +γ T =1

[0027] Where, p S [n]、p RS [n]、p SM [n] and p T [n] represents the safety index, robot singularity index, smoothness index, and tracking index at the nth time step, respectively; C represents the preset first weighting coefficient; γ SM This represents the preset second specific gravity coefficient, γ. T denoted by the preset third weighting coefficient, p[n] represents the comprehensive evaluation index at the nth time step.

[0028] Optionally, a dynamic trust model can be established using a weighted moving average algorithm based on the comprehensive evaluation index at each time step, implemented through the following formula:

[0029]

[0030] Where n represents the total number of time steps that have been performed, the current time step is the nth time step, α is the preset time decay factor, p[nk] represents the comprehensive evaluation index of the nkth time step, and Trust[n] represents the dynamic trust model.

[0031] Optionally, the trust level values ​​of the dynamic trust model are classified based on fuzzy control, and the operator's control authority over the robot is adjusted according to the classification results, using the following formula:

[0032]

[0033] Where AL represents the operator's control authority over the robot, low represents low control authority, med represents medium control authority, high represents high control authority, and max represents the highest control authority.

[0034] Using the above methods, circular motion tracking experiments and rectangular straight-line tracking experiments were built and completed. Operators observed the reference path and controlled the robot to move. The robot's end-effector pose and joint angles were recorded in real time during the movement. Based on the real-time recorded end-effector pose and joint angles, four main dimensions of safety, singularity, smoothness, and tracking in the human-computer interaction process were quantitatively analyzed. The operator's technical level was comprehensively and objectively evaluated. A dynamic trust estimation mechanism, or dynamic trust model, was constructed using multi-factor fusion technology and weighted moving average algorithm. The trust value of the dynamic trust model was classified based on fuzzy control, and the operator's control authority over the robot was adjusted in a timely and accurate manner according to the classification results. The higher the trust value, the higher the control authority granted to the operator, thereby effectively improving the trust estimation accuracy, ensuring that the human-computer trust model can adapt to diverse human-computer interaction scenarios, and enhancing the robustness of the human-computer trust model in the face of uncertainty and dynamic changes.

[0035] In a second aspect, embodiments of this application provide an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to implement a human-machine trust model based on a fuzzy control system as described in the first aspect above.

[0036] Thirdly, embodiments of this application propose a computer-readable storage medium storing a computer program that, when executed by a processor, enables a human-machine trust model based on a fuzzy control system as described in the first aspect above.

[0037] It is understood that the beneficial effects of the second and third aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies of this application will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This is a flowchart of a human-machine trust model based on a fuzzy control system provided in one embodiment of this application;

[0040] Figure 2 This is a schematic diagram of a circular motion tracking experiment provided in one embodiment of this application;

[0041] Figure 3 This is a schematic diagram of a rectangular line tracking experiment provided in one embodiment of this application;

[0042] Figure 4 This is a distribution diagram of the motion nodes of a robot in a circular motion tracking experiment provided in one embodiment of this application;

[0043] Figure 5 This is a diagram showing the distribution of motion nodes of a robot in a rectangular linear tracking experiment provided in one embodiment of this application;

[0044] Figure 6 This is a schematic diagram of a fifth-order polynomial difference function provided in one embodiment of this application;

[0045] Figure 7 This is a schematic diagram of a minimum singularity exponential function provided in one embodiment of this application;

[0046] Figure 8 This is a schematic diagram illustrating how trust level changes with robot movement, provided in one embodiment of this application;

[0047] Figure 9 This is a schematic diagram of the structure of an electronic device provided in another embodiment of this application. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the various embodiments of this application will be described in detail below with reference to the accompanying drawings. In the various embodiments of this application, many technical details are presented to enable the reader to better understand this application. However, even without these technical details and various variations and modifications based on the following embodiments, the technical solutions claimed in this application can be implemented. The division of the following embodiments is only for convenience of description and should not constitute any limitation on the specific implementation of this application. The various embodiments can be combined with and referenced by each other without contradiction.

[0049] One embodiment of this application proposes a human-machine trust model based on a fuzzy control system. The implementation details of this fuzzy control system-based human-machine trust model are described below. These details are provided for ease of understanding and are not essential for implementing this solution. The specific implementation process of the fuzzy control system-based human-machine trust model proposed in this embodiment can be as follows: Figure 1 As shown, it includes:

[0050] Step 101: Set up and complete the circular motion tracking experiment and the rectangular straight line tracking experiment. The operator observes the reference path and controls the robot to move, and records the end pose and joint angles of the robot in real time during the movement.

[0051] In the specific implementation, it is first necessary to build and complete the circular motion tracking experiment and the rectangular straight line tracking experiment. The operator observes the reference path and controls the virtual Kuka robot on the Simulink platform to move by controlling the Touch controller. The server records the end pose and joint angles of the robot in real time during the movement.

[0052] In one example, a circular motion tracking experiment is as follows: Figure 2 As shown, the rectangular line tracking experiment is as follows: Figure 3 As shown, the distribution of motion nodes of the robot in the circular motion tracking experiment is as follows: Figure 4 As shown, the distribution of motion nodes of the robot in the rectangular line tracking experiment is as follows: Figure 5 As shown.

[0053] Step 102: Based on the real-time recorded end-effector pose and joint angles of the robot during its movement, calculate the safety index, robot singularity index, smoothness index, and tracking index for each time step during the human-computer interaction process.

[0054] In practical implementation, the process of a robot moving from one node to the next is called a time step. In the two experiments, the server calculated the safety index, robot singularity index, smoothness index and tracking index of each time step in the human-computer interaction process based on the real-time recorded end pose and joint angles of the robot during the movement.

[0055] In one example, safety metrics and robot singularity metrics can be calculated based on data recorded in a rectangular straight-line tracking experiment, while smoothness metrics and tracking metrics can be calculated based on data recorded in a circular motion tracking experiment.

[0056] In one example, the safety metric is defined based on the probability of the robot colliding with surrounding objects under operator control. The safety metric can be calculated using the following formula:

[0057] p S [n] = f(a[n], x - ,x + ,y - ,y + )

[0058]

[0059] Where, x - and x+ These represent the lower and upper critical values ​​of the robot's acceleration when safety indicators change, respectively. - and y + These are the lower and upper critical values ​​of the safety index, respectively. a[n] represents the magnitude of the constant acceleration required for the robot to decelerate from its current position to the obstacle and stop. v[n] represents the robot's velocity at the time step when it collides with the obstacle. v0[n] represents the robot's velocity at the current time step (set to 0). Δs[n] represents the distance between the robot's end effector and the obstacle at the current time step. n represents the current time step. f(·) is a fifth-order polynomial. f(·) in x - and x + The first and second derivatives at p are both 0. S [n] represents the safety index at the current time step. Figure 6 This is a schematic diagram of a fifth-order polynomial interpolation function.

[0060] In one example, the robot singularity index is calculated through the following steps.

[0061] First, the server needs to construct the Jacobian matrix for the current time step based on the linear and angular velocities of the robot's end effector and the robot's joint angles. The definition of the Jacobian matrix is ​​as follows:

[0062]

[0063] Among them, v e =[v ex ,v ey ,v ez ] represents the linear velocity of the robot's end effector, ω. e =[ω ex ,ω ey ,ω ez ] represents the angular velocity of the robot's end effector, Θ represents the joint angle of the robot's robotic arm, and J represents the Jacobian matrix.

[0064] Subsequently, the server performs singular value decomposition (SVD) on the Jacobian matrix at the current time step to obtain the minimum singular value at the current time step. SVD is an important tool in modern numerical analysis, especially numerical computation. A Jacobian matrix J must have orthogonal matrices U and V such that J = UΣV. The minimum singular value of the Jacobian matrix is ​​often used to represent the degree of singularity. Numerous experiments have shown that the minimum singular value of the Jacobian matrix can well represent the degree of singularity in robot joint configurations. Therefore, the server performs SVD on J = UΣV and searches for the minimum singular value σ among the obtained singular values. min .

[0065] Finally, the server calculates the robot singularity index based on the minimum singularity value at the current time step. This calculation can be performed using the following formula:

[0066]

[0067] σ - <σ min [n]<σ +

[0068] Where, σ min [n] represents the minimum singular value at the current time step, σ min [n-1] represents the minimum singular value of the previous time step, σ - and σ + These represent the lower and upper critical values ​​for the degree of singularity of the robot as its singularity index changes. and These are the lower high threshold and the upper high threshold, respectively. and These are the lower and upper critical values, respectively. β is a preset adjustment parameter, 0 < β << 1. s(·) is a fifth-order polynomial, and s(·) in σ - and σ + The first and second derivatives at p are both 0. RS [n] represents the robot singularity index at the current time step. Figure 7 A schematic diagram of an exponential function with a minimum singular value is shown.

[0069] Near the singular configuration, σ min →0, the joint angular velocity will be very large, σ min The larger the value, the faster the robotic arm's end effector responds to joint movements; therefore, we hope that σ... min The larger the better. When σ min →σ - The robot's configuration is close to a bizarre configuration, posing a risk of excessive joint angular velocity. Furthermore, when σ... min →σ + This indicates that the robot is unlikely to become a strange configuration.

[0070] In addition, whether the robot is close to a singular configuration is also a factor in calculating p. RS Another important factor to consider when [n] is that p... when the robotic arm moves away from the singular configuration. RS [n] should be relatively high, especially when the robotic arm approaches a singular configuration, p RS [n] should be low, therefore σ during robot movement min p when it is increasing RS [n] should be greater than σ min p when there is a decreasing trendRS [n], therefore when σ min [n]≥σ min When [n-1], use the higher-lower critical value and the higher-upper critical value, while σ min [n]<σ min When [n-1], use the lower threshold and the upper threshold.

[0071] In one example, the smoothness index is used to quantify the smoothness of the robot's movements when an operator manipulates the robot to perform a task; the smoothness index is denoted as p. SM p SM ∈[0,1], p SM =0 is considered as highly non-smooth motion, p SM A smoothness index of 1 indicates smooth operator control and good robot control. The smoothness index can be calculated using the following formula:

[0072] p SM [n] = g(S|S) - ,S + ,1,0)

[0073]

[0074] Where S represents the smoothness of the robot's motion, S - and S + These represent the lower and upper critical values ​​of smoothness for the robot as the smoothness index changes. Let g(·) represent the velocity vector of the robot's end effector in Cartesian space at the current time step. g(·) is a fifth-order polynomial with values ​​ranging from [0,1]. g(·) in S... - and S + The first and second derivatives at p are both 0. sM [n] represents the smoothness index of the current time step.

[0075] In one example, the degree to which the robot's end effector trajectory fits the reference trajectory reflects the robot's tracking ability to a certain extent. The tracking ability index can be calculated using the following formula:

[0076] p T [n] = h(T|T) - ,T + ,1,0)

[0077]

[0078] 0 <T - <T + <<1

[0079] Where T represents the goodness of fit of the robot's motion, that is, the goodness of fit between the robot's end effector trajectory and the reference trajectory. - and T + Let x and y represent the lower and upper critical values ​​of the goodness of fit of the robot as the tracking performance index changes, respectively, and let X represent the maximum Cartesian space of the robot's end effector during motion. e and x d Let f(t) represent the actual position and the desired position of the robot's end effector, respectively. h(·) is a fifth-order polynomial, and the function value of h(·) ranges from [0,1]. h(·) in T - and T + The first and second derivatives at p are both 0. T [n] represents the tracking metric for the current time step.

[0080] Step 103: Conduct a comprehensive performance evaluation of the safety index, robot singularity index, smoothness index, and tracking index for each time step to obtain a comprehensive evaluation index for each time step.

[0081] In practical implementation, after calculating the safety index, robot singularity index, smoothness index, and tracking index for each time step, the server can perform a comprehensive performance evaluation of the safety index, robot singularity index, smoothness index, and tracking index for each time step to obtain a comprehensive evaluation index for each time step.

[0082] In one example, the server performs a comprehensive performance evaluation of safety metrics, robot singularity metrics, smoothness metrics, and tracking metrics for each time step, resulting in a comprehensive evaluation metric for each time step, achieved through the following formula:

[0083] p[n]=p S [n]·p RS [n]·{C+(1-C)·(γ SM p SM [n]+γ T p T [n])}

[0084] γ SM +γ T =1

[0085] Where, p S [n]、p RS [n]、p SM [n] and p T [n] represents the safety index, robot singularity index, smoothness index, and tracking index at the nth time step, respectively; C represents the preset first weighting coefficient; γ SM This represents the preset second specific gravity coefficient, γ. Tp represents the preset third weighting coefficient, and p[n] represents the comprehensive evaluation index at the nth time step. S [n] and p RS [n] is closely related to the safe execution of the task and is considered a key factor, p SM [n] and p T [n] is considered a secondary factor.

[0086] Step 104: Using the weighted moving average algorithm, a dynamic trust model is established based on the comprehensive evaluation index at each time step.

[0087] In practice, after the server calculates the comprehensive evaluation index for each time step, it can use a weighted moving average algorithm to establish a dynamic trust model based on the comprehensive evaluation index for each time step.

[0088] In one example, the server uses a weighted moving average algorithm to build a dynamic trust model based on a comprehensive evaluation index at each time step, which can be achieved through the following formula:

[0089]

[0090] Where n represents the total number of time steps that have been performed, the current time step is the nth time step, α is the preset time decay factor, p[nk] represents the comprehensive evaluation index of the nkth time step, and Trust[n] represents the dynamic trust model.

[0091] Trust ∈ [0,1], where Trust = 0 indicates that the robot has no trust in the operator, and Trust = 1 indicates that the robot has complete trust in the operator. The setting of the time decay factor reduces the sensitivity to historical performance, meaning that recent performance has a greater impact on the dynamic trust model.

[0092] Step 105: Classify the trust level value of the dynamic trust model through fuzzy control, and adjust the operator's control authority over the robot according to the classification result. The higher the trust level value, the higher the control authority granted to the operator over the robot.

[0093] In practice, after the server establishes a dynamic trust model, it can classify the trust value of the dynamic trust model through fuzzy control, and adjust the operator's control permissions over the robot according to the classification results. The higher the trust value, the higher the control permissions granted to the operator over the robot, and the lower the trust value, the lower the control permissions granted to the operator over the robot.

[0094] In one example, the server classifies the trust level values ​​of the dynamic trust model based on fuzzy control and adjusts the operator's control permissions over the robot according to the classification results. This can be achieved using the following formula:

[0095]

[0096] Where AL represents the operator's control authority over the robot, low represents low control authority, med represents medium control authority, high represents high control authority, and max represents the highest control authority.

[0097] In one example, the change in trust level with robot movement during the experiment can be shown as follows: Figure 8 As shown, as the operator's skill level improves over time, the robot's trust in the operator also increases, leading to a greater degree of control over the robot granted to the operator.

[0098] In this embodiment, circular motion tracking experiments and rectangular straight-line tracking experiments were set up. The operator observed the reference path and controlled the robot to move. The end-effector pose and joint angles of the robot during the movement were recorded in real time. Based on the real-time recorded end-effector pose and joint angles of the robot during the movement, four main dimensions of safety, singularity, smoothness and tracking in the human-computer interaction process were quantitatively analyzed. The operator's technical level was comprehensively and objectively evaluated. Multi-factor fusion technology and weighted moving average algorithm were used to construct a dynamic trust estimation mechanism, that is, to construct a dynamic trust model. The trust value of the dynamic trust model was classified based on fuzzy control, and the operator's control authority over the robot was adjusted in a timely and accurate manner according to the classification results. The higher the trust value, the higher the control authority granted to the operator over the robot, thereby effectively improving the trust estimation accuracy, ensuring that the human-computer trust model can adapt to diverse human-computer interaction scenarios, and enhancing the robustness of the human-computer trust model in the face of uncertainty and dynamic changes.

[0099] The steps of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this application. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, but without changing the core design of the algorithm and process, are also within the scope of protection of this application.

[0100] Another embodiment of this application provides an electronic device, the specific structure of which can be as follows: Figure 9 As shown, it includes: at least one processor 301; and a memory 302 communicatively connected to the at least one processor 301; wherein the memory 302 stores instructions executable by the at least one processor 301, the instructions being executed by the at least one processor 301 to enable the at least one processor 301 to implement a human-machine trust model based on a fuzzy control system as described in the above embodiments.

[0101] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and the memory. The bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.

[0102] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.

[0103] Another embodiment of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, can implement a human-machine trust model based on a fuzzy control system as described in the above embodiments.

[0104] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, ROM (Read-Only Memory), RAM (Random Access Memory), a magnetic disk, or an optical disk.

[0105] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing this application, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of this application.

Claims

1. A human-machine trust model based on a fuzzy control system, characterized in that, include: Build and complete circular motion tracking experiments and rectangular line tracking experiments. Operators observe the reference path and control the robot to move, and record the robot's end pose and joint angles in real time during the movement. Based on the robot's end-effector pose and joint angles recorded in real time during its movement, safety indicators, robot singularity indicators, smoothness indicators, and tracking indicators are calculated for each time step in the human-computer interaction process. A comprehensive performance evaluation of the safety index, robot singularity index, smoothness index, and tracking index at each time step is performed to obtain a comprehensive evaluation index for each time step. A dynamic trust model is established using a weighted moving average algorithm based on a comprehensive evaluation index at each time step. Based on fuzzy control, the trust level values ​​of the dynamic trust model are classified, and the operator's control authority over the robot is adjusted according to the classification results. The higher the trust level value, the higher the control authority granted to the operator over the robot.

2. The human-machine trust model based on a fuzzy control system according to claim 1, characterized in that, The safety index is calculated using the following formula: p S [n]=f(a[n],x - ,x + ,y - ,y + ) Where, x - and x + These represent the lower and upper critical values ​​of the robot's acceleration when safety indicators change, respectively. - and y + These are the lower and upper critical values ​​of the safety index, respectively. a[n] represents the magnitude of the constant acceleration required for the robot to decelerate from its current position to the obstacle and stop. v[n] represents the robot's velocity at the time step when it collides with the obstacle. v0[n] represents the robot's velocity at the current time step. Δs[n] represents the distance between the robot's end effector and the obstacle at the current time step. n represents the current time step. f(·) is a fifth-order polynomial. f(·) in x - and x + The first and second derivatives at p are both 0. S [n] represents the safety index at the current time step.

3. The human-machine trust model based on a fuzzy control system according to claim 1, characterized in that, The robot singularity index is calculated through the following steps: Based on the linear and angular velocities of the robot's end effector at the current time step, as well as the robot's joint angles, the Jacobian matrix for the current time step is constructed. Perform singular value decomposition on the Jacobian matrix at the current time step to obtain the minimum singular value at the current time step; Based on the minimum singular value at the current time step, the robot singularity index is calculated using the following formula: Where, σ min [n] represents the minimum singular value at the current time step, σ min [n-1] represents the minimum singular value of the previous time step, σ - and σ + These represent the lower and upper critical values ​​for the degree of singularity of the robot as its singularity index changes. and These are the lower high threshold and the upper high threshold, respectively. and These are the lower and upper critical values, respectively; β is a preset adjustment parameter; s(·) is a fifth-order polynomial; s(·) in σ - and σ + The first and second derivatives at p are both 0. RS [n] represents the robot singularity index at the current time step.

4. The human-machine trust model based on a fuzzy control system according to claim 1, characterized in that, The smoothness index is calculated using the following formula: p SM [n]=g(S|S - ,S + ,1,0) Where S represents the smoothness of the robot's motion, S- and S- + These represent the lower and upper critical values ​​of smoothness for the robot as the smoothness index changes. Let g(·) represent the velocity vector of the robot's end effector in Cartesian space at the current time step, and g(·) be a fifth-order polynomial. - and S + The first and second derivatives at p are both 0. SM [n] represents the smoothness index of the current time step.

5. A human-machine trust model based on a fuzzy control system according to claim 1, characterized in that, The tracking performance index is calculated using the following formula: p T [n]=h(T|T-,T + ,1,0) Where T represents the goodness of fit of the robot's motion, that is, the goodness of fit between the trajectory of the robot's end effector and the reference trajectory, and T- and T- + Let x and y represent the lower and upper critical values ​​of the goodness of fit of the robot as the tracking performance index changes, respectively, and let X represent the maximum Cartesian space of the robot's end effector during motion. e and x d Let f(t) represent the actual position and the desired position of the robot's end effector, respectively. h(-) is a fifth-order polynomial, and h(·) is expressed in T- and T-. + The first and second derivatives at p are both 0. T [n] represents the tracking metric for the current time step.

6. A human-machine trust model based on a fuzzy control system according to any one of claims 1 to 5, characterized in that, A comprehensive performance evaluation is performed on the safety metrics, robot singularity metrics, smoothness metrics, and tracking metrics at each time step to obtain the comprehensive evaluation index for each time step, which is achieved through the following formula: p[n]=p S [n]·p RS [n]·{C+(1-C)·(γ SM p SM [n]+Y T pT[n])} Y SM +g T =1 Where, p S [n]、p RS [n]、p SM [n] and p T [n] represents the safety index, robot singularity index, smoothness index, and tracking index at the nth time step, respectively; C represents the preset first weighting coefficient; β SM This represents the preset second specific gravity coefficient, γ. T denoted by the preset third weighting coefficient, p[n] represents the comprehensive evaluation index at the nth time step.

7. A human-machine trust model based on a fuzzy control system according to claim 6, characterized in that, Using a weighted moving average algorithm, a dynamic trust model is established based on the comprehensive evaluation index at each time step, implemented through the following formula: Where n represents the total number of time steps that have been performed, the current time step is the nth time step, α is the preset time decay factor, p[nk] represents the comprehensive evaluation index of the nkth time step, and Trust[n] represents the dynamic trust model.

8. A human-machine trust model based on a fuzzy control system according to claim 7, characterized in that, The trust level values ​​of the dynamic trust model are classified based on fuzzy control, and the operator's control authority over the robot is adjusted according to the classification results. This is achieved through the following formula: Where AL represents the operator's control authority over the robot, low represents low control authority, med represents medium control authority, high represents high control authority, and max represents the highest control authority.

9. An electronic device, characterized in that, include: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor, which are executed by the at least one processor to enable the at least one processor to implement a human-machine trust model based on a fuzzy control system as described in any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it can implement a human-machine trust model based on a fuzzy control system as described in any one of claims 1 to 8.

Images