Exoskeleton robot, anomaly detection method and device thereof and medium

Abstract

The invention discloses an exoskeleton robot, an anomaly detection method and device thereof and a medium. The method comprises the following steps: detecting actual pressure data of a joint part of the exoskeleton robot in an action period; determining an abnormal detection card control line; based on the actual pressure data in the action periods, the minimum accumulated distance is solved through a dynamic time adjustment algorithm, whether the minimum accumulated distance exceeds the abnormal detection card control line or not is judged, and if the minimum accumulated distance exceeds the abnormal detection card control line for a preset number of times in the multiple action periods, an abnormal alarm is given. According to the invention, real-time and intelligent detection of the exoskeleton robot can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of exoskeleton robots, and in particular to an exoskeleton robot and an abnormality detection method, device and medium thereof. Background Art

[0002] An exoskeleton robot is a wearable device that is usually composed of a mechanical structure, sensors, batteries, and electronic control systems. It uses external structures to provide protection, supplementation, and support for organisms, thereby enhancing the user's ability to move, protect, and adapt to the environment. They are designed to cooperate with the joints and muscles of the human body to assist, enhance, or restore the body's ability to move. Exoskeleton robots can be used for rehabilitation therapy to help rehabilitation patients restore muscle function, and can also be used to help disabled people achieve better independence in life, such as helping stroke patients learn to walk again, or helping spinal cord injury patients enhance muscle function.

[0003] When an exoskeleton robot is abnormal, it needs to be discovered and an alarm needs to be issued in time to avoid causing harm to the human body. The force condition of the exoskeleton robot is an important indicator for measuring the equipment. In particular, the joints are very sensitive to force. However, as the action progresses, the force condition of the skeleton robot will change over time; and it is impossible to perform unified control at the same time, because there will be differences in speed and time dislocation between multiple repetitions of the same action. This makes it difficult to detect abnormalities in the exoskeleton robot. Summary of the invention

[0004] The present invention aims to solve the technical problems in the related art at least to a certain extent. To this end, the first object of the present invention is to provide an abnormality detection method for an exoskeleton robot, which can realize real-time and intelligent detection of the exoskeleton robot.

[0005] A second object of the present invention is to provide an abnormality detection device for an exoskeleton robot.

[0006] A third object of the present invention is to provide a computer-readable storage medium.

[0007] The fourth object of the present invention is to provide an exoskeleton robot.

[0008] To achieve the above object, the present invention is implemented through the following technical solutions:

[0009] An abnormality detection method for an exoskeleton robot, comprising:

[0010] Detect the actual pressure data of the joints of the exoskeleton robot during the action cycle;

[0011] Determine the abnormal detection card control line;

[0012] Based on the actual pressure data within the action cycle, the minimum cumulative distance is calculated through a dynamic time adjustment algorithm to determine whether the minimum cumulative distance exceeds the abnormal detection card control line. If the minimum cumulative distance exceeds the abnormal detection card control line for a preset number of times within multiple action cycles, an abnormal alarm is issued.

[0013] Preferably, the method further comprises determining the action cycle, and determining the action cycle comprises:

[0014] Collect pressure data when multiple actions are not triggered as noise data when the action does not occur;

[0015] Performing statistics on the noise data, and determining the upper and lower boundaries of the noise data using Chebyshev's inequality;

[0016] The upper boundary and the lower boundary of the noise data are used to perform action triggering and end judgment to determine the action triggering time and the action end time, and the action cycle is determined based on the action triggering time and the action end time.

[0017] Preferably, the use of the upper boundary and the lower boundary of the noise data to perform action triggering and end judgment to determine the action triggering time and the action end time includes:

[0018] When the pressure data is continuously within the noise boundary, if the pressure data detected three times in succession exceeds the upper boundary and / or the lower boundary of the noise data, the time of first exceeding the upper boundary and / or the lower boundary of the noise data is determined as the action trigger time;

[0019] After the action is triggered, if the pressure data detected three times in succession are located within the noise boundary, the time when the pressure data detected for the first time is located within the noise boundary is determined as the action end time.

[0020] Preferably, the determining of the abnormal detection card control line includes:

[0021] Detect multiple sets of normal pressure data of the joints of the exoskeleton robot during normal motion within the motion cycle;

[0022] Normalize multiple groups of normal pressure data;

[0023] Averaging the normalized multiple groups of normal pressure data to obtain multiple groups of standardized pressure data;

[0024] Performing dynamic time adjustment on the normalized multiple groups of normal pressure data and the multiple groups of standardized pressure data through the dynamic time adjustment algorithm, and calculating the cumulative distance of each group of normal pressure data;

[0025] The mean and standard deviation of the accumulated distances of each group of normal pressure data are calculated so as to determine the abnormal detection control line according to the mean and standard deviation.

[0026] Preferably, the step of obtaining the minimum cumulative distance through a dynamic time adjustment algorithm based on actual pressure data within the action cycle includes:

[0027] Normalize multiple groups of actual pressure data within the action cycle;

[0028] The normalized multiple sets of actual pressure data and the multiple sets of standardized pressure data are dynamically time adjusted by the dynamic time adjustment algorithm to obtain the minimum cumulative distance.

[0029] Preferably, the specific steps of the dynamic time adjustment algorithm include:

[0030] Acquire multiple sets of normalized pressure data, the multiple sets of normalized pressure data form a first array; acquire multiple sets of standardized pressure data, the multiple sets of standardized pressure data form a second array, wherein any data in the first array and any data in the second array constitute a data point;

[0031] Create a cumulative distance matrix;

[0032] The cumulative distances of all data points are determined and the cumulative distance matrix is ​​populated so that the minimum cumulative distance is determined from the cumulative distance matrix.

[0033] Preferably, determining the cumulative distance of the data points comprises:

[0034] determining the Euclidean distance of the data points;

[0035] Determine the cumulative distances of three adjacent data points to the left, below and lower left of the data point;

[0036] Determining a minimum cumulative distance from the cumulative distances of three adjacent data points of the data point;

[0037] The cumulative distance of the data point is obtained by summing the Euclidean distance of the data point and a minimum cumulative distance determined from the cumulative distances of three adjacent data points of the data point.

[0038] To achieve the above object, the second aspect of the present invention provides an abnormality detection device for an exoskeleton robot, comprising:

[0039] The pressure sensor is arranged at the joint of the exoskeleton robot and is used to detect the actual pressure data of the joint of the exoskeleton robot during the action cycle;

[0040] A central processing unit and a buzzer alarm, wherein the central processing unit is connected to the pressure sensor and the buzzer alarm respectively, and the central processing unit is used to determine the abnormal detection card control line, and based on the actual pressure data within the action cycle, the minimum cumulative distance is calculated through a dynamic time adjustment algorithm, and it is judged whether the minimum cumulative distance exceeds the abnormal detection card control line. If within multiple action cycles, it is judged that the minimum cumulative distance exceeds the abnormal detection card control line for a preset number of times, the buzzer alarm is controlled to issue an abnormal alarm.

[0041] To achieve the above-mentioned purpose, the third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the above-mentioned method for detecting anomalies of an exoskeleton robot is implemented.

[0042] To achieve the above-mentioned objectives, the fourth aspect of the present invention provides an exoskeleton robot, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned exoskeleton robot abnormality detection method is implemented.

[0043] The present invention has at least the following technical effects:

[0044] The present invention provides an abnormality detection scheme for an exoskeleton robot, which specifically includes first determining an action cycle, then detecting multiple groups of normal pressure data of the joint parts of the exoskeleton robot in the action cycle during normal action, and determining multiple groups of standardized pressure data and card control lines based on the multiple groups of normal pressure data, and then detecting the actual pressure data of the joint parts of the exoskeleton robot in the action cycle, and then performing dynamic time adjustment through a dynamic time adjustment algorithm based on the multiple groups of actual pressure data and the multiple groups of standardized pressure data, and finding an indicator, namely the minimum cumulative distance, and comparing the minimum cumulative distance with the determined card control line. If the card control line is exceeded for a preset number of times, an alarm is sounded through a buzzer alarm. Thus, the present invention can realize real-time and intelligent detection of the exoskeleton robot.

[0045] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 Flow chart of an abnormality detection method for an exoskeleton robot according to an embodiment of the present invention.

[0047] Figure 2 Schematic diagram of action triggering and end judgment according to an embodiment of the present invention.

[0048] Figure 3Schematic diagram of the structure of an abnormality detection device for an exoskeleton robot according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The present embodiment is described in detail below, and examples of the embodiment are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0050] The exoskeleton robot and its abnormality detection method, device and medium of this embodiment are described below with reference to the accompanying drawings.

[0051] Figure 1 FIG. 4 is a flow chart of an abnormality detection method for an exoskeleton robot according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0052] Step S1: Detecting actual pressure data of the joints of the exoskeleton robot during an action cycle.

[0053] Step S2: Determine the abnormal detection card control line.

[0054] Step S3: Based on the actual pressure data within the action cycle, the minimum cumulative distance is calculated through a dynamic time adjustment algorithm to determine whether the minimum cumulative distance exceeds the abnormal detection card control line. If the minimum cumulative distance exceeds the abnormal detection card control line for a preset number of times within multiple action cycles, an abnormal alarm is issued.

[0055] Specifically, a pressure sensor and a buzzer alarm can be placed at the joint of the exoskeleton robot, and the exoskeleton robot can also be provided with a central processing unit, on which an abnormality detection algorithm can be built-in. The pressure sensor can be used to detect the force condition of the joint of the exoskeleton robot, and will collect the pressure signal, that is, the actual pressure data, in real time, and convert it into an electrical signal. Then, the collected electrical signal is transmitted to the central processing unit. The central processing unit will process the collected electrical signal, use the dynamic time adjustment algorithm to calculate the index, that is, the minimum cumulative distance, and compare it with the preset card control line determined, and the mark exceeding the card control line is A. When multiple consecutive actions are marked as A within multiple action cycles, the buzzer alarm at the joint is triggered to alarm, otherwise the buzzer alarm will not be triggered to alarm.

[0056] In one embodiment of the present invention, the method further includes determining an action cycle, and determining the action cycle includes: collecting pressure data when the action is not triggered for a plurality of times as noise data when the action does not occur; performing statistics on the noise data, and using Chebyshev's inequality to determine the upper and lower boundaries of the noise data, such as Figure 2As shown; the upper and lower boundaries of the noise data are used to trigger and end the action to determine the action trigger time and the action end time, and the action cycle is determined based on the action trigger time and the action end time.

[0057] Among them, the upper boundary and lower boundary of the noise data are used to trigger and end the action to determine the action trigger time and the action end time, including: when the pressure data is continuously located within the noise boundary, if the pressure data detected three times in a row exceeds the upper boundary and / or lower boundary of the noise data, then the time when the upper boundary and / or lower boundary of the noise data is exceeded for the first time is determined as the action trigger time; after the action is triggered, if the pressure data detected three times in a row is located within the noise boundary, then the time when the pressure data detected for the first time is located within the noise boundary is determined as the action end time.

[0058] Specifically, the received values ​​of the pressure sensor when the action is not triggered for multiple times are collected. The received values ​​at this time are the noise data when the action does not occur. The noise data is statistically analyzed to calculate the average value μ and variance σ 2 Using Chebyshev's inequality, the upper bound is μ+5σ and the lower bound is μ-5σ, then 96% of the noise is within this interval, where Chebyshev's inequality is as follows:

[0059]

[0060] Among them, P represents the probability of an event occurring, x represents the data collected by the pressure sensor when the action is not triggered, and σ represents the standard deviation.

[0061] Furthermore, the pressure data collected by the pressure sensor is traversed in chronological order. When the pressure data is continuously within the noise boundary and suddenly exceeds the noise boundary three times in a row, it is considered as an action trigger. After the action is triggered, if the pressure data is within the noise boundary three times in a row, it is considered as the end of the action. Using the time of action triggering and action end, the signal value in the middle time is intercepted as the pressure sensor signal value of an action cycle.

[0062] In one embodiment of the present invention, determining an abnormality detection control line includes: detecting multiple groups of normal pressure data of the joint parts of the exoskeleton robot during normal action within an action cycle; normalizing the multiple groups of normal pressure data; averaging the multiple groups of normal pressure data after normalization to obtain multiple groups of standardized pressure data; dynamically adjusting the multiple groups of normal pressure data after normalization and the multiple groups of standardized pressure data through a dynamic time adjustment algorithm to calculate the cumulative distance of each group of normal pressure data; calculating the mean and standard deviation of the cumulative distance of each group of normal pressure data, so as to determine the abnormality detection control line based on the mean and standard deviation.

[0063] Specifically, firstly, a plurality of groups of normal pressure sensor signal data of the exoskeleton robot in the action cycle are collected, that is, a plurality of groups of normal pressure data are collected, and each group of normal pressure data is normalized separately.

[0064] P i represents the normal pressure data at the i-th moment, P max Indicates the maximum value of normal pressure data within the action cycle, P min represents the minimum value of normal pressure data within the action cycle, then the normalized normal pressure data can be expressed as

[0065] The average of multiple groups of normalized normal pressure data is calculated in time order to obtain a new pressure-time two-dimensional array as the standardized pressure data. The normalized value of each group of normal pressure data and the standardized pressure data are dynamically adjusted in time to calculate the cumulative distance of each group of normal pressure data. The cumulative distance of each group of normal pressure data is counted to obtain the average value u of the cumulative distance of each group of normal pressure data d and standard deviation σ d Then, the upper and lower boundaries are calculated as the control lines for abnormal judgment. The upper limit is u d +5σ d , the lower limit is u d -5σ d .

[0066] In the actual operation process, the minimum cumulative distance can be calculated based on the actual pressure data in the action cycle through the dynamic time adjustment algorithm, and then the minimum cumulative distance can be compared with the card control line. The method of calculating the minimum cumulative distance based on the actual pressure data in the action cycle through the dynamic time adjustment algorithm includes: normalizing multiple groups of actual pressure data in the action cycle; and dynamically adjusting the multiple groups of normalized actual pressure data and multiple groups of standardized pressure data through the dynamic time adjustment algorithm to calculate the minimum cumulative distance.

[0067] Specifically, the data of the pressure sensor when the action occurs is collected in real time. The sensor data within the action cycle is obtained according to the action trigger and action end time. The data within the action cycle collected in real time is normalized. The normalized actual pressure data and the standardized pressure data are dynamically adjusted in time, and the minimum cumulative distance of the two sets of data after dynamic time adjustment is calculated. The calculated minimum cumulative distance is compared with the preset card control line. If it exceeds the upper and lower limits of the card control, it is considered that the action is abnormal and marked as A, otherwise it is marked as G. When the A mark occurs five times in a row, the buzzer alarm is triggered to alarm for abnormality. Five times is an empirical value and can be adjusted according to actual conditions. If there are many false alarms, the number can be increased; if there are many missed alarms, the number can be reduced.

[0068] In one embodiment of the present invention, the specific steps of the dynamic time adjustment algorithm include: obtaining multiple sets of normalized pressure data, the multiple sets of normalized pressure data form a first array; obtaining multiple sets of standardized pressure data, the multiple sets of standardized pressure data form a second array, wherein any data in the first array and any data in the second array constitute a data point; creating a cumulative distance matrix; determining the cumulative distance of all data points, and filling the cumulative distance matrix so as to determine the minimum cumulative distance from the cumulative distance matrix.

[0069] Wherein, determining the cumulative distance of the data point includes: determining the Euclidean distance of the data point; determining the cumulative distances of three adjacent data points to the left, below and lower left of the data point; determining the minimum cumulative distance from the cumulative distances of the three adjacent data points of the data point; and summing the Euclidean distance of the data point and the minimum cumulative distance determined from the cumulative distances of the three adjacent data points of the data point to obtain the cumulative distance of the data point.

[0070] The purpose of dynamic time adjustment is to compress or stretch the normalized signal value in the time dimension so that the actions of different groups are aligned as much as possible to avoid misjudgment caused by the speed of the action or the displacement of time. The specific method is:

[0071] (1) Obtain multiple sets of normalized pressure data to form a first array R = [r1, r2...r m ], and then obtain multiple sets of standardized pressure data to form a second array B = [b1, b2...b n ], where r m is the normalized pressure data of the mth group, b n Normalize the pressure data for the nth group.

[0072] (2) Determine the measurement method used when calculating the distance. Here, Euclidean distance is used for measurement.

[0073] (3) Initialization of dynamic programming table: Create a cumulative distance matrix, i.e. a two-dimensional array, of size (n+1)(m+1). Initialize the first row and the first column so that the cumulative distance D(0,0) of the starting point, i.e. the data point (0,0) = 0, and set D(b i ,0) and D(0,r j ) is infinite or a very large number, indicating an initially invalid value.

[0074] (4) Filling the cumulative distance matrix: Use dynamic programming to fill the cumulative distance matrix. For each data point (b i ,r j ), calculate the cumulative distance D(b i ,r j).

[0075] (5) The method for calculating the value of the data point is the Euclidean distance d(b) of the current data point. i ,r j ) and the smallest cumulative distance value among the three adjacent points on its left, bottom, and bottom left. The specific calculation formula is:

[0076] D(b i ,r j )=d(b i ,r j )+min{D(b i-1 ,r j ), D(b i ,r j-1 ), D(b i-1 ,r j-1 )} (2)

[0077] Among them, D(b i-1 ,r j )、D(b i ,r j-1 )、D(b i-1 ,r j-1 ) are data points (b i ,r j )The cumulative distance values ​​of the three adjacent points on the left, bottom, and lower left.

[0078] (6) Calculate the minimum cumulative distance for filling. After the cumulative distance matrix is ​​filled, D(b n ,r m ) is the minimum cumulative distance between the two time series data B and R, which is the dynamic time warping distance between the two sets of time series data B and R. It can be compared with the card control line.

[0079] Figure 3 FIG. 1 is a schematic diagram of the structure of an abnormality detection device of an exoskeleton robot according to an embodiment of the present invention. Figure 3 As shown, the abnormality detection device of the exoskeleton robot includes a pressure sensor, a central processing unit and a buzzer alarm which are connected in sequence. The pressure sensor is arranged at the joint of the exoskeleton robot, and is used to detect the actual pressure data of the joint of the exoskeleton robot in the action cycle; the central processing unit is connected to the pressure sensor and the buzzer alarm respectively, and the central processing unit is used to determine the abnormality detection card control line, and based on the actual pressure data in the action cycle, the minimum cumulative distance is calculated through the dynamic time adjustment algorithm, and it is judged whether the minimum cumulative distance exceeds the abnormality detection card control line. If it is judged that the minimum cumulative distance exceeds the abnormality detection card control line for a preset number of times in multiple action cycles, the buzzer alarm is controlled to give an abnormality alarm.

[0080] It should be noted that the specific implementation of the abnormality detection device of the exoskeleton robot in this embodiment can refer to the specific implementation of the abnormality detection method of the exoskeleton robot mentioned above, and will not be repeated here to avoid redundancy.

[0081] In summary, the present invention can timely detect the abnormality of the exoskeleton robot and avoid danger; the present invention adopts a calculation method based on Chebyshev's inequality to judge the complete stage of action triggering and ending through signals; the present invention adopts a dynamic time adjustment method to evaluate the pressure signal of the action, avoiding false alarms caused by signals at different speeds and time displacements, and realizing intelligent abnormality detection of the exoskeleton robot.

[0082] Furthermore, the present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the above-mentioned abnormality detection method for the exoskeleton robot is implemented.

[0083] Furthermore, the present invention also provides an exoskeleton robot, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned exoskeleton robot abnormality detection method is implemented.

[0084] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0085] Although the content of the present invention has been described in detail through the above preferred embodiments, it should be appreciated that the above description should not be considered as a limitation of the present invention. After reading the above content, it will be apparent to those skilled in the art that various modifications and substitutions of the present invention will occur. Therefore, the protection scope of the present invention should be limited by the appended claims.

Claims

1. A method for detecting anomalies of an exoskeleton robot, characterized in that: include: Detect the actual pressure data of the joints of the exoskeleton robot during the action cycle; Determine the abnormal detection card control line; Based on the actual pressure data within the action cycle, the minimum cumulative distance is calculated through a dynamic time adjustment algorithm to determine whether the minimum cumulative distance exceeds the abnormal detection card control line. If the minimum cumulative distance exceeds the abnormal detection card control line for a preset number of times within multiple action cycles, an abnormal alarm is issued.

2. The method for detecting anomalies of an exoskeleton robot according to claim 1, wherein: The method further includes determining the action period, wherein determining the action period includes: Collect pressure data when multiple actions are not triggered as noise data when the action does not occur; Performing statistics on the noise data, and determining the upper and lower boundaries of the noise data using Chebyshev's inequality; The upper boundary and the lower boundary of the noise data are used to perform action triggering and end judgment to determine the action triggering time and the action end time, and the action cycle is determined based on the action triggering time and the action end time.

3. The abnormality detection method of the exoskeleton robot according to claim 2, characterized in that: The use of the upper boundary and the lower boundary of the noise data to perform action triggering and end judgment to determine the action triggering time and the action end time includes: When the pressure data is continuously within the noise boundary, if the pressure data detected three times in succession exceeds the upper boundary and / or the lower boundary of the noise data, the time of first exceeding the upper boundary and / or the lower boundary of the noise data is determined as the action trigger time; After the action is triggered, if the pressure data detected three times in succession are located within the noise boundary, the time when the pressure data detected for the first time is located within the noise boundary is determined as the action end time.

4. The method for detecting anomalies of an exoskeleton robot according to claim 1, wherein: The step of determining an abnormal detection card control line includes: Detect multiple sets of normal pressure data of the joints of the exoskeleton robot during normal motion within the motion cycle; Normalize multiple groups of normal pressure data; Averaging the normalized multiple groups of normal pressure data to obtain multiple groups of standardized pressure data; Performing dynamic time adjustment on the normalized multiple groups of normal pressure data and the multiple groups of standardized pressure data by using the dynamic time adjustment algorithm, and calculating the cumulative distance of each group of normal pressure data; The mean and standard deviation of the accumulated distances of each group of normal pressure data are calculated so as to determine the abnormal detection control line according to the mean and standard deviation.

5. The method for detecting anomalies of an exoskeleton robot according to claim 4, wherein: The method of calculating the minimum cumulative distance based on the actual pressure data within the action cycle through a dynamic time adjustment algorithm includes: Normalize multiple groups of actual pressure data within the action cycle; The normalized multiple sets of actual pressure data and the multiple sets of standardized pressure data are dynamically time adjusted by the dynamic time adjustment algorithm to obtain the minimum cumulative distance.

6. The method for detecting anomalies of an exoskeleton robot according to claim 5, wherein: The specific steps of the dynamic time adjustment algorithm include: Acquire multiple sets of normalized pressure data, the multiple sets of normalized pressure data form a first array; acquire multiple sets of standardized pressure data, the multiple sets of standardized pressure data form a second array, wherein any data in the first array and any data in the second array constitute a data point; Create a cumulative distance matrix; The cumulative distances of all data points are determined and the cumulative distance matrix is ​​populated so that the minimum cumulative distance is determined from the cumulative distance matrix.

7. The method for detecting anomalies of an exoskeleton robot according to claim 6, wherein: Determine the cumulative distance of data points, including: determining the Euclidean distance of the data points; Determine the cumulative distances of three adjacent data points to the left, below and lower left of the data point; Determining a minimum cumulative distance from the cumulative distances of three adjacent data points of the data point; The cumulative distance of the data point is obtained by summing the Euclidean distance of the data point and a minimum cumulative distance determined from the cumulative distances of three adjacent data points of the data point.

8. An abnormality detection device for an exoskeleton robot, characterized in that: include: The pressure sensor is arranged at the joint of the exoskeleton robot and is used to detect the actual pressure data of the joint of the exoskeleton robot during the action cycle; A central processing unit and a buzzer alarm, wherein the central processing unit is connected to the pressure sensor and the buzzer alarm respectively, and the central processing unit is used to determine the abnormal detection card control line, and based on the actual pressure data within the action cycle, the minimum cumulative distance is calculated through a dynamic time adjustment algorithm, and it is judged whether the minimum cumulative distance exceeds the abnormal detection card control line. If within multiple action cycles, it is judged that the minimum cumulative distance exceeds the abnormal detection card control line for a preset number of times, the buzzer alarm is controlled to issue an abnormal alarm.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the abnormality detection method for the exoskeleton robot as described in any one of claims 1 to 7 is implemented.

10. An exoskeleton robot, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the abnormality detection method for the exoskeleton robot as described in any one of claims 1-7.

Images