Intelligent robot arm control system and method based on internet of things

Abstract

The application discloses an intelligent robot arm control system and method based on an Internet of Things, relates to the technical field of equipment control, and comprises an equipment operation test and optimization module, a historical task information analysis module, an information comparison module and an abnormal early warning processing module. The equipment operation test and optimization module is used for performing equipment operation test on the intelligent robot arm and performing test optimization. The historical task information analysis module is used for predicting the time required by the intelligent robot arm to complete a current object grabbing task, which is recorded as a grabbing prediction time. The information comparison module is used for collecting load information of a current object to be grabbed, judging a range in which the load of the current object to be grabbed is located, and calling test information, which is recorded as a temperature rise test time. The grabbing prediction time and the temperature rise test time are compared. When the grabbing prediction time is greater than the temperature rise test time, the abnormal early warning processing module is used for predicting that there is a risk of object falling and performing abnormal early warning, so that the probability of failure of the robot arm operation task is reduced.

Description

Technical Field

[0001] This invention relates to the field of equipment control technology, specifically to an intelligent robotic arm control system and method based on the Internet of Things. Background Technology

[0002] With the development of science and technology, the application fields of intelligent robot arms are becoming more and more extensive, including object handling, parts assembly, inspection, welding and other fields. Intelligent robot arms are equipped with various sensors to perceive the surrounding environment and the position, shape, weight and other information of objects. The data perceived by the sensors will be transmitted to the control system of the robot arm. The control system can perform motion planning based on the received feedback information to control the robot arm to complete the operation task.

[0003] When using a smart robotic arm to grasp objects, it is necessary to ensure stable gripping force to avoid abnormal phenomena such as objects falling during the grasping process. However, in heavy-load scenarios, the reducer of the robotic arm may experience a rapid temperature rise during the grasping of heavy objects. When the temperature rises to a certain level, the frictional fluctuations may occur due to a sudden drop in the viscosity of the lubricating oil, which in turn causes fluctuations in the gripping force of the robotic arm. There is a certain probability that the object will fall during the grasping process, which is not conducive to the smooth completion of the operation task. Existing technology cannot predict such abnormal situations in advance through comprehensive analysis of sensing data, thereby reducing the probability of failure of the robotic arm operation task. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent robotic arm control system and method based on the Internet of Things (IoT) to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent robot arm control system based on the Internet of Things, the system including a device operation testing and optimization module, a historical task information analysis module, an information comparison module, and an anomaly early warning processing module;

[0006] The device operation test and optimization module is used to perform device operation tests and optimize the test process for the intelligent robotic arm.

[0007] The historical task information analysis module collects the load information of the object to be grasped, collects the time information of the intelligent robot arm to complete the object grasping task, and predicts the time required for the intelligent robot arm to complete the current object grasping task, which is recorded as the grasping prediction time.

[0008] The information comparison module determines the range of the load of the object to be grasped in the test data, retrieves the time required for the reducer of the robot arm to rise to the temperature threshold when grasping an object with a load within the corresponding range, and records it as the temperature rise test time. The grasping prediction time and the temperature rise test time are compared, and the load is the weight of the object to be grasped.

[0009] When the predicted grabbing time exceeds the temperature rise test time, the abnormal early warning processing module predicts that there is a risk of the object falling during the grabbing process and issues an abnormal task early warning.

[0010] Preferably, the equipment operation testing and optimization module includes an equipment operation testing unit and a testing optimization unit;

[0011] The equipment operation test unit is used to conduct equipment operation tests on the intelligent robot arm. The first few tests are conducted by controlling the robot arm to grasp objects with progressively increasing loads and sensing the temperature rise information of the robot arm's reducer during the grasping process.

[0012] The test optimization unit is used to optimize subsequent test methods and organize the test data after several tests.

[0013] Preferably, the historical task information analysis module includes a historical information acquisition unit and a capture time prediction unit;

[0014] The historical information collection unit is used to collect the load information of the current object to be grasped and the time information of the intelligent robot arm in the past when it completed the object grasping task with the same load as the current object to be grasped.

[0015] The grasping time prediction unit is used to predict the time required for the intelligent robot arm to complete the grasping task of the current object based on the time information collected for the grasping task, and the predicted time is recorded as the grasping prediction time.

[0016] Preferably, the information comparison module includes a test data retrieval unit and a time comparison unit;

[0017] The test data retrieval unit is used to determine the range of the load of the object to be grasped in the test data, retrieve the time required for the reducer of the robot arm to rise to the temperature threshold when grasping an object with a load within the corresponding range, and record the retrieved time as the temperature rise test time.

[0018] The time comparison unit is used to compare the temperature rise test time and the capture prediction time and obtain the comparison result.

[0019] The IoT-based intelligent robot arm control method includes the following steps:

[0020] S1: Conduct equipment operation tests on the intelligent robotic arm: control the robotic arm to grasp objects with progressively increasing loads, sense the temperature rise of the robotic arm's reducer during the grasping process, optimize subsequent test work after several tests, and organize the test data.

[0021] S2: Collect the load information of the object to be grasped, collect the time information of the intelligent robot arm to complete the object grasping task, and predict the time required for the intelligent robot arm to complete the current object grasping task, which is recorded as the grasping prediction time.

[0022] S3: Determine the range of the load of the object to be grasped in the test data, retrieve the time required for the reducer of the robot arm to rise to the temperature threshold when grasping an object with a load within the corresponding range, record it as the temperature rise test time, and compare the grasping prediction time and the temperature rise test time.

[0023] S4: If the predicted grabbing time is longer than the temperature rise test time, it is predicted that there is a risk of the object falling during the grabbing process, and a task abnormality warning is issued; otherwise, it is predicted that there is no risk.

[0024] Preferably, in step S1: Testing is conducted with load A as the initial load value and the load value k is increased sequentially. In the first n tests, the robot arm is controlled to grasp objects with loads of {A, A+k, A+2k, ... A+(n-1)k}. A temperature threshold of L is set. Testing shows that the gripping force of the robot arm fluctuates when the reducer temperature rises to L. The time taken for the reducer temperature to rise to L during the grasping process is detected by a temperature sensor as {T1, T2, T3, ... T... n}, where T1 represents the time taken for the temperature of the reducer to rise to L when the robot arm grasps an object with load A. The n time values ​​are randomly divided into m groups, where any time value in the first group is less than that in the second group. The set of average values ​​for the m groups after any grouping method is then {Q1, Q2, ..., Q}. m}, Q1 represents the average of all time values ​​in group 1. The data adoption confidence level D for any grouping method is calculated using the formula: D = [∑ m i=1 (Q i -(∑ m i=1 Q i ) / m) 2 ] 1 / 2The data adoption confidence level of all grouping methods is calculated in the same way. The grouping result with the highest data adoption confidence level is retrieved, and the set of average time values ​​of m groups in the grouping result with the highest data adoption confidence level is obtained as {Q1}. ’ Q2 ’ ,...Q m ’}, The minimum load corresponding to the time value in group 1 of the group with the highest data adoption confidence is A, and the maximum load is A+b1k. The first test data is then compiled as follows: When the robot arm grasps an object with a load within the range [A, A+b1k], the time it takes for the reducer temperature to rise to L is Q1. ’ Then, the test data from the 2nd to the mth time were compiled, and the load ranges of the first m items were obtained as {[A,A+b1k],(A+b1k,A+b2k],(A+b2k,A+b3k]...(A+b m-1 Let {k, A+(n-1)k]} be the range of loads for the first m terms. Calculate the difference between the maximum and minimum values ​​in the range of loads for the first m terms. The result is a set of differences {J1, J2, ... J}. m}, predict the difference for the (m+1)th term: the predicted value is J m+1 =J m *δ+(1-δ)*Z m Where δ represents the smoothing coefficient, 0 < δ < 1, Z m Z represents the exponentially smoothed value of the m-th difference, obtained by solving for it in the following way. m The exponentially smoothed value Z1 of the first difference is calculated using Z1=J1*δ+(1-δ)*[(J1+J2+J3) / 3]. The exponentially smoothed value Z2 of the second difference is calculated using Z2=J1*δ+(1-δ)*Z1. The exponentially smoothed value Z3 of the third difference is calculated using Z3=J2*δ+(1-δ)*Z2, and so on, until Z is obtained. m J1=b1k, J m =(n-1-b m-1 In the (n+1)th test, the object with load H1 is directly gripped, and the time taken for the reducer temperature to rise to L is recorded as Q. m+1 ’ The (m+1)th test data is: when the robot arm grasps an object with a load in the range (A+(n-1)k, H1], the time it takes for the reducer temperature to rise to L is Q. m+1 ’ H1=A+(n-1)k+J m+1Starting from the (n+2)th test, each subsequent test will use the same method as the (n+1)th test to confirm and test the load of the grabbed objects, and the data obtained from the subsequent tests will be organized.

[0025] To confirm the time taken for the reducer to reach the temperature threshold when the robotic arm grasps heavy objects with varying loads, a robotic arm grasping operation test was conducted. In the initial test phase, objects with loads increasing progressively from k were grasped. However, considering that if the load differences between grasped objects are too small, the time taken for the reducer to reach the temperature threshold might also be too small or even the same. Therefore, by analyzing the test data from the previous phase, loads with similar time durations were grouped together to form load ranges. It was determined that when grasping objects within the same load range, the time taken for the reducer to reach the temperature threshold was similar. Several load ranges and their corresponding time durations were compiled. As test data, in the later stages of testing, by analyzing the variation pattern of the difference between the maximum and minimum values ​​within several load ranges, it is predicted that the time it takes for the reducer to reach the threshold temperature when grasping objects with load differences within a certain range will be similar. Therefore, the test is directly conducted on objects with load difference threshold value, i.e., load H, and the time it takes for the reducer to reach the threshold temperature is obtained. This time is used as the test time for the reducer to reach the threshold temperature when grasping objects with loads within the corresponding load range. In the later stages of testing, the method of gradually increasing the load is no longer used. The equipment operation test of the robot arm is optimized, which saves test times and improves test efficiency while ensuring the accuracy of test data.

[0026] Preferably, in step S2: the load of the current object to be grasped is collected as F, and the intelligent robot arm has previously completed u object grasping tasks with a load of F. The set of times taken to complete u object grasping tasks with a load of F is {t1, t2, ... t}. u Let t1 represent the time spent completing the first object grasping task with a load of F. The grasping task refers to the process of the robotic arm grasping the object from its current location and moving it to the target location. The predicted time required for the intelligent robotic arm to complete the current object grasping task is t. u+1 , t u+1 =t u *δ+(1-δ)*P u , where P u P represents the exponentially smoothed value of the time taken by the intelligent robotic arm to complete the u-th object grasping task with a load of F. u The solution method and Z m The same, tu+1 This is recorded as the capture prediction time.

[0027] Preferably, in step S3: if the load F is determined to be within the load range (A+(n-1)k, H1) in the test data, then Q is... m+1 ’ Recorded as the temperature rise test time, t is compared. u+1 and Q m+1 ’ .

[0028] Preferably, in step S4: if t u+1 Q m+1 ’ If the prediction is correct, then there is a risk of the object falling during the grabbing process, and a task anomaly warning will be issued; otherwise, it is determined that there is no risk.

[0029] If the predicted grasping time is longer than the temperature rise test time, it indicates that the reducer temperature may rise to the temperature threshold during object grasping, leading to fluctuations in the gripping force of the robot arm and the risk of the object falling. Otherwise, it means that the object grasping task has been completed before the reducer temperature reaches the temperature threshold, and there is no risk of the object falling. This invention analyzes the time taken by the intelligent robot arm to grasp objects with the same load as the current object to be grasped in the past. Considering that the performance of the intelligent robot arm decreases over time, even when grasping objects with the same load, the time taken to complete the grasping task may vary. The time spent on grasping tasks can vary or even increase. Therefore, by dynamically predicting the grasping time based on historical data, a grasping prediction time can be obtained. This prediction time is then compared with the temperature rise test time. Based on the comparison results, it can be determined whether there is a risk of the object falling when grasping the current object. This helps to predict abnormal situations in advance and prompt appropriate measures to be taken. For example, during the grasping process, the speed of the robotic arm moving the object can be increased to avoid fluctuations in the gripping force that could cause the object to fall. This reduces the probability of the robotic arm failing to perform the task and improves the accuracy of predicting abnormal situations.

[0030] Compared with the prior art, the beneficial effects of the present invention are:

[0031] This invention aims to determine the time it takes for the reducer temperature to rise to a threshold when a robotic arm grasps heavy objects with varying loads. To investigate this, robotic arm grasping operation tests were conducted. In the initial testing phase, objects with loads increasing progressively with a base load of k were grasped. However, considering that if the load differences between grasped objects are too small, the time it takes for the reducer temperature to rise to the threshold might also be too small or even identical. Therefore, by analyzing the previous test data, loads with similar temperatures were grouped together to form load ranges. It was determined that when grasping objects within the same load range, the time it takes for the reducer temperature to rise to the threshold is similar. Several load ranges and their corresponding temperatures were then identified. Using this information as test data, in the later stages of testing, by analyzing the variation pattern of the difference between the maximum and minimum values ​​within several load ranges, we can predict the time it takes for the reducer to reach the threshold temperature when grasping objects with load differences within a certain range. Therefore, we directly use objects with load differences at the critical value (load H) for grasping tests, and obtain the time it takes for the reducer to reach the threshold temperature. This time is then used as the time it takes for the reducer to reach the threshold temperature when grasping objects with loads within the corresponding load range. In the later stages of testing, we no longer use the method of gradually increasing the load. This optimizes the operation of the robot arm, ensuring the accuracy of the test data while saving the number of tests and improving testing efficiency.

[0032] By analyzing the time taken by the intelligent robotic arm to grasp objects with the same load as the current object, and considering that the performance of the equipment decreases over time, even when grasping objects with the same load, the time taken to complete the grasping task may vary or even increase. Therefore, the grasping time is dynamically predicted based on historical data to obtain the grasping prediction time. The grasping prediction time is then compared with the temperature rise test time. Based on the comparison results, it is determined whether there is a risk of the object falling when grasping the current object. This helps to predict abnormal situations in advance and prompt appropriate measures, such as increasing the speed of the robotic arm's movement during the grasping process to avoid fluctuations in gripping force that could cause the object to fall. This reduces the probability of the robotic arm's operation task failure and improves the accuracy of predicting abnormal situations. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the structure of the IoT-based intelligent robot arm control system of the present invention;

[0034] Figure 2 This is a flowchart illustrating the IoT-based intelligent robot arm control method of the present invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example

[0036] like Figure 1 As shown, this embodiment provides an IoT-based intelligent robot arm control system, which includes: an equipment operation testing and optimization module, a historical task information analysis module, an information comparison module, and an anomaly warning processing module. The equipment operation testing and optimization module performs equipment operation testing on the intelligent robot arm and optimizes the test process. The historical task information analysis module collects the load information of the object to be grasped and the time information taken by the intelligent robot arm to complete the object grasping task, predicting the time required for the intelligent robot arm to complete the current object grasping task, and recording this as the grasping prediction time. The information comparison module determines the range of the load of the object to be grasped in the test data, retrieves the time required for the robot arm's reducer temperature to rise to the temperature threshold when grasping objects with loads within the corresponding range, and records this as the temperature rise test time. The grasping prediction time and the temperature rise test time are compared, with the load being the weight of the grasped object. The anomaly warning processing module, when the grasping prediction time exceeds the temperature rise test time, predicts a risk of the object falling during grasping and issues a task anomaly warning.

[0037] The equipment operation testing and optimization module includes an equipment operation testing unit and a test optimization unit. The equipment operation testing unit is used to conduct equipment operation tests on the intelligent robot arm. The first few tests are conducted by controlling the robot arm to grasp objects with progressively increasing loads and sensing the temperature rise information of the robot arm's reducer during the grasping process. The test optimization unit is used to optimize the subsequent test methods after several tests and to organize the test data.

[0038] The historical task information analysis module includes a historical information acquisition unit and a grasping time prediction unit. The historical information acquisition unit is used to collect the load information of the current object to be grasped and the time information of the intelligent robot arm in the past when it completed the grasping task of the same object load as the current object to be grasped. The grasping time prediction unit is used to predict the time required for the intelligent robot arm to complete the grasping task of the current object to be grasped based on the time information of the grasping task collected, and the predicted time is recorded as the grasping prediction time.

[0039] The information comparison module includes a test data retrieval unit and a time comparison unit. The test data retrieval unit is used to determine the range of the load of the object to be grasped in the test data, retrieve the time required for the reducer of the robot arm to rise to the temperature threshold when grasping an object with a load within the corresponding range, and record the retrieved time as the temperature rise test time. The time comparison unit is used to compare the temperature rise test time and the grasping prediction time and obtain the comparison result. Example

[0040] like Figure 2 As shown, this embodiment provides a smart robot arm control method based on the Internet of Things, which is implemented based on the control system in this embodiment, and specifically includes the following steps:

[0041] S1: Conduct equipment operation tests on the intelligent robotic arm: Control the robotic arm to grasp objects with progressively increasing loads, sense the temperature rise of the robotic arm's reducer during the grasping process, optimize subsequent tests after several tests, and organize the test data: Test with load A as the initial load value and progressively increasing the load value k: In the first n tests, control the robotic arm to grasp objects with loads of {A, A+k, A+2k, ... A+(n-1)k}, set the temperature threshold to L, and test to find that the gripping force of the robotic arm fluctuates when the reducer temperature rises to L. Use the temperature sensor to sense the time taken for the reducer temperature to rise to L during the grasping process as {T1, T2, T3, ... T n}, where T1 represents the time taken for the temperature of the reducer to rise to L when the robot arm grasps an object with load A. The n time values ​​are randomly divided into m groups, where any time value in the first group is less than that in the second group. The set of average values ​​for the m groups after any grouping method is then {Q1, Q2, ..., Q}. m}, Q1 represents the average of all time values ​​in group 1. The data adoption confidence level D for any grouping method is calculated using the formula: D = [∑ m i=1 (Q i -(∑ m i=1 Q i ) / m) 2 ] 1 / 2 The data adoption confidence level of all grouping methods is calculated in the same way. The grouping result with the highest data adoption confidence level is retrieved, and the set of average time values ​​of m groups in the grouping result with the highest data adoption confidence level is obtained as {Q1}. ’ Q2 ’ ,...Q m ’}, The minimum load corresponding to the time value in group 1 of the group with the highest data adoption confidence is A, and the maximum load is A+b1k. The first test data is then compiled as follows: When the robot arm grasps an object with a load within the range [A, A+b1k], the time it takes for the reducer temperature to rise to L is Q1. ’ Then, the test data from the 2nd to the mth time were compiled, and the load ranges of the first m items were obtained as {[A,A+b1k],(A+b1k,A+b2k],(A+b2k,A+b3k]...(A+b m-1 Let {k, A+(n-1)k]} be the range of loads for the first m terms. Calculate the difference between the maximum and minimum values ​​in the range of loads for the first m terms. The result is a set of differences {J1, J2, ... J}. m}, predict the difference for the (m+1)th term: the predicted value is J m+1 =J m *δ+(1-δ)*Z m Where δ represents the smoothing coefficient, 0 < δ < 1, Z m Z represents the exponentially smoothed value of the m-th difference, obtained by solving for it in the following way. m The exponentially smoothed value Z1 of the first difference is calculated using Z1=J1*δ+(1-δ)*[(J1+J2+J3) / 3]. The exponentially smoothed value Z2 of the second difference is calculated using Z2=J1*δ+(1-δ)*Z1. The exponentially smoothed value Z3 of the third difference is calculated using Z3=J2*δ+(1-δ)*Z2, and so on, until Z is obtained. m J1=b1k, J m =(n-1-b m-1 In the (n+1)th test, the object with load H1 is directly gripped, and the time taken for the reducer temperature to rise to L is recorded as Q. m+1 ’ The (m+1)th test data is: when the robot arm grasps an object with a load in the range (A+(n-1)k, H1], the time it takes for the reducer temperature to rise to L is Q. m+1 ’ H1=A+(n-1)k+J m+1 Starting from the (n+2)th test, each subsequent test will use the same method as the (n+1)th test to confirm and test the load of the grabbed objects, and the data obtained from the subsequent tests will be organized.

[0042] For example: In the (n+2)th test, the object with a load of H2 is directly grasped, and the time taken for the reducer to heat up to L during the grasping process is measured, where H2 = H1 + J. m+2If the test shows that the time it takes for the reducer to reach temperature L is 10 minutes, then the data for the (n+2)th test is: when using an intelligent robotic arm to grasp an object with a load in the range of (H1, H2), the time it takes for the reducer to reach temperature L is 10 minutes.

[0043] S2: Collect the load information of the object to be grasped, collect the time information of the intelligent robot arm in completing the object grasping task, and predict the time required for the intelligent robot arm to complete the current object grasping task, denoted as the grasping prediction time: The load of the current object to be grasped is collected as F, and the intelligent robot arm has completed u object grasping tasks with a load of F in the past is counted. The set of times taken to complete u object grasping tasks with a load of F is {t1, t2, ... t}. u Let t1 represent the time spent completing the first object grasping task with a load of F. The grasping task refers to the process of the robotic arm grasping the object from its current location and moving it to the target location. The predicted time required for the intelligent robotic arm to complete the current object grasping task is t. u+1 , t u+1 =t u *δ+(1-δ)*P u , where P u P represents the exponentially smoothed value of the time taken by the intelligent robotic arm to complete the u-th object grasping task with a load of F. u The solution method and Z m The same, t u+1 This is recorded as the capture prediction time;

[0044] S3: Determine the range of the load of the object to be grasped in the test data. When grasping an object with a load within the corresponding range, retrieve the time required for the robot arm's reducer temperature to rise to the temperature threshold, recorded as the temperature rise test time. Compare the grasping prediction time and the temperature rise test time: If the load F is determined to be within the load range (A+(n-1)k, H1) in the test data, then Q... m+1 ’ Recorded as the temperature rise test time, t is compared. u+1 and Q m+1 ’ ;

[0045] S4: If the predicted grasping time is longer than the temperature rise test time, a risk of the object falling during the grasping process is predicted, and a task anomaly warning is issued; otherwise, no risk is predicted: if t u+1 Q m+1 ’ If the system predicts that there is a risk of the object falling during the grabbing process, an abnormal task warning will be issued; otherwise, if no risk is predicted, the system will issue a warning.

[0046] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for controlling an intelligent robot arm based on the Internet of Things, characterized in that: Includes the following steps: S1: Conduct equipment operation tests on the intelligent robotic arm: control the robotic arm to grasp objects with progressively increasing loads, sense the temperature rise of the robotic arm's reducer during the grasping process, optimize subsequent test work after several tests, and organize the test data. S2: Collect the load information of the object to be grasped, collect the time information of the intelligent robot arm to complete the object grasping task, and predict the time required for the intelligent robot arm to complete the current object grasping task, which is recorded as the grasping prediction time. S3: Determine the range of the load of the object to be grasped in the test data, retrieve the time required for the reducer of the robot arm to rise to the temperature threshold when grasping an object with a load within the corresponding range, record it as the temperature rise test time, and compare the grasping prediction time and the temperature rise test time. S4: If the predicted grabbing time is longer than the temperature rise test time, and the risk of the object falling during the grabbing process is predicted, an abnormal task warning will be issued. Otherwise, it is predicted that there is no risk. In step S1: Testing is conducted with load A as the initial load value and the load value k is increased sequentially. In the first n tests, the robot arm is controlled to grasp objects with loads of {A, A+k, A+2k, ..., A+(n-1)k}. A temperature threshold of L is set, and the time taken for the reducer temperature to rise to L during the grasping process is detected by a temperature sensor as {T1, T2, T3, ..., T...}. n }, where T1 represents the time taken for the temperature of the reducer to rise to L when the robot arm grasps an object with load A. The n time values ​​are randomly divided into m groups, and the set of average values ​​of the m groups after any grouping method is {Q1, Q2, ... Q}. m }, Q1 represents the average of all time values ​​in group 1. The data adoption confidence level D for any grouping method is calculated using the formula: D = [∑ m i=1 (Q i -(∑ m i=1 Q i ) / m) 2 ] 1 / 2 The data adoption confidence level of all grouping methods is calculated in the same way. The grouping result with the highest data adoption confidence level is retrieved, and the set of average time values ​​of m groups in the grouping result with the highest data adoption confidence level is obtained as {Q1}. ’ Q2 ’ ,...Q m ’ }, The minimum load corresponding to the time value in group 1 of the group with the highest data adoption confidence is A, and the maximum load is A+b1k. The first test data is then compiled as follows: When the robot arm grasps an object with a load within the range [A, A+b1k], the time it takes for the reducer temperature to rise to L is Q1. ’ Then, the test data from the 2nd to the mth time were compiled, and the load ranges of the first m items were obtained as {[A,A+b1k],(A+b1k,A+b2k],(A+b2k,A+b3k]...(A+b m-1 k,A+(n-1)k]}; Calculate the difference between the maximum and minimum values ​​in the first m load ranges to obtain the set of differences {J1, J2, ... J... m }, predict the difference for the (m+1)th term: the predicted value is J m+1 =J m *δ+(1-δ)*Z m Where δ represents the smoothing coefficient, 0 < δ < 1, Z m J1 represents the exponentially smoothed value of the m-th difference, J1=b1k, J m =(n-1-b m-1 In the (n+1)th test, the object with load H1 is directly gripped, and the time taken for the reducer temperature to rise to L is recorded as Q. m+1 ’ The (m+1)th test data is: when the robot arm grasps an object with a load in the range (A+(n-1)k, H1], the time it takes for the reducer temperature to rise to L is Q. m+1 ’ H1=A+(n-1)k+J m+1 Starting from the (n+2)th test, each subsequent test uses the same method as the (n+1)th test to confirm and test the load on the captured object, and then organizes the data obtained from the subsequent tests.

2. The IoT-based intelligent robot arm control method according to claim 1, characterized in that: In step S2: The load of the object to be grasped is collected as F. The intelligent robot arm has previously completed u object grasping tasks with a load of F. The time set for completing u object grasping tasks with a load of F is {t1, t2, ... t}. u Let t1 represent the time spent completing the first object grasping task with a load of F. The predicted time for the intelligent robotic arm to complete the current object grasping task is t. u+1 , t u+1 =t u *δ+(1-δ)*P u , where P u The exponentially smoothed value representing the time taken by the intelligent robotic arm to complete the u-th object grasping task with a load of F is denoted by t. u+1 Record this as the capture prediction time.

3. The IoT-based intelligent robot arm control method according to claim 2, characterized in that: In step S3: If the load F is determined to be within the load range (A+(n-1)k, H1) in the test data, then Q will be... m+1 ’ Recorded as the temperature rise test time, t is compared. u+1 and Q m+1 ’ .

4. The IoT-based intelligent robot arm control method according to claim 3, characterized in that: In step S4: if t u+1 Q m+1 ’ If it is predicted that there is a risk of the object falling during the grabbing process, an abnormal task warning will be issued. Otherwise, it is assumed that there is no risk.

Images