A robot collision detection method, device, electronic device, and storage medium

Abstract

This invention relates to a robot collision detection method, apparatus, electronic device, and storage medium, comprising: a threshold setting unit for pre-setting an environmental stiffness as a judgment threshold based on the object stiffness; a data acquisition unit for real-time acquisition of the contact force between each joint of the robot and the surrounding environment, the position of each joint, and the velocity of each joint; a data conversion unit for converting the real-time acquired position and velocity into Cartesian position and velocity; an estimation unit for estimating the environmental stiffness and environmental damping of the acquired contact force, Cartesian position, and velocity by adjusting a forgetting factor; and a judgment unit for comparing the estimated environmental stiffness with the pre-set environmental stiffness and judging and executing the robot action. This enhances the stability of robot collision detection and judgment, reducing the problem of false alarms.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of robot collision technology, and particularly relates to a robot collision judgment method and device, electronic equipment and storage medium. BACKGROUND

[0002] With the development of science and technology, the development speed of robots is getting faster and faster, which is an important field of contemporary new technology and has a wide range of applications, such as the medical industry, the military industry, the education industry, and production activities. In view of the service life of the robot, when the robot collides with the surrounding environment, it is necessary to reduce the damage to the robot, which requires the robot to have a collision detection function. The existing collision detection schemes include a double-encoder position deviation, acceleration, torque, or current.

[0003] However, the double-encoder position deviation-based collision detection has a flexible reducer itself, and when the robot arm moves at different speeds, even if the robot arm does not collide, a double-encoder deviation will be generated, which is irregular, so it is difficult to determine the collision detection threshold. When the detection threshold is set too small, the robot may automatically start the shutdown protection without collision. When the detection threshold is set too large, the robot may have a very serious collision, but the shutdown protection is not triggered. It is difficult to achieve the protection purpose of the robot body.

[0004] The acceleration-based collision detection is obtained by differentiating the encoder at each joint to obtain the speed of each joint of the robot, and the acceleration of each joint of the robot is obtained by differentiating the speed again. The twice differentiation makes the noise of the acceleration signal very large. Due to the existence of noise, it is difficult to determine the acceleration threshold for collision detection. The problem is the same as the double-encoder. The acceleration signal will be delayed by filtering out the acceleration noise, and this delay will also cause the collision detection to be delayed, and the purpose of protecting the robot body cannot be achieved.

[0005] The problem of torque-based collision detection is that it is difficult to determine the torque threshold. Generally, the torque detection threshold is a fixed value, which means that when we set a very small torque detection threshold (to prevent over-collision), the robot will not be damaged when it collides with an object with a large damping but a small stiffness (such as a piston cylinder), although the torque exceeds the set threshold. Therefore, simply relying on torque cannot simply measure the degree of collision of the robot arm.

[0006] The current-based collision detection is obtained by installing a current sensor to collect the current. The current sensor is generally installed at each joint of the robot, and the external force received by the robot is indirectly measured through the current. When the robot arm collides with the surrounding environment, the current exceeds the set threshold, and it is considered that a collision has occurred. The current signal has a large noise, and the problem is the same as the acceleration-based collision detection. SUMMARY

[0007] To solve the above problems, the purpose of the present application is to provide a robot collision judgment method, device, electronic equipment and storage medium. Because the noise of the external force signal has little effect on the estimation of the environmental stiffness, and because the environmental stiffness generally has a large difference, the signal with a large signal-to-noise ratio will not cause false positives, and for sudden signals, the influence of sudden signals can be reduced through the adjustment of the forgetting factor, that is, although the torque signal suddenly changes, the estimated environmental stiffness will not suddenly change, so the robot collision detection based on environmental stiffness is stable, and because different environmental stiffness (object stiffness) has a fixed data range, the difficulty of setting the threshold value is obviously reduced compared to setting the threshold value based on the double encoder, acceleration and current.

[0008] Specifically, a robot collision judgment method comprises:

[0009] S100: setting the environmental stiffness as a judgment threshold according to the object stiffness; because the threshold value is set through the object stiffness, and the stiffness of the object can be queried, the difficulty of setting the threshold value is greatly reduced.

[0010] The setting of the environmental stiffness as a judgment threshold according to the object stiffness comprises setting at least one environmental stiffness, setting a collision judgment rule, or setting the environmental stiffness as a judgment threshold according to the object stiffness comprises setting at least one environmental stiffness and at least one environmental damping, dividing the collision level, and setting a collision judgment rule.

[0011] S200: collecting the contact force of each joint of the robot and the surrounding environment, the position of each joint, and the speed of each joint in real time; wherein the collected data can be transmitted in the form of torque, double encoder, current, etc., but the present application does not judge the collision based on acceleration, torque, double encoder or current.

[0012] S300: converting the collected position and speed into position and speed in Cartesian coordinates;

[0013] S400: estimating the environmental stiffness and environmental damping of the collected contact force, position and speed in Cartesian coordinates by adjusting the forgetting factor;

[0014] S500: comparing the estimated environmental stiffness with the pre-set environmental stiffness and judging the robot action and performing. The comparing the estimated environmental stiffness with the pre-set environmental stiffness and judging the robot action and performing includes comparing the estimated environmental stiffness with the pre-set at least one environmental stiffness and judging the robot action according to the judgment rule, or the comparing the estimated environmental stiffness with the pre-set environmental stiffness and judging the robot action and performing includes comparing the estimated at least one environmental stiffness with the pre-set at least one environmental stiffness and comparing the estimated environmental damping with the pre-set at least one environmental damping, and judging the robot action according to the judgment rule.

[0015] Further, the real-time acquisition of the contact force around each joint of the robot, the position of each joint, and the speed of each joint further includes real-time acquisition of the acceleration of each joint of the robot; the real-time acquisition of the contact force around each joint of the robot is performed by a pressure sensor, and the real-time acquisition of the position of each joint, the speed of each joint, and / or the acceleration of each joint is performed by a pose sensor.

[0016] Further, the S400 further includes the following steps of adjusting the forgetting factor to estimate the environmental stiffness and the environmental damping of the acquired contact force, the position in Cartesian, and the speed.

[0017] S410: establishing an estimation function of the contact force between the robot and the surrounding environment, as shown in formula (1):

[0018]

[0019] wherein i represents a discrete signal processed, for example, a time interval, i = 0, 1, 2, 3, …. represents the contact force, represents the environmental stiffness, represents the environmental damping. The hat in the formula represents an estimated quantity, x(i) and respectively represent the position and the speed in Cartesian;

[0020] S420: converting the estimated function into matrix (2), and ε(i) matrix (3),

[0021]

[0022]

[0023] S430: recursively estimating the matrix. When the result of inputting the evaluation function as formula (4) is zero, the environmental stiffness and / or the environmental damping are estimated.

[0024]

[0025] wherein, λ is a forgetting factor, used to strengthen the effect of current observation data on parameter estimation, and weaken the previous data, the weight of the previous residual square is exponentially attenuated according to λ in the present application, y(i) is the contact force collected by the robot and the surrounding environment; represents the estimated contact force, that is,

[0026] Further, the recursion is shown in formula (5) and formula (6);

[0027]

[0028]

[0029] wherein, λ is a forgetting factor, used to strengthen the effect of current observation data on parameter estimation; y(i) is the contact force collected by the robot and the surrounding environment.

[0030] Further, the Γ initial matrix is preset as

[0031] Further, the S100 pre-sets the environmental stiffness according to the object stiffness as a judgment threshold, including pre-setting at least one environmental stiffness, setting a collision judgment rule;

[0032] The S500 compares the estimated environmental stiffness with the pre-set environmental stiffness and judges the robot action and executes, including comparing the estimated environmental stiffness with the pre-set at least one environmental stiffness and judging the robot action according to the judgment rule.

[0033] Further, the S100 pre-sets the environmental stiffness according to the object stiffness as a judgment threshold, including pre-setting at least one environmental stiffness and at least one environmental damping, and dividing the collision level, setting a collision judgment rule;

[0034] The S500 compares the estimated environmental stiffness with the pre-set environmental stiffness and judges the robot action and executes, including comparing the estimated environmental stiffness with the pre-set at least one environmental stiffness and comparing the estimated environmental damping with the pre-set at least one environmental damping, and judging the robot action according to the judgment rule.

[0035] Specifically, a device for applying a robot collision judgment method comprises:

[0036] A threshold setting unit is configured to pre-set the environmental stiffness according to the object stiffness as a judgment threshold;

[0037] A collection unit is configured to collect the contact force between each joint of the robot and the surrounding environment, the position of each joint, and the speed of each joint in real time.

[0038] a data conversion unit for converting the real-time collected position and speed into position and speed in Cartesian;

[0039] an estimation unit for estimating the environmental stiffness and the environmental damping of the collected contact force, position and speed in Cartesian by adjusting the forgetting factor;

[0040] a judgment unit for comparing the estimated environmental stiffness with the pre-set environmental stiffness and judging the robot action and executing.

[0041] Specifically, a computer device comprises:

[0042] a memory for storing a processing program;

[0043] a processor for realizing the robot collision judgment method when executing the processing program.

[0044] Specifically, a readable storage medium has a processing program stored thereon, and the processing program realizes the robot collision judgment method when executed by a processor.

[0045] The present application has the advantages of:

[0046] The present application discloses a robot collision judgment method, device, electronic device and storage medium. Since the noise of the external force signal has little effect on the estimation of the environmental stiffness, and because the environmental stiffness generally has large differences, the signal with high signal-to-noise ratio will not cause false positives, and for sudden signals, the influence of the sudden signals can be reduced through the adjustment of the forgetting factor, that is, although the torque signal suddenly changes, the estimated environmental stiffness will not suddenly change, so the robot collision detection and judgment based on the environmental stiffness has high stability. Since different environmental stiffnesses (object stiffnesses) have fixed data ranges, the difficulty of pre-setting the threshold is obviously reduced compared to setting the threshold based on the double encoder, acceleration and current, and setting the threshold based on the environmental stiffness also reduces the problem of false positives. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 is a flowchart of a robot collision judgment method of the present application;

[0048] Figure 2 is a schematic diagram of an experiment for determining the convergence speed ratio of the forgetting factor of the present application;

[0049] Figure 3 is a schematic diagram of an experiment for determining the convergence speed ratio of Γ of the present application;

[0050] Figure 4 is a schematic diagram of a robot collision judgment device of the present application;

[0051] Figure 5 is an embodiment schematic diagram of the computer device of the present application.

[0052] The main element symbols in the embodiments of the present application are as follows:

[0053] Robot collision judgment device-400, threshold setting unit-410, acquisition unit-420, data conversion unit-430, estimation unit-440, judgment unit-450, computer device-500, processor-510, memory-520, storage medium-530, operating system-531, data-532, application program-533, power supply-540, wired or wireless network interface-550, input and output interface-560 DETAILED DESCRIPTION

[0054] The technical solutions of the present application will be described in more detail below in combination with the drawings. The present application includes but is not limited to the following embodiments.

[0055] The present application uses the RLS model, i.e. uses the recursive least square method to perform iterative calculation according to time. This itself is equivalent to a filter, except that the filter is equivalent to the input position and torque to derive the stiffness of the environment. The noise of the external force signal has little effect on the estimation of the stiffness of the environment. For white noise, in general, the signal-to-noise ratio is relatively large, and an environmental stiffness can be directly estimated. Although there is jitter, it does not affect collision detection, because the environmental stiffness generally has a large difference, and a signal with a large signal-to-noise ratio will not cause false positives. For sudden signals, the influence of sudden signals can be reduced through the adjustment of the forgetting factor, that is, although the torque signal is suddenly changed, the estimated environmental stiffness will not be suddenly changed. This can reduce the noise interference that cannot be avoided in collision detection based on torque, speed, etc., making the collision detection more accurate, and making it easier to determine the preset parameters. It also enables the robot controller to distinguish the type of object collided with, making the application scenarios of the robot more flexible.

[0056] The recursive least square method has the phenomenon of data saturation, that is, with the increase of the recursive times, the accumulation of the old data is more and more, the information amount is more and more, so that the new data cannot be embodied in the numerous data and the information of the new data cannot be embodied, which leads to meaningless parameter estimation and cannot obtain the required information. Therefore, in the process of recursion, the weight of the old data and the new data needs to be adjusted, the old data is faded to reduce the influence of the old data on the new data, and the new data is strengthened to increase the effect of the new data, and the weight parameter of the fading of the old data and / or the strengthening of the new data is the forgetting factor. The forgetting factor has the characteristics of fast convergence speed, appropriate forgetting factor can make the system reach the expected target faster, and strong tracking ability, in the random input function, the fluctuation is relatively small. In order to reflect the time-varying nature of the parameter, it is obvious that the effect of the current observation data on the parameter estimation should be strengthened, and the influence of the previous observation data should be weakened. In the present application, the forgetting factor strengthens the weight of the current data in the process of iteration, so that the weight of the current data is always large. In the evaluation function of the present application, the weight of the previous residual square is attenuated according to the lambda index, so as to weaken the influence of the old data on the new data, strengthen the information of the new data, and facilitate the estimation of the evaluation function.

[0057] In order to control the robot, it is necessary to use the robot dynamics to study the relationship between motion and force. At present, there are many kinds of dynamics methods, such as Lagrange method. The derived dynamics equation is converted to Cartesian space to facilitate the design of the controller. After the robot contacts with the environment, it is subjected to the action force of the environment, and the robot is no longer an independent control object, but is integrated with the environment into a new system.

[0058] The kinematics equation is converted to Cartesian space to obtain the position and velocity in Cartesian space. The dynamics equation formula of the present application is: Wherein, tau d = J T (q) f ext , wherein, M (q), G (q), are the inertia matrix, the coriolis force matrix, the gravity matrix and the friction matrix of the robot respectively. q, are the position, velocity and acceleration of the joints of the robot respectively. tau d is the external force acting on the joints of the robot, J T is the transpose of the Jacobian matrix of the robot (its function is to convert the external force to the joints of the robot), f ext is the external force applied by the environment on the robot. The equation for converting the dynamics equation to Cartesian space is: Wherein, x, are the position, velocity and acceleration of the robot respectively, M d (q), G d(q), I, C, G, F are inertia matrix, coriolis matrix, gravity matrix, friction matrix in cartesian space respectively.

[0059] The application discloses a robot collision judgment method, comprising:

[0060] S100: set the environmental stiffness as a judgment threshold according to the object stiffness. The threshold is set through the object stiffness, and the stiffness of the object can be inquired, so that the difficulty of threshold setting is greatly reduced.

[0061] Table 1 is an object stiffness table:

[0062] Parameter Sponge Foamed sponge Rubber Wood Iron Stiffness (N / m) 260 960 2400 2500 20000

[0063] It can be known from Table 1 that the stiffness of the object is very different. For example, the stiffness of a sponge is close to 260 N / m, the stiffness of a foamed sponge is about 960 N / m, the stiffness of wood is about 2500 N / m, and the stiffness of iron material is about 20000 N / m.

[0064] The threshold is preset according to the known material stiffness, compared with the threshold preset of the collision detection based on the moment, speed and the like, the threshold preset in the application has practical significance, the difficulty of threshold setting is reduced, and the collided object type can be distinguished, and corresponding actions can be made.

[0065] For example, the stiffness of wood 2500 N / m is taken as the preset judgment threshold, or for example, the stiffness of wood 2500 N / m and the stiffness of iron material 20000 N / m are taken as the preset judgment threshold.

[0066] The threshold is preset according to the known material stiffness, compared with the threshold preset of the collision detection based on the moment, speed and the like, the threshold preset in the application has practical significance, the difficulty of threshold setting is reduced, and the collided object type can be distinguished, and corresponding actions can be made.

[0067] S200: real-time collection of contact forces of each joint of the robot and the surrounding environment, positions of each joint, and speeds of each joint. The application further comprises real-time acquisition of accelerations of each joint of the robot for constructing a dynamic formula. The contact forces of each joint of the robot and the surrounding environment are collected in real time through a pressure sensor, the positions of each joint, the speeds and / or accelerations of each joint are collected in real time through a pose sensor, and the collected data can be transmitted in the form of a moment, a double encoder or a current, but the application does not perform collision judgment based on the accelerations, the moment, the double encoder or the current.

[0068] S300: converting the collected positions and speeds into positions and speeds in cartesian.

[0069] S400: Estimate the environmental stiffness and environmental damping of the collected contact force, position and velocity in Cartesian by adjusting the forgetting factor. Specifically, the following steps are included:

[0070] S410: Establish the estimation function of the contact force between the robot and the surrounding environment as shown in equation (1):

[0071]

[0072] Wherein, i represents the discrete signal processed, for example, time interval, i = 0, 1, 2, 3, …. represents the contact force, represents the environmental stiffness, represents the environmental damping. The hat in the formula represents the estimated quantity, x(i) and respectively represent the position and velocity in Cartesian.

[0073] S420: Convert the estimated function into matrix (2) and ε(i) matrix (3) for estimating the environmental stiffness and environmental damping.

[0074]

[0075]

[0076] S430: Recursively estimate the matrix, when the recursive result is input into the evaluation function as shown in equation (4). When the evaluation function result is zero, the environmental stiffness and / or environmental damping is estimated.

[0077]

[0078] Wherein, λ is the forgetting factor, which is used to strengthen the effect of the current observation data on the parameter estimation and weaken the previous data. In the present application, the weight of the previous residual square is exponentially attenuated according to λ, and y(i) is the collected contact force between the robot and the surrounding environment. represents the estimated contact force, that is,

[0079] Wherein the recursion is shown in equation (5) and equation (6):

[0080]

[0081]

[0082] Wherein, λ is the forgetting factor, which is used to strengthen the effect of the current observation data on the parameter estimation; y(i) is the collected contact force between the robot and the surrounding environment.

[0083] As attached Figure 2 The diagram illustrating the convergence speed of RLS with respect to λ (where the horizontal axis represents time and the vertical axis represents object stiffness) shows that the smaller the value of λ, the shorter the time to reach the same estimated stiffness, i.e., the faster the convergence speed. This is because the current joint position velocity and force have a greater impact on the algorithm, leading to faster convergence. It can be seen that when λ is greater than 1, the accuracy of the algorithm decreases. Therefore, the range of λ is (0,1), i.e., greater than 0 and less than 1. It should also be noted that when λ is too small, the system's ability to filter noise also decreases. Therefore, based on the comprehensive experimental results, this invention uses λ values ​​of 0.6, 0.7, 0.8, or 0.9.

[0084] When λ takes the value 0.9, as shown in the attached figure. Figure 3 The diagram illustrating the influence of the Γ eigenvalue on the RLS convergence speed (where the horizontal axis represents time and the vertical axis represents object stiffness) shows that a larger Γ indicates a shorter time to reach the same estimated stiffness, i.e., a faster convergence speed. However, when Γ is too large, the line segments in the attached diagram are not smooth and continuous, reflecting a decrease in the algorithm's stability when Γ is too large. In this case, the presence of noise, such as torque noise, can lead to instability in the RLS model, affecting the estimation results. Therefore, based on the experimental results, this invention uses a Γ value of 10. 5 Or 10 4 When the initial value of Γ is set to 0, due to the forgetting factor, this value will not significantly affect the algorithm itself, but it can speed up the convergence of the next iteration. Therefore, the initial matrix of Γ in this invention is preset to...

[0085] S500: comparing the estimated environmental stiffness with the pre-set environmental stiffness and determining the robot action and executing. Wherein, the S100 pre-sets the environmental stiffness as the determination threshold according to the object stiffness includes pre-setting at least one environmental stiffness and setting the collision determination rule. The S500 comparing the estimated environmental stiffness with the pre-set environmental stiffness and determining the robot action and executing includes comparing the estimated environmental stiffness with the pre-set at least one environmental stiffness and determining the robot action according to the determination rule. One embodiment is to set the stiffness of wood as 2500 N / m as the pre-set determination threshold, and set the collision determination rule as: when the stiffness of the collided object is greater than or equal to 2500 N / m, it is determined that the collided object is hard, and the robot stops working; when the stiffness of the collided object is less than 2500 N / m, it is determined that the collided object is not hard, and the robot continues to work. Another embodiment is to set the stiffness of wood as 2500 N / m and the stiffness of iron material as 20000 N / m as the pre-set determination threshold, and set the collision determination rule as: when the stiffness of the collided object is greater than or equal to 20000 N / m, it is determined that the collided object is hard, and the robot immediately stops working. When the stiffness of the collided object is greater than or equal to 2500 N / m and less than 20000 N / m, it is determined that the collided object is relatively hard, which may damage the robot, and further frequency determination is performed; if the robot collides with the object with the stiffness in the interval for more than 5 times within the pre-set time, for example, 3s, 3.5s, 4s, the robot stops working; if the robot collides with the object with the stiffness in the interval for less than 5 times within the pre-set time, for example, 3s, the frequency is cleared, and the robot continues to work. When the stiffness of the collided object is less than 2500 N / m, it is determined that the collided object is not hard, and the robot continues to work.

[0086] In addition, comparing the estimated environmental stiffness with the pre-set environmental stiffness and determining the robot action and executing also includes that the S100 pre-sets the environmental stiffness as the determination threshold according to the object stiffness includes pre-setting at least one environmental stiffness and at least one environmental damping, and dividing the collision level, and setting the collision determination rule. The S500 comparing the estimated environmental stiffness with the pre-set environmental stiffness and determining the robot action and executing includes comparing the estimated at least one environmental stiffness with the pre-set at least one environmental stiffness and comparing the estimated environmental damping with the pre-set at least one environmental damping, and determining the robot action according to the determination rule.

[0087] One embodiment is to set the stiffness of iron material as about 20000 and the damping as about 280 as the pre-set determination threshold, and set the collision determination rule as: when the stiffness of the collided object is greater than or equal to 20000 and the damping is greater than or equal to 280, it is determined that the impedance of the collided object is large, and the robot stops working; when one of the stiffness and the damping of the collided object is less than the pre-set determination threshold, it is determined that the impedance of the collided object is not large, and the robot continues to work.

[0088] AsFigure 4 As shown, based on the same concept, the present application also provides a robot collision judging device 400, which comprises:

[0089] a threshold setting unit 410 for setting the environmental stiffness as a judging threshold according to the object stiffness in advance;

[0090] a collecting unit 420 for collecting the contact force between each joint of the robot and the surrounding environment, the position of each joint, and the speed of each joint in real time;

[0091] a data conversion unit 430 for converting the collected position and speed into the position and speed in Cartesian;

[0092] an estimating unit 440 for estimating the environmental stiffness and environmental damping of the collected contact force, position and speed in Cartesian by adjusting the forgetting factor;

[0093] a judging unit 450 for comparing the estimated environmental stiffness with the environmental stiffness set in advance and judging the robot action and executing.

[0094] As shown, Figure 5 As shown, based on the same concept, the present application also provides a computer device 500, which can have a big difference due to different configurations or performances, and can comprise one or more than one processor (central processing units, CPU) 510 (for example, one or more than one processor) and a memory 520, one or more than one storage medium 530 (for example, one or more than one mass storage device) for storing application programs 533 or data 532. Wherein, the memory 520 and the storage medium 530 can be temporary storage or persistent storage. The programs stored in the storage medium 530 can comprise one or more than one module (not shown in the figure), and each module can comprise a series of instruction operations in the computer device 500. Further, the processor 510 can be arranged to communicate with the storage medium 530, and execute the series of instruction operations in the storage medium 530 on the computer device 500.

[0095] The computer device 500 can also comprise one or more than one power supply 540, one or more than one wired or wireless network interface 550, one or more than one input and output interface 560, and / or one or more than one operating system 531, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc.

[0096] Those skilled in the art can understand, Figure 5The illustrated computer device architecture is not meant to imply architectural limitations but is intended to represent various architec tures that can be utilized for a computer device. For example, a computer device can have more or fewer components than shown, can lay out such components differently, or can have different, additional, or missing components altogether.

[0097] The computer readable instructions, when executed by the processor, cause the processor to perform the following steps: setting the environment stiffness as a judgment threshold according to the object stiffness in advance; collecting the contact force of each joint of the robot and the surrounding environment, the position of each joint, and the speed of each joint in real time; converting the collected position and speed into the position and speed in Cartesian; estimating the environment stiffness and the environment damping of the collected contact force, the position and speed in Cartesian by adjusting the forgetting factor; comparing the estimated environment stiffness with the pre-set environment stiffness and judging the robot action and performing.

[0098] In one embodiment, a readable storage medium is provided, and the computer readable instructions, when executed by one or more processors, cause the one or more processors to perform the above steps, and the specific steps are not described here.

[0099] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which are not described here.

[0100] The integrated unit, if realized in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0101] The above-described and above-embodied examples are only used to illustrate the technical solutions of the present application, rather than limit them. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A robot collision determination method characterized by comprising: Comprising: S100: setting the environmental stiffness as a judgment threshold according to the object stiffness in advance; S2 S00: collecting the contact force between each joint of the robot and the surrounding environment, the position of each joint, and the speed of each joint in real time; S300: converting the collected position and speed into the position and speed in Cartesian; S400: estimating the environmental stiffness and the environmental damping of the collected contact force, the position in Cartesian, and the speed by adjusting the forgetting factor; S500: comparing the estimated environmental stiffness with the environmental stiffness set in advance and judging the robot action and executing; The S400 further comprises the following steps: S410: establishing the estimation function of the contact force between the robot and the surrounding environment as shown in formula (1); （1） where i denotes the discrete signal being processed, i = 0, 1, 2, 3,...; represents the contact force, represents the environmental stiffness, represents the environmental damping, where ^ denotes the estimated quantity, and represent the position and velocity in Cartesian coordinates, respectively; S420: convert the estimated function into matrix (2), matrix (3), （2） （3） S430: estimating the environment stiffness and / or the environment damping based on the estimated matrix recursion, when the result of the recursion is input into the evaluation function as the result of formula (4), the environment stiffness and / or the environment damping is estimated; （4） wherein, is a forgetting factor, used to reinforce the effect of current observation data on parameter estimation and to weaken previous data, in the present aspect the weight of previous residual square is eliminated according to exponential decay, is the contact force between the robot and the surrounding environment collected; represents the estimated contact force, i.e. .

2. The robot collision determination method according to claim 1, characterized in that, The real-time collection of the contact force between each joint of the robot and the surrounding environment, the position of each joint, and the speed of each joint further comprises real-time acquisition of the acceleration of each joint of the robot; the real-time collection of the contact force between each joint of the robot and the surrounding environment is realized by a pressure sensor, and the real-time collection of the position of each joint, the speed of each joint, and / or the acceleration is realized by a pose sensor.

3. The robot collision determination method according to claim 1, characterized in that, The recursion is shown in formula (5) and formula (6); （5） （6） wherein, is a forgetting factor to enforce the effect of current observation data on the parameter estimation; is the contact force of the robot with the surrounding environment.

4. The robot collision determination method according to claim 3, wherein The The initial matrix is preset as .

5. The robot collision determination method of claim 1, wherein The S100 comprises setting at least one environmental stiffness and at least one environmental damping, dividing the collision level, and setting the collision judgment rule; The S500 comprises comparing the estimated at least one environmental stiffness with the preset at least one environmental stiffness and comparing the estimated environmental damping with the preset at least one environmental damping, and judging the robot action according to the judgment rule.

6. The robot collision determination method of claim 1, wherein The S100 comprises setting at least one environmental stiffness and at least one environmental damping, dividing the collision level, and setting the collision judgment rule; The S500 comprises comparing the estimated at least one environmental stiffness with the preset at least one environmental stiffness and comparing the estimated environmental damping with the preset at least one environmental damping, and judging the robot action according to the judgment rule.

7. An apparatus for applying a robot collision determination method according to any one of claims 1 to 6, characterized in that, Comprising: A threshold setting unit for setting the environmental stiffness as a judgment threshold according to the object stiffness in advance; A collection unit for collecting the contact force between each joint of the robot and the surrounding environment, the position of each joint, and the speed of each joint in real time; A data conversion unit for converting the collected position and speed into the position and speed in Cartesian; An estimation unit for estimating the environmental stiffness and the environmental damping of the collected contact force, the position in Cartesian, and the speed by adjusting the forgetting factor; A judgment unit for comparing the estimated environmental stiffness with the environmental stiffness set in advance and judging the robot action and executing.

8. A computer device, comprising: Comprising: A memory for storing a processing program; A processor for implementing the robot collision judgment method according to any one of claims 1 to 6 when executing the processing program.

9. A readable storage medium, characterized by, The readable storage medium has a processing program stored thereon, and the processing program is executed by the processor to implement the robot collision judgment method according to any one of claims 1 to 6.

Images