Strain clamp detection control method and device of intelligent robot

Abstract

The invention relates to the technical field of robot application, and particularly discloses a strain clamp detection control method and device for an intelligent robot, and the method comprises the steps: obtaining the installation position of a strain clamp, and carrying out the clustering of the strain clamp according to the installation position, and obtaining a detection region; performing enumeration type arrangement on the detection areas, taking each arrangement scheme as a detection scheme, and synchronously generating a number sequence of each detection scheme; acquiring a motion path of each detection scheme, calculating energy consumption, and selecting an optimal scheme according to the energy consumption and the number sequence; receiving detection data fed back by the robot, generating a control instruction based on the optimal scheme and the detection data fed back by the robot, and sending the control instruction to the robot; the detection data is an image containing shooting time; according to the invention, the path planning process and the data uploading process of the robot are optimized, the energy consumption of the robot is greatly reduced, the endurance is improved, and the strain clamp in the actual use stage can be detected more quickly.

Description

Technical Field

[0001] The present invention relates to the technical field of robot applications, and specifically to a detection control method and device for strain clamps of intelligent robots. Background Art

[0002] A strain clamp is a kind of electrical hardware used in overhead lines, mainly used to fix conductors and bear tension to ensure the stability and reliability of power lines. It is commonly used at the terminals, corners, crossings, and branch positions of power transmission and distribution lines.

[0003] In addition to being detected during the production stage, strain clamps also need to be detected during the actual use stage. Since their installation positions are all on high-voltage lines, the detection process during the actual use stage is extremely dangerous. However, due to the progress of robot technology, the danger level of the detection process during the actual use stage has been greatly reduced. Therefore, in the prior art, robots can be used to detect strain clamps. When using a robot to detect a strain clamp, the battery life problem is the biggest issue. If the battery life can be improved, then the extra power can be used to increase the speed, thereby increasing the detection speed and performing one or more additional detection cycles. Summary of the Invention

[0004] The purpose of the present invention is to provide a detection control method and device for strain clamps of intelligent robots to solve the problems raised in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solutions: A detection control method for strain clamps of an intelligent robot, the method comprising: Obtaining the installation positions of the strain clamps, clustering the strain clamps according to the installation positions to obtain detection areas; the clustering process is a three-dimensional clustering process; Enumeratively arranging the detection areas, taking each arrangement scheme as a detection scheme, and synchronously generating a quantity sequence for each detection scheme; the dimension of the quantity sequence is the same as the total number of detection areas, and each element value in the quantity sequence represents the number of strain clamps in the corresponding detection area in the arrangement scheme; Obtaining the movement path of each detection scheme, calculating the energy consumption, and selecting the optimal scheme according to the energy consumption and the quantity sequence; wherein, the energy consumption is the total energy consumption of the robot working along the movement path; Receiving the detection data fed back by the robot, generating a control instruction based on the optimal scheme and the detection data fed back by the robot, and sending it to the robot; the detection data is an image containing the shooting time.

[0006] As a further scheme of the present invention: the step of obtaining the installation positions of the strain clamps, clustering the strain clamps according to the installation positions to obtain detection areas comprises: Query the installation location of the strain clamp in the filing database; Calculate the spatial distance between any two strain clamps; Query the power parameters of the lines where any two strain clamps are located, and adjust the spatial distance according to the differences in the power parameters; Cluster the strain clamps according to the adjusted spatial distance to construct a detection area.

[0007] As a further solution of the present invention: the steps of obtaining the movement path of each detection scheme, calculating the energy consumption, and selecting the optimal scheme according to the energy consumption and the quantity sequence include: Obtain the movement path of each detection scheme based on a preset path planning algorithm; Obtain the total height change and the total path length of the movement path, and calculate the energy consumption according to the total height change and the total path length; the energy consumption is proportional to the total height change, and the energy consumption is proportional to the total path length; Read the quantity sequence of the detection scheme, calculate the inverse ordinal number of the quantity sequence, and calculate the final score of the detection scheme according to the energy consumption and the inverse ordinal number; Select the detection scheme with the highest final score as the optimal scheme; The calculation process of the final score is: ; where, represents the final score, and are preset constants, both of which are positive values, and are preset correction coefficients, both of which are positive values, is the total number of coordinate points in the movement path, represents the eigenvalue of the th coordinate point in the movement path, ; represents the absolute value of the height difference corresponding to the th coordinate point in the movement path, takes the value: when , ; when , ; represents the height corresponding to the th coordinate point, represents the height corresponding to the previous coordinate point of the th coordinate point; represents the total path length; represents the inverse ordinal number.

[0008] As a further solution of the present invention, the steps of generating a control instruction based on the optimal solution, sending it to the robot, and receiving the detection data fed back by the robot include: Send the optimal solution to the robot, and obtain the position of the robot according to a preset time period; after receiving the optimal solution, the robot locally adjusts the movement direction based on the optimal solution; Compare the position with the optimal solution, generate a higher-level control instruction, and send it to the robot; the time period is a dynamic period and is related to the comparison result of the position and the optimal solution; Receive the detection data fed back by the robot, identify the detection data, and generate a speed control instruction for the robot according to the identification result, and send it to the robot.

[0009] As a further solution of the present invention, the steps of receiving the detection data fed back by the robot, generating a control instruction based on the optimal solution and the detection data fed back by the robot, and sending it to the robot include: Receive the detection data fed back by the robot, and calculate the feedback frequency according to the shooting time of the detection data; Perform a higher-level identification on the detection data. When the higher-level identification result is that there is no abnormality, determine the speed adjustment coefficient according to the feedback frequency, and use it as the speed control instruction to send to the robot; When the higher-level identification result is that there is an abnormality, query the position of the robot at the shooting time, use the robot position as the target point to generate a higher-level control instruction, and send it to the robot; Receive the images fed back by the robot at the passing points. When the number of images reaches a preset number threshold, generate a continue detection instruction and send it to the robot; Among them, during the movement of the robot, perform local identification on the detection data, and determine whether to feedback images according to the local identification result; the local identification adopts a convolution identification scheme, and the convolution kernel adopts the convolution features of the image of the strain clamp under a preset standard state; the higher-level identification adopts a neural network model, the feature of the sample of the neural network model is the image of the strain clamp, and the label of the sample is the abnormal type; the order of magnitude of the number of samples is not less than the preset order-of-magnitude threshold.

[0010] As a further solution of the present invention, the method further includes: On the robot side, obtain the battery life in real time. When the battery life is less than the preset battery life threshold, query other robots within a preset distance range, insert its own identifier into the feedback result, and then feedback the feedback result to other robots; when the identifier of the feedback result received by the master end is different from the sender, generate a warning message pointing to the robot corresponding to the identifier.

[0011] The technical solution of the present invention also provides a strain clamp detection control device for an intelligent robot, and the device includes: The detection area creation module is used to obtain the installation positions of strain clamps, cluster the strain clamps according to the installation positions to obtain detection areas; the clustering process is a three-dimensional clustering process; The scheme enumeration module is used to perform an enumerative arrangement on the detection areas, regard each arrangement scheme as a detection scheme, and synchronously generate a quantity sequence for each detection scheme; the dimension of the quantity sequence is the same as the total number of detection areas, and each element value in the quantity sequence represents the number of clamps in the detection area corresponding to the position in the arrangement scheme; The scheme selection module is used to obtain the movement path of each detection scheme, calculate the energy consumption, and select the optimal scheme according to the energy consumption and the quantity sequence; wherein, the energy consumption is the total energy consumption of the robot working along the movement path; The instruction generation and sending module is used to receive the detection data fed back by the robot, generate a control instruction based on the optimal scheme and the detection data fed back by the robot, and send it to the robot; the detection data is an image containing the shooting time.

[0012] As a further scheme of the present invention: the detection area creation module includes: The position query unit is used to query the installation positions of strain clamps in the record database; The distance calculation unit is used to calculate the spatial distance between any two strain clamps; The distance update unit is used to query the power parameters of the lines where any two strain clamps are located, and adjust the spatial distance according to the difference in power parameters; The clustering execution unit is used to cluster the strain clamps according to the adjusted spatial distance to construct detection areas.

[0013] As a further scheme of the present invention: the scheme selection module includes: The path determination unit is used to obtain the movement path of each detection scheme based on a preset path planning algorithm; The energy consumption calculation unit is used to obtain the total height change amount and the total path length of the movement path, and calculate the energy consumption according to the total height change amount and the total path length; the energy consumption is directly proportional to the total height change amount, and the energy consumption is directly proportional to the total path length; The score calculation unit is used to read the quantity sequence of the detection scheme, calculate the inversion number of the quantity sequence, and calculate the final score of the detection scheme according to the energy consumption and the inversion number; The score application unit is used to select the detection scheme with the highest final score as the optimal scheme; The calculation process of the final score is: ; in the formula, represents the final score, and are preset constants, both of which are positive values, and are preset correction coefficients, both of which are positive values, is the total number of coordinate points in the motion path, represents the th eigenvalue of the coordinate point in the motion path, ; represents the absolute value of the height difference corresponding to the th coordinate point in the motion path, takes values as: when , ; when , ; represents the height corresponding to the th coordinate point, represents the height corresponding to the previous coordinate point of the th coordinate point; represents the total path length; represents the inversion number.

[0014] As a further solution of the present invention: the instruction generation and sending module includes: A scheme sending unit, configured to send the optimal scheme to the robot and obtain the position of the robot according to a preset time period; after receiving the optimal scheme, the robot locally adjusts the motion direction based on the optimal scheme; A comparison and adjustment unit, configured to compare the position with the optimal scheme, generate a superior control instruction, and send it to the robot; the time period is a dynamic period and is related to the comparison result of the position and the optimal scheme; A speed adjustment unit, configured to receive the detection data fed back by the robot, identify the detection data, and generate a speed control instruction for the robot according to the identification result and send it to the robot.

[0015] Compared with the prior art, the beneficial effects of the present invention are: the present invention optimizes the robot path planning process and data uploading process, greatly reduces the energy consumption of the robot, improves the battery life, and enables it to detect the strain clamp in the actual use stage faster. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0017] Figure 1 It is a flowchart of the detection control method for the strain clamp of an intelligent robot.

[0018] Figure 2It is a block diagram of the composition structure of the strain clamp detection and control device for an intelligent robot. Specific implementation manners

[0019] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0020] Figure 1 It is a flowchart of the strain clamp detection and control method for an intelligent robot. In an embodiment of the present invention, a strain clamp detection and control method for an intelligent robot, the method includes: Step S100: Obtain the installation position of the strain clamp, cluster the strain clamps according to the installation position to obtain a detection area; the clustering process is a three-dimensional clustering process; The strain clamp belongs to the infrastructure of the power line, and its installation position is pre-filed. By obtaining the installation position of the strain clamp and clustering the strain clamps according to the installation position, similar strain clamps can be grouped into one category, and then a detection area can be constructed. The purpose of this process is that the detection process in this application is an image-based detection process. Taking one image may include multiple strain clamps. Therefore, by clustering the strain clamps and then performing shooting detection, the number of images taken can be reduced to a certain extent. In this application, when the number of images taken becomes smaller, the battery life of the robot will become longer.

[0021] It should be noted that since the strain clamp is installed in the air, its installation position is a three-dimensional coordinate, and the clustering process is also a clustering process in three-dimensional space.

[0022] Step S200: Enumerate and arrange the detection areas, take each arrangement scheme as a detection scheme, and synchronously generate a quantity sequence for each detection scheme; the dimension of the quantity sequence is the same as the total number of detection areas, and each element value in the quantity sequence represents the number of clamps in the detection area corresponding to the position in the arrangement scheme; For an intelligent robot, the number of detection areas within its inspection range is limited. By enumerating and arranging the detection areas, multiple detection schemes can be obtained. Generally speaking, a detection scheme is the order in which the detection areas are detected. Each arrangement scheme is a detection scheme. The arrangement scheme and the detection scheme are two names for the same content. At the same time, for each detection scheme, it is an ordered arrangement of detection areas. Query the number of clamps in each detection area, and arrange the number of clamps in the order of the detection areas to obtain a quantity sequence. From the above content, it can be seen that a detection scheme also corresponds to a quantity sequence.

[0023] Step S300: Obtain the motion path of each detection scheme, calculate the energy consumption, and select the optimal scheme according to the energy consumption and the quantity sequence; wherein, the energy consumption is the total energy consumption of the robot working along the motion path. Furthermore, obtaining the motion path of each detection scheme is not a complicated process. It is the path passing through each detection area, and the existing navigation service can meet this requirement. On this basis, calculate the energy consumption of each detection scheme. Combining these two parameters, namely the energy consumption and the quantity sequence, can evaluate the detection scheme, and finally select the optimal scheme from multiple detection schemes; wherein, the energy consumption is the total energy consumption of the robot working along the motion path. In addition, the "optimal" in the optimal scheme of this application refers to the optimal under certain conditions. The optimal under different conditions is different, and it is not an absolute optimal either, but a relative optimal that meets certain conditions.

[0024] Step S400: Receive the detection data fed back by the robot, generate a control instruction based on the optimal scheme and the detection data fed back by the robot, and send it to the robot; the detection data is an image containing the shooting time. During the movement of the robot, it will obtain detection data. There are two architectures for the obtaining process. One is to pre-determine the shooting points according to the positions of the strain clamps in the detection area. When the robot reaches these points, activate the camera to take pictures. The other is to turn on the camera in real time, first record the video under low-definition conditions, and detect in real time whether there are strain clamps. If there are, then improve the accuracy. Although this method has a slightly increased energy consumption, it can record the working process of the robot, and when special situations occur, the staff can conduct a retrospective analysis afterwards.

[0025] After receiving the detection data fed back by the robot, generate a control instruction jointly based on the optimal scheme and the detection data fed back by the robot, and send it to the robot to control the robot to complete the detection work of the strain clamps.

[0026] Regarding step S100, the steps of obtaining the installation position of the strain clamps and clustering the strain clamps according to the installation position to obtain the detection area include: Query the installation position of the strain clamps in the filing database. Calculate the spatial distance between any two strain clamps. Query the power parameters of the lines where any two strain clamps are located, and adjust the spatial distance according to the differences in the power parameters. Cluster the strain clamps according to the adjusted spatial distance to construct the detection area.

[0027] The above content describes the creation process of the detection area. Query the installation positions of strain clamps in the record database, calculate the spatial distance between any two strain clamps. At the same time, this application will also query the electrical parameters of the lines where any two strain clamps are located. The electrical parameters are a superordinate concept, and some easily measurable electrical parameters can be selected, such as DC resistance, inductance, line impedance, DC resistance, AC resistance, inductance, and capacitance, etc. One or several of these parameters can also be selected, which is specifically determined by the staff according to the situation; in the technical solution of the present invention, calculate the difference in electrical parameters of the lines where two strain clamps are located, and then adjust the spatial distance. The greater the difference, the greater the adjusted spatial distance; finally, cluster the strain clamps according to the adjusted spatial distance. The clustering scheme can adopt the DBSCAN clustering scheme, etc.; in addition, there is still one step from the clustering process to the construction of the detection area. The clustering process classifies the strain clamps into one category and creates an area containing the same type of strain clamps, which is the detection area; the simplest detection area construction process is to create a circular area containing the same type of strain clamps, and the diameter is A times the maximum distance of the strain clamps in the same type of strain clamps, where A is greater than one, to ensure that the circular area can cover all the same type of strain clamps.

[0028] Regarding step S300, the steps of obtaining the movement path of each detection scheme, calculating the energy consumption, and selecting the optimal scheme according to the energy consumption and the quantity sequence include: Obtain the movement path of each detection scheme based on a preset path planning algorithm; Obtain the total height change and the total path length of the movement path, and calculate the energy consumption according to the total height change and the total path length; the energy consumption is proportional to the total height change, and the energy consumption is proportional to the total path length; Read the quantity sequence of the detection scheme, calculate the inverse number of the quantity sequence, and calculate the final score of the detection scheme according to the energy consumption and the inverse number; Select the detection scheme with the highest final score as the optimal scheme; In an example of the technical solution of the present invention, the selection process of all enumerated detection schemes is described. For any detection scheme, the detection scheme is the arrangement of detection areas. By using the existing path planning algorithm to obtain the path passing through each detection area, the detection scheme is converted into a movement path; for the movement path of each detection scheme, obtain the total height change and the total path length of the movement path, calculate the energy consumption according to the total height change and the total path length. At the same time, read the quantity sequence of the detection scheme, calculate the inverse number of the quantity sequence, calculate the final score of the detection scheme according to the energy consumption and the inverse number, and select the detection scheme with the highest final score as the optimal scheme.

[0029] The calculation process of the final score is: ; where Indicates the final rating. and are preset constants, all of which are positive values. and is the preset correction coefficient, which is a positive value. is the total number of coordinate points in the motion path, Indicates the first The characteristic value of the coordinate point, ; Indicates the first The absolute value of the height difference corresponding to the coordinate points, The value of is: hour, ;when , ; Indicates The height corresponding to the coordinate point, Indicates The height corresponding to the previous coordinate point of the coordinate point; represents the total length of the path; Indicates a reverse number.

[0030] The final score is inversely proportional to the total height change and the total path length. The total height change is the total number of times the height change sign changes during the motion path, corresponding to The sign change of the height change indicates how many times the robot has switched uphill and downhill. The uphill and downhill switching indicates that the robot needs to adjust the mode, which brings a certain amount of energy consumption. Therefore, the greater the total height change, the higher the energy consumption and the lower the score.

[0031] Correspondingly, the longer the total path length is, the higher the energy consumption is, and the lower the score is; further, the reverse number represents the quantity characteristics of the tension clamps. The meaning of the scheme with the largest reverse number is that the arrangement order of the detection areas is the descending order of the number of tension clamps, and the number decreases successively. In the technical scheme of the present invention, it is believed that the more the number of tension clamps is, the more important the corresponding detection area is, and the more it needs to be detected first. That is, the larger the reverse number is, the more the detection areas with more numbers are closer to the front and the higher the score is.

[0032] Regarding step S400, the steps of generating control instructions based on the optimal solution, sending the instructions to the robot, and receiving the detection data fed back by the robot include: The optimal solution is sent to the robot, and the position of the robot is obtained according to a preset time period; after the robot receives the optimal solution, the robot adjusts the movement direction locally based on the optimal solution; The position is compared with the optimal solution, and a higher-level control instruction is generated and sent to the robot; the time period is a dynamic period, which is related to the comparison result between the position and the optimal solution.

[0033] In an example of the technical solution of the present invention, the recognition and application process of the detection data and the control process of the robot are defined. The optimal solution is sent to the robot. After receiving the optimal solution, the robot locally adjusts its movement direction based on the optimal solution. At the same time, the master terminal obtains the position of the robot according to a preset time period, that is, obtains the position of the robot every once in a while. This requires the use of a locator built into the robot. The locator belongs to the basic equipment of the robot. The master terminal compares the position with the optimal solution to determine whether there is an offset, and then generates a superior control instruction to adjust the local control process of the robot; in this process, the robot does not need to upload its position in real time, reducing energy consumption and improving battery life.

[0034] In addition, the time period adopts a dynamic period, which is related to the comparison result of the position and the optimal solution. Specifically, it is related to the generation interval of the superior control instruction. The smaller the generation interval, the smaller the dynamic period.

[0035] Furthermore, the master terminal also needs to receive the detection data fed back by the robot, recognize the detection data, and generate a speed control instruction for the robot according to the recognition result and send it to the robot.

[0036] The steps of receiving the detection data fed back by the robot, generating a control instruction based on the optimal solution and the detection data fed back by the robot, and sending it to the robot include: Receiving the detection data fed back by the robot and calculating the feedback frequency according to the shooting time of the detection data; Performing superior recognition on the detection data. When the superior recognition result is that there is no abnormality, determining a speed adjustment coefficient according to the feedback frequency as the speed control instruction and sending it to the robot; When the superior recognition result is that there is an abnormality, querying the position of the robot at the shooting time, generating a superior control instruction with the robot position as the target point, and sending it to the robot; Receiving the image fed back by the robot at the passing point. When the number of images reaches a preset number threshold, generating a continue detection instruction and sending it to the robot.

[0037] Specifically, during the movement of the robot, local recognition is performed on the detection data, and it is determined whether to feedback an image based on the local recognition result; the local recognition adopts a convolutional recognition scheme, and the convolutional kernel adopts the convolutional features of the image of the strain clamp under a preset standard state; when the robot captures an image, convolutional recognition is performed on the image. The convolutional recognition method is to traverse the image according to the preset convolutional kernel. When the convolutional kernel successfully matches a certain area in the image, it is considered that the strain clamp is in the standard state. It should be noted that only when the strain clamp is completely in the standard state will the robot determine it as standard through the image. If there are some occlusions or other unclear phenomena, it will be determined as non-standard, and the image (detection data) will be uploaded to the master terminal.

[0038] Receive the detection data fed back by the robot, calculate the feedback frequency according to the shooting time of the detection data. The shooting time is a label of the detection data, which is generated and inserted into the image when the robot acquires the image. The master terminal performs upper-level recognition on the detection data. When the upper-level recognition result shows no abnormality, determine the speed adjustment coefficient according to the feedback frequency as the speed control instruction and send it to the robot. The upper-level recognition adopts a neural network model. The features of the samples of the neural network model are the images of the strain clamp, and the labels of the samples are the types of abnormalities; the order of magnitude of the number of samples is not less than the preset magnitude threshold, and the recognition depth and recognition accuracy of the neural network model are higher.

[0039] In addition, when the upper-level recognition result is recognized and shows no abnormality, the speed adjustment coefficient also needs to be determined according to the feedback frequency. The speed adjustment coefficient is inversely proportional to the feedback frequency. The lower the feedback frequency, the more images that the robot locally recognizes as completely standard. At this time, the robot does not need to perform multiple interaction processes with the master terminal, has lower energy consumption and longer battery life, can increase the speed, and complete the detection task faster, that is, the speed adjustment coefficient is larger.

[0040] Furthermore, when the upper-level recognition result shows an abnormality, query the position of the robot at the shooting time to determine which position is abnormal. At this time, use the robot position as the target point to generate an upper-level control instruction and send it to the robot to control the robot to acquire more images; when the number of images is sufficient, generate a continue detection instruction and send it to the robot. The continue detection instruction is used to control the robot to move according to the original rules.

[0041] As a preferred embodiment of the technical solution of the present invention, the method further includes: On the robot side, the battery life is obtained in real time. When the battery life is less than the preset battery life threshold, query other robots within the preset distance range, insert its own identifier into the feedback result, and then feedback the feedback result to other robots; when the identifier of the feedback result received by the master terminal is different from the sender, generate a warning message pointing to the robot corresponding to the identifier.

[0042] In an example of the technical solution of the present invention, an emergency plan is introduced. When the battery life of a certain robot is low, its detection data can be sent to the nearest robot, and the nearest robot will upload the data on its behalf (to the main terminal) without uploading the data by itself. Sending the data to the nearest robot can utilize a short-range connection channel such as Bluetooth, with lower energy consumption. When the identifier of the feedback result received by the main terminal is different from the sender, a warning message pointing to the robot corresponding to the identifier is generated to inform the administrator of the problematic robot.

[0043] It is worth mentioning that the above solution is an emergency plan. The probability of a robot having a problem is very low, and the probability of a certain robot and its receiving robot having problems simultaneously is even lower. However, this probability still exists. For this, the image receiving port of the problematic machine is closed so that it cannot receive the images forwarded by other robots. When other robots fail to forward, they will select another robot as the receiver.

[0044] Figure 2 It is a block diagram of the composition structure of the detection and control device for the strain clamp of an intelligent robot. In an embodiment of the present invention, a detection and control device for the strain clamp of an intelligent robot, the device 10 includes: A detection area creation module 11, configured to obtain the installation position of the strain clamp, cluster the strain clamps according to the installation position to obtain a detection area; the clustering process is a three-dimensional clustering process; A scheme enumeration module 12, configured to perform an enumerative arrangement on the detection area, regard each arrangement scheme as a detection scheme, and synchronously generate a quantity sequence for each detection scheme; the dimension of the quantity sequence is the same as the total number of detection areas, and each element value in the quantity sequence represents the number of clamps in the detection area at the corresponding position in the arrangement scheme; A scheme selection module 13, configured to obtain the movement path of each detection scheme, calculate the energy consumption, and select the optimal scheme according to the energy consumption and the quantity sequence; wherein, the energy consumption is the total energy consumption of the robot working along the movement path; An instruction generation and sending module 14, configured to receive the detection data fed back by the robot, generate a control instruction based on the optimal scheme and the detection data fed back by the robot, and send it to the robot; the detection data is an image containing the shooting time.

[0045] Further, the detection area creation module 11 includes: A position query unit, configured to query the installation position of the strain clamp in the record database; A distance calculation unit, configured to calculate the spatial distance between any two strain clamps; A distance update unit, configured to query the electrical parameters of the lines where any two strain clamps are located, and adjust the spatial distance according to the difference in electrical parameters; The clustering execution unit is used to cluster the strain clamps according to the adjusted spatial distance and construct a detection area.

[0046] Specifically, the scheme selection module 13 includes: The path determination unit is used to obtain the motion path of each detection scheme based on a preset path planning algorithm; The energy consumption calculation unit is used to obtain the total height change and the total path length of the motion path, and calculate the energy consumption according to the total height change and the total path length; the energy consumption is proportional to the total height change, and the energy consumption is proportional to the total path length; The score calculation unit is used to read the number sequence of the detection scheme, calculate the inversion number of the number sequence, and calculate the final score of the detection scheme according to the energy consumption and the inversion number; The score application unit is used to select the detection scheme with the highest final score as the optimal scheme; The calculation process of the final score is as follows: ; where represents the final score, and are preset constants, both of which are positive values, and are preset correction factors, both of which are positive values, is the total number of coordinate points in the motion path, represents the th coordinate point in the motion path, ; represents the absolute value of the height difference corresponding to the th coordinate point in the motion path, The value of is: when ; when , ; represents the height corresponding to the th coordinate point, represents the height corresponding to the previous coordinate point of the th coordinate point; represents the total path length; represents the inversion number.

[0047] Furthermore, the instruction generation and sending module 14 includes: The scheme sending unit is used to send the optimal scheme to the robot and obtain the position of the robot according to a preset time period; after receiving the optimal scheme, the robot locally adjusts its motion direction based on the optimal scheme; The comparison and adjustment unit is used to compare the position with the optimal solution, generate a superior control instruction, and send it to the robot; the time period is a dynamic period and is related to the comparison result of the position and the optimal solution. The speed adjustment unit is used to receive the detection data fed back by the robot, identify the detection data, generate a speed control instruction for the robot according to the identification result, and send it to the robot.

[0048] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for detecting and controlling a tension clamp of an intelligent robot, characterized in that: The method comprises: Acquire the installation position of the tension clamp, cluster the tension clamp according to the installation position, and obtain the detection area; the clustering process is a three-dimensional clustering process; The detection areas are arranged enumerably, each arrangement scheme is regarded as a detection scheme, and a quantity sequence of each detection scheme is generated synchronously; the dimension of the quantity sequence is the same as the total number of detection areas, and each element value in the quantity sequence represents the number of wire clamps in the detection area at the corresponding position in the arrangement scheme; Obtain the motion path of each detection scheme, calculate the energy consumption, and select the optimal scheme according to the energy consumption and quantity sequence; wherein the energy consumption is the total energy consumption of the robot working along the motion path; Receive the detection data fed back by the robot, generate control instructions based on the optimal solution and the detection data fed back by the robot, and send them to the robot; the detection data is an image containing the shooting time.

2. The method for detecting and controlling the tension clamp of an intelligent robot according to claim 1, characterized in that: The step of obtaining the installation position of the tension clamp, clustering the tension clamps according to the installation position, and obtaining the detection area comprises: Query the installation location of the tension clamp in the record database; Calculate the spatial distance between any two tension clamps; Query the power parameters of the lines where any two tension clamps are located, and adjust the spatial distance according to the difference in power parameters; The tension clamps are clustered according to the adjusted spatial distance and the detection area is constructed.

3. The method for detecting and controlling the tension clamp of an intelligent robot according to claim 1, characterized in that: The steps of obtaining the motion path of each detection scheme, calculating the energy consumption, and selecting the optimal scheme according to the energy consumption and the quantity sequence include: Obtain the motion path of each detection scheme based on a preset path planning algorithm; Obtaining a total amount of height change and a total length of the motion path, and calculating energy consumption according to the total amount of height change and the total length of the path; the energy consumption is proportional to the total amount of height change, and the energy consumption is proportional to the total length of the path; Read the quantity sequence of the detection scheme, calculate the reverse number of the quantity sequence, and calculate the final score of the detection scheme according to the energy consumption and the reverse number; Select the detection scheme with the highest final score as the optimal scheme; The calculation process of the final score is: ; In the formula, Indicates the final rating. and are preset constants, all of which are positive values. and is the preset correction coefficient, which is a positive value. is the total number of coordinate points in the motion path, Indicates the first The characteristic value of the coordinate point, ; Indicates the first The absolute value of the height difference corresponding to the coordinate points, The value of is: hour, ;when , ; Indicates The height corresponding to the coordinate point, Indicates The height corresponding to the previous coordinate point of the coordinate point; represents the total length of the path; Indicates a reverse number.

4. The method for detecting and controlling the tension clamp of an intelligent robot according to claim 1, characterized in that: The steps of generating control instructions based on the optimal solution, sending the instructions to the robot, and receiving the detection data fed back by the robot include: The optimal solution is sent to the robot, and the position of the robot is obtained according to a preset time period; after the robot receives the optimal solution, the robot adjusts the movement direction locally based on the optimal solution; Compare the position with the optimal solution, generate a higher-level control instruction, and send it to the robot; the time period is a dynamic period, which is related to the comparison result of the position with the optimal solution; Receive the detection data fed back by the robot, identify the detection data, generate the robot's speed control command according to the identification result, and send it to the robot.

5. The method for detecting and controlling the tension clamp of an intelligent robot according to claim 4, characterized in that: The step of receiving the detection data fed back by the robot, generating a control instruction based on the optimal solution and the detection data fed back by the robot, and sending the control instruction to the robot comprises: Receive the detection data fed back by the robot, and calculate the feedback frequency according to the shooting time of the detection data; Performing upper-level recognition on the detection data, and when the upper-level recognition result shows that there is no abnormality, determining a speed adjustment coefficient according to the feedback frequency, and sending the speed control instruction to the robot as the speed control instruction; When the upper-level recognition result is abnormal, the robot position at the time of shooting is queried, and the robot position is used as the target point to generate an upper-level control instruction, which is sent to the robot; Receive images fed back by the robot at the waypoints, and when the number of images reaches a preset threshold, generate a continue detection instruction and send it to the robot; Among them, during the movement, the robot performs local recognition on the detection data and determines whether to feed back an image based on the local recognition result; the local recognition adopts a convolution recognition scheme, and the convolution kernel adopts the convolution feature of the image of the tension clamp under a preset standard state; the upper-level recognition adopts a neural network model, and the characteristics of the samples of the neural network model are the tension clamp images, and the labels of the samples are abnormal types; the magnitude of the sample number is not less than a preset magnitude threshold.

6. The method for detecting and controlling the tension clamp of an intelligent robot according to claim 1, characterized in that: The method further comprises: On the robot side, the battery life is obtained in real time. When the battery life is less than the preset battery life threshold, other robots are queried within the preset distance range. After inserting its own identification in the feedback result, the feedback result is fed back to other robots. When the identification of the feedback result received by the main end is different from that of the sender, a warning message is generated pointing to the robot corresponding to the identification.

7. A tension clamp detection and control device for an intelligent robot, characterized in that: The device comprises: A detection zone creation module is used to obtain the installation position of the tension clamp, cluster the tension clamp according to the installation position, and obtain the detection zone; the clustering process is a three-dimensional clustering process; A scheme enumeration module is used to enumerate and arrange the detection areas, taking each arrangement scheme as a detection scheme and synchronously generating a quantity sequence for each detection scheme; the dimension of the quantity sequence is the same as the total number of detection areas, and each element value in the quantity sequence represents the number of wire clamps in the detection area at the corresponding position in the arrangement scheme; A scheme selection module is used to obtain the motion path of each detection scheme, calculate the energy consumption, and select the optimal scheme according to the energy consumption and quantity sequence; wherein the energy consumption is the total energy consumption of the robot working along the motion path; The instruction generation and sending module is used to receive the detection data fed back by the robot, generate control instructions based on the optimal solution and the detection data fed back by the robot, and send them to the robot; the detection data is an image containing the shooting time.

8. The intelligent robot tension clamp detection and control device according to claim 7, characterized in that: The detection zone creation module includes: A position query unit, used to query the installation position of the tension clamp in the record database; A distance calculation unit is used to calculate the spatial distance between any two tension clamps; A distance updating unit is used to query the power parameters of the lines where any two tension clamps are located and adjust the spatial distance according to the difference in power parameters; The clustering execution unit is used to cluster the tension clamps according to the adjusted spatial distance and construct a detection area.

9. The intelligent robot tension clamp detection and control device according to claim 7, characterized in that: The scheme selection module includes: A path determination unit, used to obtain a motion path for each detection scheme based on a preset path planning algorithm; An energy consumption calculation unit, used to obtain a total amount of height change and a total length of the motion path, and calculate energy consumption according to the total amount of height change and the total length of the path; the energy consumption is proportional to the total amount of height change, and the energy consumption is proportional to the total length of the path; A scoring calculation unit, used for reading the quantity sequence of the detection scheme, calculating the reverse number of the quantity sequence, and calculating the final score of the detection scheme according to the energy consumption and the reverse number; A scoring application unit, used to select the detection scheme with the highest final score as the optimal scheme; The calculation process of the final score is: ; In the formula, Indicates the final rating. and are preset constants, all of which are positive values. and is the preset correction coefficient, which is a positive value. is the total number of coordinate points in the motion path, Indicates the first The characteristic value of the coordinate point, ; Indicates the first The absolute value of the height difference corresponding to the coordinate points, The value of is: hour, ;when , ; Indicates The height corresponding to the coordinate point, Indicates The height corresponding to the previous coordinate point of the coordinate point; represents the total length of the path; Indicates a reverse number.

10. The intelligent robot tension clamp detection and control device according to claim 7, characterized in that: The instruction generation and sending module comprises: A solution sending unit is used to send the optimal solution to the robot and obtain the position of the robot according to a preset time period; after the robot receives the optimal solution, it locally adjusts the movement direction based on the optimal solution; A comparison and adjustment unit is used to compare the position with the optimal solution, generate a higher-level control instruction, and send it to the robot; the time period is a dynamic period, which is related to the comparison result of the position and the optimal solution; The speed regulating unit is used to receive the detection data fed back by the robot, identify the detection data, generate the speed control instruction of the robot according to the identification result, and send it to the robot.

Images