Physical human-computer interaction-oriented robot time domain vibration detection method and device

Abstract

The invention provides a robot time domain vibration detection method and device for physical man-machine interaction, and relates to the technical field of robot physical man-machine interaction. Aiming at the problems in the prior art, the method solves the defects of the prior art in the aspects of accuracy, robustness, real-time performance and user friendliness by acquiring an interaction force signal data block, identifying and registering an extreme point in the data block, checking the validity of the extreme point, calculating a vibration index and smoothing the vibration index; the instability degree of the physical man-machine interaction system can be effectively quantified, reliable quantitative indication is provided for a vibration suppression strategy, and it is ensured that the system stably and safely completes cooperation tasks.

Description

Technical Field

[0001] The present invention relates to the technical field of physical human-robot interaction, and particularly to a method and device for detecting robot time-domain vibration for physical human-robot interaction. Background Art

[0002] In the field of physical human-robot interaction, humans and robots share a working space and maintain physical contact, forming a complex coupled dynamic system. Such systems are widely used in tasks such as human-robot collaborative operation and robot-assisted medical treatment, aiming to combine the perceptual and decision-making abilities of humans with the high-load and high-precision capabilities of robots to achieve complementary advantages and improve operation performance. However, due to the rigid interaction between humans and robots, the system often faces vibration and instability problems, resulting in the interruption or failure of collaborative tasks, and even threatening the safety of humans and robots.

[0003] Existing physical human-robot interaction instability detection technologies are mainly divided into the following categories: First, methods based on additional devices: Use devices such as electromyogram (EMG) and force / torque sensors to measure or estimate the stiffness of the human arm in real time as a sign of instability detection. However, these methods are greatly affected by signal noise, and the use of additional devices reduces user-friendliness.

[0004] Second, methods based on energy analysis: Indirectly represent the instability of the system through energy changes. However, the energy analysis method cannot detect instability in a timely manner, and in some cases, it cannot accurately reflect the vibration intensity of the system.

[0005] Third, methods based on time-domain signal analysis: Calculate the vibration index by performing online oscillation waveform recognition on the interaction force signal or detecting extreme points of the robot end velocity. However, existing methods are sensitive to signal noise, and the calculation of the vibration index is sensitive to the number of extreme points within the time window, prone to false positive misjudgments and sudden changes in the vibration index.

[0006] Fourth, methods based on frequency-domain signal analysis: Use discrete Fourier transform (DFT) or fast Fourier transform (FFT) to perform frequency analysis on the robot end position or interaction force signal, and provide an instability indication by calculating the proportion of unstable frequency components. However, the frequency-domain analysis method has a trade-off between sampling time and frequency resolution, and it is difficult to effectively distinguish intentional low-frequency motion and unintentional high-frequency oscillation.

[0007] Fifth, heuristic methods based on admittance interaction models: Calculate the deviation from the nominal behavior of the admittance control robot according to the current state to provide an instability detection sign. However, this method highly depends on admittance parameters and the velocity and acceleration boundaries of the robot, and the detection threshold is difficult to adjust and changes over time.

[0008] In summary, the existing technologies have problems of insufficient accuracy, robustness, real-time performance, and user-friendliness, and it is difficult to meet the vibration detection requirements of complex physical human-machine interaction systems. Summary of the Invention

[0009] In view of the above problems in the existing technologies, a robot time-domain vibration detection method for physical human-machine interaction is proposed in the first aspect of the present invention, which specifically includes the following steps: Step S1: Obtain an interaction force signal data block, and online package the interaction force signal to be detected through a preset fixed time window to form a data block for subsequent processing. The width of the fixed time window is set according to the signal noise characteristics, specifically including: in each sampling period, the latest detected interaction force signal point is successively filled into the rightmost side of a fixed time window with a constant width of N sampling periods, and the detected signal point at the leftmost side of the fixed time window is deleted to finally obtain the data block within the fixed time window , where t k represents the time point of the k-th sampling period, y k represents the interaction force signal value of the k-th sampling period, N is the preset number of sampling periods, T s is the sampling period, and T N is the width of the fixed time window ; Step S2: Identify and register the extreme points in the data block. The extreme points include maximum points and minimum points, specifically including: in each sampling period, search for the maximum value and the minimum value in the data block , and judge whether the center point of the fixed time window is a maximum point or a minimum point. If so, register it and save it to the beginning of a preset extreme point container ; Step S3: Perform validity verification on the extreme points registered and saved in Step S2, and eliminate the invalid extreme points to obtain valid extreme points; Step S4: Based on a variable time window, calculate the vibration index using the valid extreme points, and the width of the variable time window is adjusted in real time according to the distribution of the valid extreme points; Step S5: Smooth the vibration index, and filter the vibration index using a second-order low-pass filter to obtain a smoothed vibration index.

[0010] In combination with the first aspect, in some implementation manners of the first aspect, in Step S1, the width of the fixed time window is set according to the signal noise characteristics, specifically including: If the intensity of the signal noise exceeds the preset noise threshold, increase the width of the fixed time window to reduce the impact of noise on the detection of extreme points; If the intensity of the signal noise does not exceed the preset noise threshold, reduce the width of the fixed time window to improve the detection sensitivity to high-frequency vibrations.

[0011] Combined with the first aspect, in some implementation manners of the first aspect, in step S2, identifying and registering the extreme points in the data block includes: Search for the maximum and minimum values in the data block, and use them as the maximum and minimum extreme points of the data block respectively; Judge the center point of the fixed time window Whether it is a maximum extreme point or a minimum extreme point. If so, register it and save it to the preset extreme point container At the beginning. If the center point is neither a maximum extreme point nor a minimum extreme point, do nothing. Among them, the center point The calculation formula is , where c is the center point index of the fixed time window.

[0012] Combined with the first aspect, in some implementation manners of the first aspect, if a new extreme point is registered and saved to the preset extreme point container in step S2 At the beginning, the new extreme point is recorded as the first extreme point in the preset extreme point container In step S3, perform an effectiveness check on the extreme points registered and saved in step S2, including: If the new extreme point is the only extreme point in the extreme point container, determine that the new extreme point is a valid extreme point; If the extreme point container contains more than one extreme point, judge the effectiveness of the new extreme point by comparing the absolute value of the difference between the new extreme point and the second extreme point in the extreme point container Whether it exceeds the preset detection threshold. Among them, the detection threshold is a preset multiple of the average absolute deviation of the interaction force signal values of all valid extreme points in the extreme point container, and the value range of the preset multiple is greater than or equal to 1.2 and less than or equal to 2.5.

[0013] Combined with the first aspect, in some implementation manners of the first aspect, if the extreme point container contains more than one extreme point, judge the effectiveness of the new extreme point by comparing the absolute value of the difference between the new extreme point and the second extreme point in the extreme point container, including: When the absolute value of the difference between the new extreme point and the second extreme point in the extreme point container exceeds the detection threshold, mark the new extreme point as a valid extreme point; When the absolute value of the difference does not exceed the detection threshold, remove the new extreme point from the extreme point container.

[0014] In combination with the first aspect, in some implementation manners of the first aspect, in step S4, the width of the variable time window is adjusted in real time according to the distribution of the effective extreme points, including: If the distribution span of the effective extreme points participating in the vibration index calculation increases, the width of the variable time window increases to avoid sudden changes in the vibration index; If the distribution span of the effective extreme points participating in the vibration index calculation decreases, the width of the variable time window decreases to improve the real-time performance of vibration detection.

[0015] In combination with the first aspect, in some implementation manners of the first aspect, in step S4, the vibration index is calculated by using the effective extreme points, including: In each sampling period, an index variable is determined according to the number of effective extreme points. The calculation formula of the index variable m is: where, is the number of effective extreme points included in the extreme point container Count is the maximum number of effective extreme points preset to participate in the vibration index calculation; Based on the index variable and the time interval between the effective extreme points, the vibration index of the current sampling period is calculated. The calculation formula of the vibration index V k is: where, Δt is the maximum time interval threshold between the preset effective extreme points, is the signal value of the i-th effective extreme point in the extreme point container, is the signal value of the (i + 1)-th effective extreme point in the extreme point container, is the time point of the i-th effective extreme point in the extreme point container, is the time point of the (i + 1)-th effective extreme point in the extreme point container, is the center point of the fixed time window of the time point, is the time point of the first effective extreme point in the extreme point container.

[0016] In combination with the first aspect, in some implementation manners of the first aspect, in step S5, a second-order low-pass filter is used to filter the vibration index, including: In each sampling period, the vibration index is input into the second-order low-pass filter to obtain a smoothed vibration index. The transfer function G(s) of the second-order low-pass filter is: where, s is the complex frequency of the Laplace transform.

[0017] In a second aspect, the present invention provides a robot time-domain vibration detection device for physical human-robot interaction. The detection device adopts the detection method provided in any one of the above embodiments, including: A signal acquisition unit, including a six-axis force sensor deployed at the end effector of the robot, for real-time acquisition of interaction force signals; A data caching unit, composed of a dual-port memory, for storing interaction force signals and forming data blocks with a fixed time window; An extreme point processing unit, including a digital signal processor, configured to perform the following operations: identify maximum and minimum extreme points in the data block, store the extreme points corresponding to the center point of the fixed time window into the extreme point container, and perform validity verification on the extreme points; A vibration index calculation unit, composed of a variable time window controller and an arithmetic logic unit, for calculating the vibration index based on valid extreme points; A filtering output unit, including a second-order low-pass filter, for smoothing the vibration index and outputting it.

[0018] In a third aspect, the present invention provides a computer-readable storage medium, and the storage medium stores a computer program, and the computer program is used to execute the detection method provided in any of the above embodiments.

[0019] In a fourth aspect, the present invention provides an electronic device, and the electronic device includes: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is used to execute the detection method provided in any of the above embodiments.

[0020] Compared with the prior art, by introducing the combination of a constant-width time window and a variable-width time window, the present invention can detect the extreme points of the interaction force signal in real time and calculate the vibration index, and has the following remarkable technical effects: First, it improves the accuracy of vibration detection: By using a constant-width time window to online detect the extreme points of the interaction force signal, it can effectively capture the local characteristics of the signal and avoid misjudgment caused by signal noise. Before calculating the vibration index, perform validity verification on the extreme points to eliminate invalid extreme points, further improving the accuracy of the vibration index. By using a variable-width time window to adjust the calculation range of the vibration index in real time, it can avoid sudden changes and distortions of the vibration index caused by fluctuations in the number of extreme points.

[0021] Second, it enhances the robustness to signal noise: The width of the constant-width time window can be dynamically adjusted according to the signal noise characteristics to ensure effective detection of extreme points in different noise environments. Eliminate invalid extreme points through validity verification to reduce the interference of noise on the calculation of the vibration index.

[0022] Third, it improves real-time performance: By adopting the method of online encapsulation of interaction force signals, data blocks are generated in real time and extreme point detection and vibration index calculation are carried out to ensure that the system can quickly respond to vibration changes. The design of the variable-width time window enables the calculation of the vibration index to be adjusted in real time according to the distribution of extreme points, avoiding response delays caused by fixed time windows.

[0023] Fourth, it enhances the smoothness and practicality of the vibration index: The vibration index is smoothed by using a second-order low-pass filter to eliminate high-frequency fluctuations in the vibration index, making it more continuous and smooth. The smoothed vibration index can more intuitively reflect the instability degree of the system and provide a reliable quantitative indication for vibration suppression strategies.

[0024] Fifth, it improves user-friendliness: Without additional equipment, the interaction force signal is directly used for vibration detection, simplifying the system structure and reducing the usage cost. By real-time monitoring and evaluating the instability of the system, vibration suppression measures can be taken in a timely manner to ensure the stability and safety of the physical human-robot interaction system. Description of the Drawings

[0025] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0026] Figure 1 The following shows a schematic flowchart of a robot time-domain vibration detection method for physical human-robot interaction provided by an embodiment of the present invention.

[0027] Figure 2 The following shows a schematic structural diagram of a robot time-domain vibration detection device for physical human-robot interaction provided by an embodiment of the present invention. Detailed Embodiments

[0028] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0029] The following will explain the detailed embodiments of the present invention.

[0030] In view of the above problems in the prior art, the present invention proposes a robot time-domain vibration detection method for physical human-robot interaction, which solves the deficiencies of the prior art in terms of accuracy, robustness, real-time performance, and user-friendliness, can effectively quantify the instability degree of the physical human-robot interaction system, provide a reliable quantitative indication for vibration suppression strategies, and ensure that the system stably and safely completes collaborative tasks.

[0031] Embodiment 1

[0032] As Figure 1 shown, the present invention proposes a robot time-domain vibration detection method for physical human-robot interaction, which specifically includes the following steps: Step S1: Obtain an interaction force signal data block, and online encapsulate the interaction force signal to be detected through a preset fixed time window to form a data block for subsequent processing. The width of the fixed time window is set according to the signal noise characteristics.

[0033] Specifically, it includes: in each sampling period, fill the latest detected interaction force signal points one by one into the rightmost side of a fixed time window with a constant width of N sampling periods, and delete the detected signal points on the leftmost side of the fixed time window , and finally obtain a data block within the fixed time window, where t k represents the time point of the k-th sampling period, y k represents the interaction force signal value of the k-th sampling period, N is the preset number of sampling periods, T s is the sampling period, and T N is the width of the fixed time window. .

[0034] The core of this design concept is to ensure that the data block always contains the latest interaction force signal through the dynamic update of the fixed time window, providing real-time data support for subsequent extreme point detection.

[0035] Step S2: Identify and register the extreme points in the data block. The extreme points include maximum points and minimum points.

[0036] Specifically, it includes: in each sampling period, search for the maximum value and the minimum value in the data block , and judge whether the center point of the fixed time window is a maximum point or a minimum point. If so, register it and save it to the beginning of a preset extreme point container .

[0037] The design concept of this step is to capture the local features in the interactive force signal through the detection of extreme points, providing basic data for the subsequent calculation of the vibration index. The process of identifying extreme points depends on the setting of the width of the fixed time window. If the window width is too small, it may lead to noise interference, and if the window width is too large, it may miss high-frequency vibration information. Therefore, the reasonable setting of the window width is the key to ensuring the accuracy of extreme point detection.

[0038] Step S3: Conduct validity tests on the extreme points registered and saved in Step S2, eliminate the invalid extreme points, and obtain valid extreme points.

[0039] The design concept of this step is to further screen out reliable extreme points through validity tests, avoiding misjudgments caused by noise or signal fluctuations, thereby improving the accuracy of the vibration index.

[0040] Step S4: Calculate the vibration index based on the valid extreme points using a variable time window, and the width of the variable time window is adjusted in real time according to the distribution of the valid extreme points.

[0041] The design concept of this step is to avoid mutations and distortions of the vibration index caused by a fixed time window through the dynamic adjustment of the variable time window, thereby enhancing the robustness and real-time performance of the vibration index.

[0042] Step S5: Smooth the vibration index, filter the vibration index using a second-order low-pass filter, and obtain the smoothed vibration index.

[0043] The design concept of this step is to eliminate high-frequency fluctuations in the vibration index through filtering, making it more continuous and smooth, thereby providing a reliable quantitative indication for the vibration suppression strategy.

[0044] The present invention solves the deficiencies of the prior art in terms of accuracy, robustness, real-time performance, and user-friendliness, can effectively quantify the instability degree of the physical human-machine interaction system, provide a reliable quantitative indication for the vibration suppression strategy, and ensure the stable and safe completion of collaborative tasks by the system.

[0045] Combined with the first aspect, in some implementation manners of the first aspect, in Step S1, the width of the fixed time window is set according to the signal noise characteristics, specifically including: If the intensity of the signal noise exceeds the preset noise threshold, increase the width of the fixed time window to reduce the influence of noise on the detection of extreme points; If the intensity of the signal noise does not exceed the preset noise threshold, reduce the width of the fixed time window to improve the detection sensitivity to high-frequency vibrations.

[0046] The embodiment of the present invention further defines a method for setting the width of a fixed time window. The width of the fixed time window is dynamically adjusted according to the signal noise characteristics to ensure the robustness against signal noise. By dynamically adjusting the width of the fixed time window, it is ensured that the system can effectively detect extreme points in different noise environments, thereby improving the accuracy and robustness of vibration detection.

[0047] In combination with the first aspect, in some implementation manners of the first aspect, in step S2, identifying and registering extreme points in the data block includes: Searching for the maximum value and the minimum value in the data block, and respectively taking them as the maximum extreme value and the minimum extreme value of the data block; Judging whether the center point of the fixed time window is a maximum extreme point or a minimum extreme point. If so, register it and save it to the beginning of a preset extreme point container. If the center point is neither a maximum extreme point nor a minimum extreme point, no operation is performed. Among them, the center point is calculated by the formula

[0048] where c is the index of the center point of the fixed time window. Through the detection of extreme points, the embodiment of the present invention captures local features in the interaction force signal, providing basic data for subsequent vibration index calculation. The identification process of extreme points depends on the setting of the width of the fixed time window. If the window width is too small, it may lead to noise interference. If the window width is too large, high-frequency vibration information may be missed. Therefore, the reasonable setting of the window width is the key to ensuring the accuracy of extreme point detection.

[0049] In combination with the first aspect, in some implementation manners of the first aspect, if a new extreme point is registered and saved to the beginning of a preset extreme point container in step S2, the new extreme point is recorded as the first extreme point in the preset extreme point container. In step S3, performing an effectiveness test on the extreme points registered and saved in step S2 includes: If the new extreme point is the only extreme point in the extreme point container, determine that the new extreme point is a valid extreme point; If there is more than one extreme point in the extreme point container, determine the effectiveness of the new extreme point by comparing the absolute value of the difference between the new extreme point and the second extreme point in the extreme point container Whether it exceeds a preset detection threshold. Among them, the detection threshold is a preset multiple of the average absolute deviation of the interaction force signal values of all valid extreme points in the extreme point container, and the value range of the preset multiple is greater than or equal to 1.2 and less than or equal to 2.5.

[0050] Through the effectiveness test, the embodiment of the present invention further screens out reliable extreme points, avoiding misjudgment caused by noise or signal fluctuations, thereby improving the accuracy of the vibration index.

[0051] In combination with the first aspect, in some implementation manners of the first aspect, if there is more than one extreme point in the extreme point container, it is determined whether the absolute value of the difference between the new extreme point and the second extreme point in the extreme point container exceeds a preset detection threshold to determine the validity of the new extreme point, including: When the absolute value of the difference between the new extreme point and the second extreme point in the extreme point container exceeds the detection threshold, the new extreme point is marked as a valid extreme point; When the absolute value of the difference does not exceed the detection threshold, the new extreme point is removed from the extreme point container.

[0052] The embodiment of the present invention further defines the validity verification process of the new extreme point. In practical applications, it is determined whether the absolute value of the difference between the new extreme point and the second extreme point in the extreme point container exceeds the detection threshold to determine the validity of the new extreme point. This process ensures that reliable extreme points can be screened out, providing reliable data support for subsequent vibration index calculation.

[0053] In combination with the first aspect, in some implementation manners of the first aspect, in step S4, the width of the variable time window is adjusted in real time according to the distribution of valid extreme points, including: If the distribution span of the valid extreme points participating in the vibration index calculation increases, the width of the variable time window increases to avoid sudden changes in the vibration index; If the distribution span of the valid extreme points participating in the vibration index calculation decreases, the width of the variable time window decreases to improve the real-time performance of vibration detection.

[0054] The embodiment of the present invention ensures that the calculation of the vibration index can adapt to changes in different vibration frequencies and amplitudes through the dynamic adjustment of the variable time window, thereby avoiding sudden changes and distortion of the vibration index caused by a fixed time window. The width adjustment process of the variable time window depends on the distribution of valid extreme points, ensuring that the system can respond to vibration changes in real time.

[0055] In combination with the first aspect, in some implementation manners of the first aspect, in step S4, the vibration index is calculated using valid extreme points, including: In each sampling period, an index variable is determined according to the number of valid extreme points. The calculation formula of the index variable m is: , where is the number of valid extreme points contained in the extreme point container , and Count is the maximum number of valid extreme points preset to participate in the vibration index calculation; Based on the index variable and the time interval of valid extreme points, the vibration index of the current sampling period is calculated. The calculation formula of the vibration index V k is: , where Δt is the maximum time interval threshold between preset effective extreme points. is the signal value of the i-th effective extreme point in the extreme point container. is the signal value of the (i + 1)-th effective extreme point in the extreme point container. is the time point of the i-th effective extreme point in the extreme point container. is the time point of the (i + 1)-th effective extreme point in the extreme point container. is the center point of the fixed time window of the time point. is the time point of the first effective extreme point in the extreme point container.

[0056] In the embodiment of the present invention, through the dynamic adjustment of the variable time window, it is ensured that the calculation of the vibration index can adapt to the changes in different vibration frequencies and amplitudes, thereby avoiding the mutation and distortion of the vibration index caused by the fixed time window. The calculation process of the vibration index depends on the number and time interval of effective extreme points, ensuring that the system can respond to vibration changes in real time, thereby improving the accuracy and robustness of the vibration index.

[0057] Combined with the first aspect, in some implementation manners of the first aspect, in step S5, a second-order low-pass filter is used to filter the vibration index, including: In each sampling period, the vibration index is input into the second-order low-pass filter to obtain a smoothed vibration index. The transfer function G(s) of the second-order low-pass filter is: , where s is the complex frequency of the Laplace transform.

[0058] In the embodiment of the present invention, through the filtering process, the high-frequency fluctuations in the vibration index are eliminated, making it more continuous and smooth, thereby providing a reliable quantitative indication for the vibration suppression strategy. The parameter setting of the second-order low-pass filter directly affects the smoothness and practicality of the vibration index, ensuring that the system can respond to vibration changes in real time.

[0059] In practical applications, the system filters the vibration index by setting the parameters of the second-order low-pass filter. The filtered vibration index is more continuous and smooth, and can more intuitively reflect the instability degree of the system. This process ensures that the system can respond to vibration changes in real time, thereby improving the accuracy and robustness of the vibration index.

[0060] Embodiment 2

[0061] As Figure 2 shown, in the second aspect, the present invention provides a robot time-domain vibration detection device for physical human-robot interaction. The detection device adopts the detection method provided in any of the above embodiments, including: The signal acquisition unit 10 includes a six-axis force sensor deployed at the end effector of the robot, and is used to collect interaction force signals in real time; The data caching unit 20 is composed of a dual-port memory, and is used to store interaction force signals and form data blocks with a fixed time window; The extreme point processing unit 30 includes a digital signal processor, and is configured to perform the following operations: identify maximum and minimum points in the data block, store the extreme points corresponding to the center point of the fixed time window into the extreme point container, and perform validity checks on the extreme points; The vibration index calculation unit 40 is composed of a variable time window controller and an arithmetic logic unit, and is used to calculate the vibration index based on valid extreme points; The filtering output unit 50 includes a second-order low-pass filter, and is used to smooth the vibration index and output it.

[0062] The present invention solves the deficiencies of the prior art in terms of accuracy, robustness, real-time performance, and user-friendliness, can effectively quantify the instability degree of the physical human-machine interaction system, provide a reliable quantitative indication for the vibration suppression strategy, and ensure that the system stably and safely completes the collaborative task.

[0063] The present invention also provides an electronic device, and the electronic device includes: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is used to execute the detection method provided by any one of the above embodiments.

[0064] The present invention also provides a computer-readable storage medium, and the storage medium stores a computer program, and the computer program is used to execute the detection method provided by any one of the above embodiments.

[0065] The computer-readable storage medium can adopt any combination of one or more readable media. The readable media can be a readable signal medium or a readable storage medium. The readable storage medium can, for example, include but is not limited to an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0066] 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; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A robot time-domain vibration detection method for physical human-robot interaction, characterized in that, Including: Step S1: Obtain the interactive force signal data block. Online encapsulate the interactive force signal to be detected through a preset fixed time window to form a data block for subsequent processing. The width of the fixed time window is set according to the signal noise characteristics, specifically including: in each sampling period, fill the latest detected interactive force signal point one by one into the rightmost side of the fixed time window with a constant width of N sampling periods, and delete the detected signal point on the leftmost side of the fixed time window . Finally, obtain the data block within the fixed time window , where t k represents the time point of the k-th sampling period, y k represents the interactive force signal value of the k-th sampling period, N is the preset number of sampling periods, T s is the sampling period, and T N is the width of the fixed time window ; Step S2: Identify and register the extreme points in the data block, where the extreme points include maximum points and minimum points, specifically including: In each sampling period, search the data block for the maximum value and the minimum value , and determine whether the center point of the fixed time window is the maximum point or the minimum point. If so, register it and save it to the beginning of the preset extreme point container; Step S3: Perform validity checks on the extreme points registered and saved in Step S2, eliminate invalid extreme points, and obtain valid extreme points; Step S4: Based on a variable time window, calculate a vibration index using the valid extreme points, and the width of the variable time window is adjusted in real time according to the distribution of the valid extreme points; Step S5: Smooth the vibration index, and filter the vibration index using a second-order low-pass filter to obtain a smoothed vibration index.

2. The detection method according to claim 1, wherein In Step S1, the width of the fixed time window is set according to the signal noise characteristics, specifically including: If the intensity of the signal noise exceeds a preset noise threshold, increase the width of the fixed time window; If the intensity of the signal noise does not exceed the preset noise threshold, decrease the width of the fixed time window.

3. The detection method according to claim 1, wherein In Step S2, identifying and registering the extreme points in the data block includes: Search for the maximum value and the minimum value in the data block, and use them as the maximum extreme point and the minimum extreme point of the data block respectively; Determine the center point of the fixed time window Whether it is the maximum point or the minimum point. If so, register it and save it to the preset extreme point container At the beginning, if the center point is neither the maximum point nor the minimum point, no operation is performed. Among them, the center point The calculation formula is , where c is the center point index of the fixed time window 4. The detection method according to claim 1, wherein If a new extreme point is registered and saved at the beginning of a preset extreme point container in Step S2, the new extreme point is recorded as the first extreme point in the preset extreme point container. In Step S3, performing validity checks on the extreme points registered and saved in Step S2 includes: If the new extreme point is the only extreme point in the extreme point container, determine the new extreme point as the valid extreme point; If there is more than one extreme point in the extreme point container, judge the validity of the new extreme point by comparing the absolute value of the difference between the new extreme point and the second extreme point in the extreme point container. Among them, the detection threshold is a preset multiple of the average absolute deviation of the interaction force signal values of all valid extreme points in the extreme point container, and the value range of the preset multiple is greater than or equal to 1.2 and less than or equal to 2.

5.

5. The detection method according to claim 4, wherein If there is more than one extreme point in the extreme point container, judge the validity of the new extreme point by comparing the absolute value of the difference between the new extreme point and the second extreme point in the extreme point container, including: When the absolute value of the difference between the new extreme point and the second extreme point in the extreme point container exceeds the detection threshold, mark the new extreme point as a valid extreme point; When the absolute value of the difference does not exceed the detection threshold, remove the new extreme point from the extreme point container.

6. The detection method according to claim 1, characterized in that, In Step S4, the width of the variable time window is adjusted in real time according to the distribution of the valid extreme points, including: If the distribution span of the valid extreme points participating in the calculation of the vibration index increases, increase the width of the variable time window; If the distribution span of the valid extreme points participating in the calculation of the vibration index decreases, decrease the width of the variable time window.

7. The detection method according to claim 1, characterized in that In Step S4, calculating the vibration index using the valid extreme points includes: In each sampling period, an index variable is determined according to the number of the valid extreme points, and the calculation formula of the index variable m is as follows: , where size is the number of the valid extreme points included in the extreme point container , Count is the maximum number of the valid extreme points preset to participate in the calculation of the vibration index; Calculate the vibration index for the current sampling period based on the index variable and the time interval of the valid extreme points, where the vibration index is V k The calculation formula is as follows: , where Δt is the maximum time interval threshold between the preset valid extreme points, is the signal value of the i-th valid extreme point in the extreme point container, is the signal value of the (i + 1)-th valid extreme point in the extreme point container, is the time point of the i-th valid extreme point in the extreme point container, is the time point of the (i + 1)-th valid extreme point in the extreme point container, is the center point of the fixed time window is the time point of the first valid extreme point in the extreme point container.

8. The detection method according to claim 1, characterized in that, In Step S5, filtering the vibration index using a second-order low-pass filter includes: In each sampling period, the vibration index is input into the second-order low-pass filter to obtain the smoothed vibration index, and the transfer function G(s) of the second-order low-pass filter is: , where s is the complex frequency of the Laplace transform.

9. A robot time-domain vibration detection device for physical human-robot interaction, characterized in that, The detection device adopts the detection method described in any one of claims 1 to 8, and includes: A signal acquisition unit, which includes a six-axis force sensor deployed at the end effector of the robot and is used to collect interaction force signals in real time; A data cache unit, which is composed of a dual-port memory and is used to store the interaction force signals and form data blocks with a fixed time window; An extreme point processing unit, which includes a digital signal processor and is configured to perform the following operations: identify maximum points and minimum points in the data block, store the extreme points corresponding to the center point of the fixed time window into an extreme point container, and perform validity tests on the extreme points; A vibration index calculation unit, which consists of a variable time window controller and an arithmetic logic unit and is used to calculate the vibration index based on valid extreme points; A filtering and output unit, which includes a second-order low-pass filter and is used to smooth the vibration index and output it.

Images