Robot collision detection method, robot system, and computer-readable medium

Abstract

The application relates to a robot collision detection method, a robot system and a computer readable medium. The method obtains current values of each joint every preset time interval within a preset monitoring period, preanalyzes the current values of the current values of each joint, if the analysis obtains that the current value fluctuation within the preset monitoring period is abnormal, further calculates a theoretical joint torque and a theoretical current value, and compares the theoretical current value with a current current value to verify. If the analysis obtains that the current value fluctuation within the preset monitoring period is not abnormal, the theoretical current value does not need to be calculated. The current value within the preset monitoring period is preanalyzed, unnecessary calculation amount and calculation time during collision detection are reduced, and the real-time requirement of a mechanical arm is met.

Description

Technical Field

[0001] This application relates to the field of robot control technology, and in particular to a robot collision detection method, a robot system, and a computer-readable medium. Background Technology

[0002] In endoscopic surgery, medical robots can assist surgeons in automatically adjusting the surgical field of view. Through visual feedback, the medical robot achieves rapid endoscope positioning and then autonomously adjusts the surgical field of view to assist the surgeon in completing the procedure. When medical robots perform human-computer interaction tasks in a collaborative shared space, ensuring human-machine safety is a prerequisite for human-computer interaction. This requires real-time detection of collisions between the medical robot and the human, and between the robot and the environment. In the event of a collision, safety measures must be implemented to prevent accidents.

[0003] Traditional collision detection for medical robots involves acquiring the current value of each joint at the current moment within each control cycle and calculating the actual joint torque based on the current value. Furthermore, it requires acquiring the angular displacement, angular velocity, and angular acceleration of each joint within each control cycle and inputting them into a kinematic model for analysis. This allows the kinematic model to output the theoretical torque at the current moment, and the actual joint torque (current) is compared with the theoretical torque (current) to determine whether a collision has occurred. Considering the extremely short control cycle, typically around 1 ms, calculating the theoretical torque every control cycle would consume a significant amount of computational resources, making it difficult to meet the real-time requirements of robotic arms. Summary of the Invention

[0004] Therefore, it is necessary to address the problem that traditional solutions require calculating the actual joint torque and theoretical torque once in each control cycle, which consumes a lot of computing resources and makes it difficult to meet the real-time requirements of robotic arms. To address this, a robot collision detection method, robot system, and computer-readable medium should be provided.

[0005] This application provides a robot collision detection method, applied to a robot system, the robot system including a robotic arm having multiple rotatable joints; the robot collision detection method includes:

[0006] The current value of each joint is acquired at preset time intervals within the preset monitoring period;

[0007] Select a joint and obtain the current current value of that joint;

[0008] Calculate the average current change of the joint within a preset monitoring period;

[0009] Return to the selected joint and obtain the current current value of that joint, until all joints have been selected;

[0010] Determine whether there is at least one joint where the average current change is greater than a preset current change threshold.

[0011] If the average current change of at least one joint is greater than the preset current change threshold, then further determine whether at least one joint is abnormal based on the current angular displacement, current angular velocity, current angular acceleration and current value of each joint.

[0012] If at least one joint is abnormal, the abnormal joint is defined as an abnormal joint, and the fault information of the corresponding abnormal joint is obtained.

[0013] Output the fault information corresponding to the abnormal joint to the display device.

[0014] This application also provides a robot system, including:

[0015] The processor is used to execute the robot collision detection method mentioned above;

[0016] A display device, wherein the display device is communicatively connected to the processor;

[0017] A robotic arm with multiple rotatable joints;

[0018] The robotic arm is electrically connected to the processor.

[0019] This application also provides a computer-readable medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the robot collision detection method as described above.

[0020] This application relates to a robot collision detection method, a robot system, and a computer-readable medium. The method acquires the current values ​​of each joint at preset time intervals within a preset monitoring period, performs pre-analysis on the current values ​​of each joint, and if the analysis reveals abnormal fluctuations in the current values ​​within the preset monitoring period, further calculates the theoretical joint torque and theoretical current values, comparing the theoretical current values ​​with the current values ​​for verification. If the analysis reveals no abnormal fluctuations in the current values ​​within the preset monitoring period, there is no need to calculate the theoretical current values. By employing pre-analysis of current values ​​within a preset monitoring period, unnecessary computational load and time consumption during collision detection are reduced, thereby meeting the real-time requirements of the robotic arm. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating a robot collision detection method provided in an embodiment of this application.

[0022] Figure 2 This is a schematic diagram of the joint state matrix in a robot collision detection method provided in an embodiment of this application.

[0023] Figure 3 This is a structural diagram of a robot system provided in one embodiment of this application.

[0024] Figure label:

[0025] 100-Robot System;

[0026] 110 - Processor; 120 - Display device; 130 - Robotic arm. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0028] This application provides a robot collision detection method.

[0029] It should be noted that the robot collision detection method provided in this application does not limit the subject of execution. Optionally, the field-of-view control method provided in this application is applied to a robot system, the robot system including a robotic arm having multiple rotatable joints.

[0030] like Figure 1 As shown, in one embodiment of this application, the robot collision detection method includes the following steps S100 to S810:

[0031] S100 acquires the current value of each joint at preset time intervals within a preset monitoring cycle.

[0032] Specifically, the preset time period can be set according to system requirements, and can be 1ms. The preset monitoring cycle can also be set according to system requirements, and the current values ​​of each joint during the preset monitoring cycle can be stored for selection in subsequent steps.

[0033] S200: Select a joint and obtain the current current value of that joint.

[0034] Specifically, the current current value corresponds to the current system time.

[0035] S300, calculate the average current change of the joint within a preset monitoring period.

[0036] Specifically, the average current change of the joint within the preset monitoring period is calculated according to the following formula 1:

[0037]

[0038] Where I0 represents the average current change, I n Indicates the current value, In-j This represents the current value obtained in the previous j times, where j is an integer greater than 2.

[0039] S400, return to the step of selecting a joint and obtaining the current current value of that joint, until all joints have been selected.

[0040] S500, determine whether there is at least one joint whose average current change is greater than a preset current change threshold.

[0041] S610, if the average current change of at least one joint is greater than the preset current change threshold, then further determine whether at least one joint is abnormal based on the current angular displacement, current angular velocity, current angular acceleration and current value of each joint.

[0042] S710, if at least one joint is abnormal, the abnormal joint is defined as an abnormal joint, and the fault information of the corresponding abnormal joint is obtained.

[0043] Specifically, the fault information includes alarm fault codes. For example, if a robotic arm has 7 rotatable joints, the fault codes and corresponding abnormal conditions are as follows:

[0044] E11: Joint 1 current abnormality. E12: Joint 2 current abnormality. E13: Joint 3 current abnormality. E14: Joint 4 current abnormality. E15: Joint 5 current abnormality. E16: Joint 6 current abnormality. E17: Joint 7 current abnormality.

[0045] E21: Abnormal position of joint 1. E22: Abnormal position of joint 2. E23: Abnormal position of joint 3. E24: Abnormal position of joint 4. E25: Abnormal position of joint 5. E26: Abnormal position of joint 6. E27: Abnormal position of joint 7. The position indicates the angular displacement of the joint.

[0046] Other alarm fault codes include: E01: Image recognition device connection failure. E02: Robotic arm self-test failure. E03: Robotic arm collision failure (determined based on joint abnormality). E04: Software verification failure.

[0047] S810, output the fault information corresponding to the abnormal joint to the display device.

[0048] In this embodiment, the current values ​​of each joint are acquired at preset time intervals within a preset monitoring period. A pre-analysis of the current values ​​of each joint is performed. If the analysis reveals abnormal fluctuations in the current values ​​within the preset monitoring period, the theoretical joint torque and theoretical current values ​​are further calculated and compared with the current values ​​for verification. If the analysis shows no abnormal fluctuations in the current values ​​within the preset monitoring period, there is no need to calculate the theoretical current values. This pre-analysis of current values ​​within the preset monitoring period reduces unnecessary computational load and time consumption during collision detection, thereby meeting the real-time requirements of the robotic arm.

[0049] In one embodiment of this application, S610 includes the following S611a to S611f.

[0050] S611a, input the current angular displacement, current angular velocity and current angular acceleration of each joint into the joint dynamics model, run the joint dynamics model, and obtain the theoretical joint torque of each joint output by the joint dynamics model.

[0051] Specifically, the joint dynamics model is as follows:

[0052]

[0053] Where, τ d =[τ1,τ2,...τ i ] T For the theoretical joint torque, q = [q1, q2, ... q i ] T This represents the current angular displacement of each joint, where i represents the joint number and is an integer greater than or equal to 2. This indicates the current angular velocity of each joint. This represents the current angular acceleration of each joint. The term representing the inertial force of the joint. G(q) represents the Cosian force and centripetal force terms of the joint, and G(q) represents the gravity term.

[0054] S611b, Select a joint.

[0055] S611c, Input the theoretical torque of the joint into the torque-current dynamic model, run the torque-current dynamic model, and obtain the theoretical current value output by the torque-current dynamic model;

[0056] Specifically, the torque-current dynamic model is as follows:

[0057] I ai =k i τ i

[0058] Where i represents the joint number, and i is an integer greater than or equal to 2, τ i I represents the theoretical joint torque corresponding to the joint number. ai This represents the current value corresponding to the joint number, k. i This represents the conversion coefficient corresponding to the joint number. The conversion coefficient is the current and torque coefficient corresponding to the joint motor.

[0059] S611d, determine whether the current difference between the current value and the theoretical current value of the joint is greater than a preset current difference threshold.

[0060] S611e, if the current difference between the current value and the theoretical current value of the joint is greater than a preset current difference threshold, then the joint is marked as abnormal.

[0061] S611f, return to the selected joint, until all joints have been selected.

[0062] In this embodiment, the actual joint torque of each joint is compared with the theoretical joint torque. If the torque difference exceeds the threshold range, it indicates a joint abnormality. Then, the abnormality of each joint can be combined to determine whether a collision has occurred, thus ensuring the safety of robot operation.

[0063] In one embodiment of this application, S610 includes the following S612a to S612b.

[0064] S612a constructs a joint state matrix by combining the current angular displacement, current angular velocity, current angular acceleration, and current current value of each joint.

[0065] Specifically, the joint state matrix is ​​as follows: Figure 2 As shown, the rows of the joint state matrix are composed of the current angular displacement, current angular velocity, current angular acceleration, and current current value of each joint.

[0066] S612b, Input the joint state matrix into an anomaly analysis model, run the anomaly analysis model, and obtain the analysis results output by the anomaly analysis model.

[0067] Specifically, the analysis results include the serial numbers and fault information of the joints that are normal or abnormal.

[0068] In this embodiment, the state of each joint (the sequence number of joints with no abnormality or abnormality and fault information) is directly analyzed based on the data matrix composed of the current angular displacement, current angular velocity, current angular acceleration and current value, and the anomaly analysis model is used to determine whether each joint of the robotic arm is abnormal. There is no need to calculate the theoretical torque and actual torque, which better meets the real-time requirements of the robotic arm.

[0069] Specifically, the anomaly analysis model uses a convolutional neural network model.

[0070] The training method for the anomaly analysis model is as follows: A sufficient number of joint state matrices are obtained from the system database. These joint state matrices consist of the current angular displacement, current angular velocity, current angular acceleration, and current value of each joint during normal operation of the robotic arm. The joint state matrices are used as input data for the training set of the anomaly analysis model. The actual joint states corresponding to these joint state matrices (the serial numbers of joints without anomalies or with anomalies, and fault information) are used as output data for the training set of the anomaly analysis model. The anomaly analysis model is trained using these input and output data, and then tested and verified to obtain a converged anomaly analysis model.

[0071] In one embodiment of this application, after S600, the robot collision detection method further includes S620.

[0072] S620, if the average current change of all joints is less than or equal to the preset current change threshold, then return to the current value of each joint obtained at preset time intervals within the preset monitoring period.

[0073] In this embodiment, the average current change of the joints within a preset monitoring period is calculated to make a preliminary judgment on the fluctuation of the joint current value. If the average current change of all joints is less than or equal to the preset current change threshold, it means that the current value fluctuation is within a safe range, and no further verification calculation is required, thus reducing unnecessary calculations.

[0074] In one embodiment of this application, prior to S810, the robot collision detection method further includes the following S721 to S724.

[0075] S721, Select an abnormal joint.

[0076] S722, determine whether the abnormal joints all exhibit abnormal behavior within a preset monitoring period.

[0077] S723, if the abnormal joints all exhibit abnormal behavior within the preset monitoring period, then execute the output of the fault information corresponding to the abnormal joint to the display device.

[0078] S724, return to the step of selecting an abnormal joint, until all abnormal joints have been selected.

[0079] In this example, if the abnormal state of the joint persists for a certain period of time, it is considered that a contact collision has occurred, thus avoiding the possibility of misjudgment in a single detection result and improving the accuracy of collision detection.

[0080] In one embodiment of this application, after S810, the robot collision detection method further includes the following S821 to S824.

[0081] S821, obtain the operator's identity information and the historical number of operations corresponding to that identity information.

[0082] Specifically, the operator wears an identification bracelet that records the operator's identity information and the corresponding number of historical operations, which can be obtained by the system after interaction.

[0083] S822, determine whether the number of historical operations is less than the operation count threshold.

[0084] S823: If the number of historical operations is less than the operation count threshold, then an alarm signal and a shutdown signal will be output.

[0085] Specifically, warning signals include audible alarms or visual alarms.

[0086] S824: If the number of historical operations is greater than or equal to the operation count threshold, only an alarm signal will be output.

[0087] In this embodiment, the operator's proficiency is assessed based on their historical number of operations. If the historical number of operations is less than a threshold (i.e., low proficiency), a warning signal is issued and the machine is stopped to prevent secondary harm to the patient. If the historical number of operations is less than the threshold (i.e., high proficiency), only a warning signal is issued, and the operator's judgment is required to determine whether to stop the machine, thus preventing interruption of the surgical procedure.

[0088] This application also provides a robot system.

[0089] like Figure 3 As shown, in one embodiment of this application, the robot system 100 includes: a processor 110, a display device 120, and a robotic arm 130.

[0090] Specifically, processor 110 is used in the robot collision detection method mentioned above. Display device 120 is communicatively connected to the processor. Robotic arm 130 has multiple rotatable joints and is electrically connected to the processor.

[0091] This application also provides a computer-readable medium.

[0092] In one embodiment of this application, a computer program is stored on a computer-readable medium, wherein the computer program, when executed by a processor, implements the target following control method as mentioned above.

[0093] The technical features of the above embodiments can be combined arbitrarily, and the execution order of the method steps is not restricted. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0094] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A robot collision detection method, characterized in that, The application is to a robotic system, which includes a robotic arm having multiple rotatable joints; The robot collision detection method includes: The current value of each joint is acquired at preset time intervals within the preset monitoring period; Select a joint and obtain the current current value of that joint; Calculate the average current change of the joint within a preset monitoring period; Return to the selected joint and obtain the current current value of that joint, until all joints have been selected; Determine whether there is at least one joint where the average current change is greater than a preset current change threshold. If the average current change of at least one joint is greater than the preset current change threshold, then further determine whether at least one joint is abnormal based on the current angular displacement, current angular velocity, current angular acceleration and current value of each joint. If at least one joint is abnormal, the abnormal joint is defined as an abnormal joint, and the fault information of the corresponding abnormal joint is obtained. Output the fault information corresponding to the abnormal joint to the display device.

2. The robot collision detection method according to claim 1, characterized in that, The average current change of the joint within the preset monitoring period is calculated according to the following formula 1: Official 1; Where I0 represents the average current change, I n Indicates the current value, I n-j This represents the current value obtained in the previous j times, where j is an integer greater than 2.

3. The robot collision detection method according to claim 1, characterized in that, The further determination of whether at least one joint is abnormal based on the current angular displacement, current angular velocity, current angular acceleration, and current current value of each joint includes: The current angular displacement, current angular velocity, and current angular acceleration of each joint are input into the joint dynamics model. The joint dynamics model is run to obtain the theoretical joint torque of each joint output by the joint dynamics model. Select a joint; The theoretical torque of the joint is input into the torque-current dynamic model, the torque-current dynamic model is run, and the theoretical current value output by the torque-current dynamic model is obtained. Determine whether the current difference between the current value and the theoretical current value of the joint is greater than a preset current difference threshold; If the difference between the current current value and the theoretical current value of the joint is greater than a preset current difference threshold, the joint is marked as abnormal. Return to the previous step of selecting a joint, and continue until all joints have been selected.

4. The robot collision detection method according to claim 3, characterized in that, The joint dynamics model is as follows: in, For theoretical joint torque, This represents the current angular displacement of each joint, where i represents the joint number. This indicates the current angular velocity of each joint. This represents the current angular acceleration of each joint. The term representing the inertial force of the joint. The Cosian force and centripetal force terms of the joint are represented. This represents the gravity term.

5. The robot collision detection method according to claim 3, characterized in that, The torque-current dynamic model is as follows: Where i represents the joint number. This represents the theoretical joint torque corresponding to the joint number. This indicates the current value for the corresponding joint number. This represents the conversion coefficient for the corresponding joint number.

6. The robot collision detection method according to claim 1, characterized in that, The further determination of whether at least one joint is abnormal based on the current angular displacement, current angular velocity, current angular acceleration, and current current value of each joint includes: A joint state matrix is ​​constructed by taking the current angular displacement, current angular velocity, current angular acceleration, and current current value of each joint. The joint state matrix is ​​input into an anomaly analysis model, the anomaly analysis model is run, and the analysis results output by the anomaly analysis model are obtained. The analysis results include the serial numbers and fault information of the joints that are normal or abnormal.

7. The robot collision detection method according to claim 1, characterized in that, Before outputting the fault information corresponding to the abnormal joint to the display device, the robot collision detection method further includes: Select an abnormal joint; Determine whether the abnormal joint exhibits abnormal behavior within a preset monitoring period; If the abnormal joints are not abnormal within the preset monitoring period, then the fault information corresponding to the abnormal joints will be output to the display device. Return to the previous step of selecting an abnormal joint, and continue until all abnormal joints have been selected.

8. The robot collision detection method according to claim 1, characterized in that, If at least one joint is abnormal, the abnormal joint is defined as an abnormal joint, and fault information corresponding to the abnormal joint is obtained. The robot collision detection method further includes: Obtain the operator's identity information and the corresponding historical number of operations; Determine whether the number of historical operations is less than the operation count threshold; If the number of historical operations is less than the operation threshold, an alarm signal and a shutdown signal will be output. If the number of historical operations is greater than or equal to the operation count threshold, only an alarm signal will be output.

9. A robot system, characterized in that, include: A processor for performing the robot collision detection method as described in any one of claims 1 to 8; A display device, wherein the display device is communicatively connected to the processor; A robotic arm with multiple rotatable joints; The robotic arm is electrically connected to the processor.

10. A computer-readable medium having a computer program stored thereon, wherein, When the computer program is executed by the processor, it implements the robot collision detection method as described in any one of claims 1 to 8.

Images