CN-122008202-A - Method and system for estimating joint external force of heavy-load industrial robot operating on large ship

CN122008202ACN 122008202 ACN122008202 ACN 122008202ACN-122008202-A

Abstract

The invention relates to a method and a system for estimating joint external force of a heavy-load industrial robot operating on a large ship, and belongs to the technical field of robot motor driving joint control. The method comprises the steps of firstly obtaining actual motor current of a target joint motor, motion parameters of a target joint and base acceleration of the moving base when the moving base of the robot contacts the moving base, predicting non-contact operation state joint motor current of the moving base of the robot according to the motion parameters, calculating joint gravity moment of the moving base and the static base according to the base acceleration, further obtaining disturbance current of the moving base joint motor caused by ship swinging, subtracting the non-contact current from the actual motor current, subtracting the disturbance current from the non-contact current, obtaining joint external force current of the joint when the moving base of the robot contacts the moving base, and further obtaining joint external force of the joint through calculation. The method can realize the external force estimation of the robot joint under the condition of not adding the joint force sensor.

Inventors

WU PEIZHANG
XIONG ZIJIAN
LI PENGFEI
GUO JIANPO
WANG WEIRU
BAO YIFEI

Assignees

中国船舶集团有限公司第七一三研究所

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (10)

1. The method for estimating the joint external force of the heavy-load industrial robot for operation on the large-scale ship is characterized by comprising the following steps of: 1) Obtaining actual motor current of a target joint motor, motion parameters of a target joint and base acceleration of a moving base of the robot when the moving base contacts the robot during operation; 2) According to the motion parameters, predicting the non-contact current of a non-contact operation state joint motor under the joint static base of the robot; The method comprises the steps of obtaining a dynamic acceleration vector of the joint under a robot base coordinate system according to the base acceleration, bringing the dynamic acceleration vector and a static acceleration vector obtained based on gravity acceleration into a joint gravity moment equation to correspondingly calculate joint gravity moment under a dynamic base and static base state respectively, obtaining power-off current values required by the joint to overcome gravity under the dynamic base and static base state respectively according to the joint transmission coefficient, and obtaining disturbance current of a dynamic base joint motor caused by ship oscillation by difference between the dynamic base and the static base state; 3) And subtracting the non-contact current from the actual motor current and subtracting the disturbance current to obtain the joint external force current of the joint when the robot moves the base and contacts the operation, and calculating to obtain the joint external force of the joint when the robot moves the base and contacts the operation according to the transmission ratio of the joint motor shaft to the joint output shaft.
2. The method for estimating the external force of the joint of the heavy-duty industrial robot operating on the large-scale ship according to claim 1, wherein the inertial sensor is adopted to collect the base acceleration, and the advanced prediction algorithm based on gray prediction is adopted to obtain an acceleration prediction result for compensating the real-time influence of the transmission time lag and the update period of the inertial sensor on the base acceleration, so as to obtain the motion acceleration vector.
3. The method for estimating an external force of a joint of a heavy-duty industrial robot operating on a large vessel according to claim 2, wherein the advance prediction algorithm comprises the steps of: (2.1) storing the acceleration state measured values uploaded by M inertial sensors as a historical data sequence X h , and taking the latest M (5 < < M > < M) acceleration data from X h as an original data sequence X (0) participating in the current gray prediction; (2.2) including historical reference predictions: a) Finding a sequence meeting |X h (k+i)-X (0) (i) | < delta in the first M-M data in X h as a reference sequence X r , wherein delta is a set error threshold value, k is more than or equal to 1 and less than or equal to M-2M, and i=1, 2. B) Calculation of the degree of similarity of sequence X (0) to sequence X r ; C) Taking the sampling data X h (k+m+1) in X h as a one-step reference predicted value; Also included is an improved grey prediction: a) Judging whether an inflection point exists in the data sequence X (0) ; b) If the inflection point exists, performing advanced one-step prediction by using a first-order differential mean value of acceleration data after the inflection point to obtain a prediction result; If the inflection point does not exist, adopting a gray prediction model GM (1, 1) to conduct advanced one-step prediction to obtain a prediction result; c) Obtaining a prediction reference value according to a prediction result; and (2.3) carrying out weighted summation on the reference predicted value obtained by the historical reference prediction and the reference predicted value obtained by the improved gray prediction to obtain the acceleration prediction result.
4. A method for estimating external force of a joint of a heavy-duty industrial robot operating on a large vessel according to claim 3, wherein in the step c) of the improved gray prediction in the step (2.2), a residual sequence between the original data sequence X (0) and a predicted result sequence formed by a past predicted result is calculated; If a plurality of continuous residual values in the residual sequence are all larger than zero or all smaller than zero, modeling a gray prediction model GM (1, 1) for the plurality of continuous residual values, performing one-step prediction in advance to obtain a residual prediction correction value, and correcting the prediction result obtained in the step b) by using the residual prediction correction value to obtain an improved gray prediction reference value; Otherwise, taking the prediction result obtained in the step b) as a prediction reference value of improved gray prediction.
5. A method for estimating external force of a heavy-duty industrial robot joint operating on a large vessel according to claim 3, wherein in the step a) of the improved gray prediction in the step (2.2), it is judged whether or not there is an inflection point in the data sequence X (0) by: Defining a data sequence The kth discrete point in (a) The first order difference of (a) is: The second order difference is: If it is Then And the inflection point is the inflection point, otherwise, the non-inflection point is the inflection point.
6. The method for estimating the external force of the joint of the heavy-duty industrial robot operating on the large-scale ship according to claim 3, wherein in the step (2.1), the latest acceleration state measurement value uploaded by the inertial sensor is obtained in each sampling period, the acceleration state measurement value is processed into a positive number by an interval proportional conversion method and then stored as a historical data sequence X h , after the inertial sensor data at the latest moment is collected each time, the latest data after processing is added to the tail part of the historical data sequence X h , and the data at the earliest moment in the historical data sequence X h is removed.
7. The method for estimating the external force of the joint of the heavy-duty industrial robot operating on the large-scale ship according to claim 1, wherein the non-contact current is predicted by using a trained CNN-LSTM model according to the motion parameters including the joint displacement and the motor rotation speed of the joint motor of the target joint.
8. The method for estimating the external force of a joint of a heavy-duty industrial robot operating on a large vessel according to claim 7, wherein the robot is allowed to reciprocate a large range of times in an operation space in a non-operation contact state in a static base environment, displacement of a target joint and motor rotation speed of a motor of the target joint during the movement are recorded, then the displacement and motor rotation speed of the joint of the robot are inputted as a model, motor current of the joint is outputted as a model, and a CNN-LSTM model is trained.
9. The method for estimating joint external force of heavy-duty industrial robot working on a large vessel according to claim 1, wherein in the step 2), the joint gravity moment equation is: Wherein, the Is a joint connecting rod coordinate system Coordinates of the z-axis direction vector of (c) under its own coordinates, i.e., [ 0001 ] T ; is from a coordinate system Origin and coordinate system The vector of origin is in the coordinate system Coordinates below; is from a coordinate system To a coordinate system Is a rotation matrix of (a); the mass of the ith connecting rod; Centroid for the ith link in coordinate system Coordinates below; Is an acceleration vector in the robot base coordinate system.
10. A system for estimating external force of a heavy-duty industrial robot joint operating on a large vessel, comprising a memory, a processor and a computer program stored on the memory, wherein the processor executes the computer program to implement the steps of the method for estimating external force of a heavy-duty industrial robot joint operating on a large vessel according to any one of claims 1 to 9.

Description

Method and system for estimating joint external force of heavy-load industrial robot operating on large ship Technical Field The invention relates to a method and a system for estimating joint external force of a heavy-load industrial robot operating on a large ship, and belongs to the technical field of robot motor driving joint control. Background Currently, industrial robots are increasingly used in marine operations such as on-board ship assembly, transportation, and replenishment. However, the operation of the industrial robot on the ship is different from the operation under the land-based static base, the ship is influenced by the environmental factors such as offshore wind, current, wave and the like in the sailing process, the swinging motion of six degrees of freedom such as swinging, pitching, heaving and rolling, pitching and rolling can be generated, the swinging motion of the ship can cause larger disturbance on the driving moment of the joints of the robot, the operation precision and the operation safety of the robot are further influenced, the robot is easy to generate overlarge contact force with an operation object during the assembly and the transportation of the robot on the ship, and the external force perception prediction capability of the heavy-duty industrial robot on the joints during the operation of the ship can provide a technical basis for the safe operation of the heavy-duty industrial robot. Most heavy-duty industrial robots are not provided with joint force sensors, and meanwhile, if the heavy-duty industrial robots which are deployed on a ship and are not provided with the force sensors have joint force sensing capability, the traditional method generally needs to modify the mechanical structure and the electrical hardware of the robot joint to additionally provide the joint force sensors, and combine the dynamic parameter identification and modeling method of the robot to accurately identify the mechanical model of the robot joint, so that the joint external force is predicted to further realize the safe force control in the operation process. The traditional method for installing the joint force sensor has the following defects that the large-range moment sensor required by ① heavy load has larger volume, the complexity and cost of robot design, processing and installation are obviously improved by adding the additional force sensor to the robot joint, even no installation space is available at the joint, the ② manual modeling dynamics process is complex, the parameter identification is complicated, the model is inaccurate, particularly the friction force model in the robot joint transmission chain is still a challenging problem up to now, the ③ robot generally completes production and factory debugging in a land-based static base environment, and the dynamics model obtained by identification under the land-based static base condition is difficult to be directly applied to the marine dynamic base environment. Therefore, how to make the heavy-duty industrial robot predict the joint external force as precisely as possible when working on a ship without changing the mechanical structure and electrical hardware of the joint of the heavy-duty industrial robot is a problem to be solved. Disclosure of Invention The invention aims to provide a method and a system for estimating joint external force of a heavy-duty industrial robot operating on a large ship, which are used for solving the problem of how to predict the joint external force of the heavy-duty industrial robot when operating on the ship under the condition of not changing the mechanical structure and the electrical hardware of the joint of the heavy-duty industrial robot. In order to achieve the above object, the present invention provides a method comprising: The invention discloses a method for estimating joint external force of a heavy-load industrial robot operating on a large ship, which comprises the following steps: 1) Obtaining actual motor current of a target joint motor, motion parameters of a target joint and base acceleration of a moving base of the robot when the moving base contacts the robot during operation; 2) According to the motion parameters, predicting the non-contact current of a non-contact operation state joint motor under the joint static base of the robot; The method comprises the steps of obtaining a dynamic acceleration vector of the joint under a robot base coordinate system according to the base acceleration, bringing the dynamic acceleration vector and a static acceleration vector obtained based on gravity acceleration into a joint gravity moment equation to correspondingly calculate joint gravity moment under a dynamic base and static base state respectively, obtaining power-off current values required by the joint to overcome gravity under the dynamic base and static base state respectively according to the joint transmission coefficient, and obtaining disturbance current of a dynamic