WO2022229785A1

WO2022229785A1 - Method and related system for safely controlling a robot

Info

Publication number: WO2022229785A1
Application number: PCT/IB2022/053632
Authority: WO
Inventors: Andrea Maria ZANCHETTIN
Original assignee: Politecnico Di Milano
Priority date: 2021-04-26
Filing date: 2022-04-19
Publication date: 2022-11-03
Also published as: IT202100010472A1

Abstract

This disclosure relates to a method of controlling at least one robot collaborating with a human operator, said method being implementable by means of a control system comprising at least one detection device, configured to detect a position of the operator's body with respect to a mobile part of the robot, a memory, a microprocessor control unit configured to process data stored in the memory and to control the movements of a collaborative robot. The method of this disclosure can be implemented by software run from a microprocessor unit.

Description

METHOD AND RELATED SYSTEM FOR SAFELY CONTROLLING A

ROBOT

TECHNICAL FIELD

The present disclosure relates to robot control methods and more particularly a method for controlling a robot intended to cooperate safely with a human operator, as well as a robot control microprocessor system and a related computer program.

BACKGROUND

With the development of robotics, it is foreseeable that in the future there will be more robots made to collaborate with human operators, in order to help operators in the execution of their work. In particular, robots can conveniently be used to quickly perform tasks in a substantially mechanical way, while the most difficult operations to be performed in complete autonomy by robots can be assigned to human operators.

Thanks to the improved motion control procedures of industrial robots, it was possible to create robots capable of autonomously performing tasks by sharing a workspace with one or more human operators. This solution allows the coexistence of a robot with a human operator in the same environment, in which, however, each of the two subjects must complete their tasks independently of the other, reaching their respective objectives.

However, this type of subdivision is not always feasible when a job to be performed is composed of both operations that require human intelligence, and mechanical operations, which can be performed individually without knowing all the work to be done and the purposes to be achieved. In general, in all those jobs where there are operations that require a high level of understanding, which human operators can easily perform but which robots cannot easily perform unless sophisticated and expensive technology is used, which would require considerable investments and would increase production costs, entrusting the entire work to a human operator appears indispensable.

To overcome this limitation, collaborative robots are created, sometimes called with the neologism "cobot" or "co-robot", which can collaborate with human operators sharing the same workspace and interacting closely and synergistically with them to reach a desired result. Since robots have to work very close to a human operator, the problem of operator safety arises. A technique for ensuring safety during operations in which robots collaborate with human operators, considered in the ISO/TS 15066 technical specification which illustrates the safety requirements for industrial robotic systems, is the technique called "Speed and Separation Monitoring" (or more briefly SSM). This technique provides that at any time the robot actuators can decelerate to eventually stop completely before violating a minimum separation distance dmin from the human operator, i.e. that the distance d between the robot and the human operator is always greater than dmin".

as schematically illustrated in figure 1.

In general, the positions of the robot and of the operator are periodically observed in certain instants of observation by means of at least one detection device, so that in these instants of observation the positions of the robot and of the operator are identified. However, it is necessary to control the position and speed of the robot even in the interval that passes between two consecutive observation instants, as well as to predict its behavior. To this end, a sampling interval of duration Tc is established and at each sampling instant the position and speed of the robot are estimated by means of an algorithm. Consequently, between two consecutive observation instants in which the detection devices provide the position of the robot and of the operator, there will be one or more sampling instants in which the speed and position of the robot will be estimated.

Given the robot position Sk on a path given at the k-th sampling instant, it is possible to estimate the position Sk+1 and the speed Vk+i at the future sampling instant k + 1 as a function of the acceleration a, of the velocity Vk, of the current position Sk and of the sampling period Tc that elapses between one sampling and the next with the following formulas:

In the SSM technique, the distance between the robot and the operator is calculated as a function of the position that the robot will have at the future sampling instant k + 1, as well as at the subsequent sampling instants, and it is verified if the condition

is satisfied. In practice, a simulation is performed to check whether it is possible to stop the robot in time respecting the minimum safety distance from the human operator d_min applying the minimum possible acceleration a_min: if this occurs, it is considered as acceleration a the maximum possible value U_max, otherwise the minimum value a_min is considered to predict the position of the robot at the future sampling instant k + 1 and the robot speed reduction is commanded.

The distance between the robot and the human operator, however, also depends on the movements that the operator makes to perform his work, that is, it depends on the Ok+1 position that the operator will assume in the future sampling instant k + 1. However, this position is not predictable and for this reason, according to the current technique, the Ok position is considered at the present sampling instant k to estimate the future distance dk+i.

SUMMARY

However, the Ok position of the operator at the present instant is not a reliable estimate of the Ok+1 position that will actually be assumed by the operator at the future sampling instant k + 1. Furthermore, the human operator can move in an unpredictable way so that the current SSM technique described above is unsatisfactory.

An object of the present disclosure is to provide a method of controlling an industrial robot used in collaborative applications. The control method of the present disclosure can be implemented by means of a control device comprising detection devices which detect the position of the operator in the workspace, a memory in which to store data of speed and position of the robot as well as data of the position of the human operator in a fixed reference system of the control device, as well as a microprocessor control unit that processes such data stored in the memory according to the method of this disclosure to predict whether the robot in the next sampling instant k + 1 will be at a distance less than the minimum safe distance or not, and to adjust the speed and acceleration of the robot accordingly based on this prediction information.

The method makes use of a technique that has proved to be extraordinarily effective for taking into account the position of the moving operator, updating the comparison threshold Ck+i with which to compare the estimated distance value dk+i between the robot and the human operator, in which the comparison threshold Ck+i is updated iteratively at each sampling period on the basis of the comparison threshold value C_k at the previous cycle (starting from d_min for k = 0) and on the basis of a maximum speed of movement of the operator.

The method of this disclosure can be implemented by means of software installed in a microprocessor unit of a relative control system of at least one robot collaborating with a human operator.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows an example of a workstation of a human operator flanked by a robot that cooperates with the operator.

Figure 2 is a flowchart illustrating a procedure for determining speed and position of the robot and operator at sampling instant k + 1, employed in the control method of this disclosure.

DETAILED DESCRIPTION

To improve the performance of execution of a job that requires the synergistic cooperation of a human operator with a collaborative robot ("cobot"), a control method has been devised that avoids that the distance between the collaborative robot and the human operator can be reduced below a minimum nominal value d_min due to movements of the operator.

To identify the position of the human operator, for example, a point corresponding to an area of the operator's body, or an area that extends indefinitely in the vertical direction, or a defined volume within which the operator acts.

In the method according to the present disclosure, the future position that the operator will assume at the sampling instant k + 1 is taken into account by updating the comparison threshold Ck+i of the estimated distance dk+i between the robot and the human operator at each sampling period iteratively, on the basis of the comparison threshold Ck estimated at the previous sampling period. Initially (k = 0), the position so of the robot and the position oo of the operator are detected and the comparison threshold co is equal to the minimum distance d_min. Then, for the subsequent sampling instants up to a new observation instant, proceed as follows.

According to one aspect, the most unfavorable case arises in which the operator between the sampling instants k and k + 1 moves in the exact opposite direction to the direction of motion of the collaborative robot. In this situation, the operator due to his own motion will reduce the distance with respect to the robot by a length equal to the product between its speed ov and the sampling period Tc. By indicating with ovmax the maximum speed at which the human operator is likely to be able to move, we can consider a maximum movement of the human operator in the direction of the robot equal to ovmax * Tc. According to the method of this disclosure, this shift is taken into account by updating the comparison threshold C_k+1 in the way that will be illustrated below.

A flow diagram of a procedure for estimating the position and speed of a robot in the reference system of the control device is shown in Figure 2. This flow diagram was obtained under the assumption that the maximum speed ovmax at which an operator can move human is 1.6 m/s, which is an exemplary speed value suggested by the ISO 13855 standard currently in force, for which the maximum movement that the operator can make during a sampling period Tc is equal to 1.6 * Tc. Indicating with C_k the value of the comparison threshold at the k-th sampling instant, the new comparison threshold at the sampling instant k + 1 will be greater than the threshold C_k by a value corresponding to the maximum speed, for example the product of the velocity maximum ovmax and tire sampling period Tc. In practice, we consider the worst case in which at each sampling interval the human operator is moving at its maximum speed ovmax in the direction of the robot.

As in the classic SSM technique, in step 1 a first simulation is carried out to verify if it is possible to stop the robot on the assigned path by applying the maximum acceleration a_max for a single sampling interval and then repeatedly the minimum acceleration a_min, which is evaluated iteratively on the basis of the corresponding state of motion, until the stop state is reached. If, in this condition, it is possible to stop the robot without violating, at any moment, the condition on the minimum separation distance, the maximum acceleration value a_max is chosen, otherwise the minimum acceleration value is chosen a_min.

According to one aspect, the nominal minimum a_min and maximum a_max accelerations are determined on the basis of the technical characteristics of the actuators that move the robot. For example, such maximum a_max and minimum a_min accelerations can be determined by the maximum and minimum torques that can be developed by the actuators, and / or by the maximum power of the actuators, etc.

Assuming that in step 1 the maximum acceleration a_max has been chosen, in step 2 we estimate the position Sk+i and the speed Vk+1 that would be reached by the robot at the sampling instant k + 1, on the basis of the position S_k and of the speed V_k at the current sampling instant k.

Unlike the SSM technique, a comparison threshold Ck+i is estimated with which to compare the distance k + 1 by increasing the comparison threshold C_k estimated at the previous sampling instant k, starting from d_min for k = 0, by an equal quantity to the product of the maximum speed ovmax at which a human operator can move and the duration of the sampling period Tc. In the exemplary case illustrated in Figure 2, this maximum speed has been established as 1.6 m/s, for which the comparison threshold Ck+i is determined by increasing the minimum distance by a term proportional to the maximum speed ovmax of a human operator:

If the robot is expected to stop ( V_k+1 = 0) (step 3), then the robot can be controlled to move it with a maximum travel acceleration a_max. If, on the other hand, at the sampling instant k + 1, or at one of the following, the robot will be in motion, then it is necessary to verify (step 4) whether the estimated position S_k+1 of the robot will go beyond a maximum distance L, in which case it is necessary to check the robot with the minimum acceleration a_min.

If the robot is expected to be moving at the sampling instant k + 1 (step 3: Vk+i> 0) and is not leaving the assigned path going beyond a predetermined maximum distance L (step 4: Sk+1 <L), according to the method of this disclosure it is estimated (step 5) the distance dk+i that there will be between the robot and the operator at the next sampling instant k + 1 on the basis of the estimated future position Sk+i of the robot and the current Ok position of the human operator as detected by the detection devices.

Instead of comparing the estimated distance dk+1 with the minimum safe distance from the human operator d_min, according to the method of this disclosure it is verified (step 6) whether the future estimated distance dk+i is lower than the comparison threshold Ck+i, in which case the robot is controlled with the minimum acceleration a_min, otherwise (step 7) a new minimum acceleration value a_min is determined, at which the robot can be moved, depending on the position Sk+i and the speed Vk+i estimate k + 1 at the future sampling instant. If this new minimum acceleration value a_min exceeds (step 8) the maximum acceleration value a_max, the previous minimum acceleration value will be maintained, otherwise the robot is checked to give it (he speed Vk+i and the determined acceleration, and iteratively repeats the algorithm for future cycles k + i (step 9) using for each cycle k + i the values established in the previous cycle k+i-1, in order to estimate the position and speed of the robot in several sampling instants between two consecutive observation instants.

According to one aspect, it is possible to introduce a maximum limit Vmax to the speed V_k at which the robot can be moved. In this case, the maximum acceleration value a_max that can be imparted to the robot must also be adjusted so that at the end of the upcoming sampling period the robot has not assumed a speed that exceeds the maximum limit Vmax. To this end, the maximum acceleration that can actually be imparted to the a_maxeff robot will be the minimum value between the maximum nominal a_max acceleration and the ratio between: the difference between the maximum limit Vmax and the speed Vk, and the duration To of the period sampling.

At the next observation instant, the position of the human operator and the position of the robot so are detected and the algorithm described above is repeated from the beginning.

Thanks to the method of this disclosure, the robot is controlled in such a way as to always maintain a minimum distance from the human operator, even if the latter were to make sudden movements.

The method of this disclosure can be implemented through a control system comprising at least one detection device configured to detect a position of the operator's body with respect to a moving part of the robot, a memory, and a microprocessor control unit which runs a computer program having software code installed in the microprocessor unit for executing the method, configured to process data stored in the memory and to control movements of the collaborative robot.

Claims

1. A control method of at least one collaborative robot with a human operator, said method being implementable by means of a control system comprising at least one detection device, configured to detect a position of the operator's body with respect to a moving part of the robot, a memory, a microprocessor control unit configured to process data stored in the memory and to control movements of a collaborative robot, the method comprising detecting a position (oo)of the operator's body with respect to at least a moving part of the robot in a first instant of observation with said at least one detection device, then cyclically performing the following operations at each sampling step (k) up to a second observation instant subsequent to said first observation instant:

1) determining an estimated comparison threshold (ck) for a current sampling step (k);

2) determining a current speed (v_k) and a current position (s_k) of the moving part of the robot at the current sampling step (k);

3) estimating an estimated speed (v_k+1) and an estimated position (s_k+1) of the moving part of the robot at the next sampling step (k + 1) as a function of a desired acceleration to be imparted to the moving part of the robot;

4) estimating an estimated distance (dk+i) between the estimated position (s_k+1) and the position (ok) of the operator's body;

5) determining an estimated comparison threshold ( C_k+1) for the next sampling step (k + 1) as a function of a nominal safety distance (d_min), of a maximum estimated speed of movement (ovmax) of the human operator and of the comparison estimated threshold (c_k) for said current sampling step (k);

6) if said estimated distance (dk+i) is less than an estimated comparison threshold (C_k+1), checking said moving part of the robot by giving the moving part a minimum nominal acceleration (a_min), otherwise carry out the following operations from 7) to 9);

7) estimating a new nominal minimum acceleration value (a_min) for the next sampling step (k + 1) as a function of the estimated position (s_k+1) and the estimated speed (v_k+1);

8) if said new nominal minimum acceleration value (a_min) exceeds a nominal maximum acceleration value (a_max), determining the new nominal minimum acceleration value (a_min) for the next sampling step (k + 1) equal to the value of minimum nominal acceleration (a_min) for the current sampling step (k); 9) controlling the robot to impart said desired acceleration to the moving part of the robot.

2. The method according to claim 1, wherein said comparison estimated threshold (ci) for a first sampling step (k = 1) is determined as the sum of said nominal safety distance (drain) and the product of said maximum estimated speed of movement (ovmax) of the human operator for a duration of a sampling period (Tc), and said comparison estimated threshold (C_k+1) for said subsequent sampling step (k + 1) is determined as the sum of the estimated threshold comparison (ck) to said current sampling step (k) and of the product between said maximum estimated speed of movement (ovmax) of the human operator for the duration of a sampling period (Tc).

3. The method according to one of the preceding claims, comprising the operations of: if said estimated speed (v_k+1) exceeds a nominal maximum speed (vmax), controlling the robot to move said moving part of the robot with said maximum nominal speed (vmax); controlling the robot to impart to said moving part of the robot a maximum effective acceleration (a_maxeff) as the minimum value between a maximum rated acceleration (amax) and the ratio between: the difference between said maximum rated speed (vmax) and the current speed (vk), and a duration of a sampling period (Tc).

4. The method according to one of the preceding claims, wherein said minimum nominal acceleration (amin) and said maximum nominal acceleration (amax) are determined on the basis of technical characteristics of actuators that move the robot.

5. The method according to one of the preceding claims, wherein said position (ok) of the operator's body is a position of a point of an area of the body of the human operator, or alternatively it is a position of a point of a surface intended to be occupied by the human operator, or alternatively it is a point of a volume intended to be occupied by the human operator.

6. A computer program loadable into a memory of a microprocessor unit, comprising a software code configured to operate the microprocessor unit to perform operations of the method of one of claims 1 to 5 when the software code is executed by the microprocessor unit.

7. A control system of at least one collaborative robot with a human operator, comprising: at least one detection device configured to detect a position of the operator's body with respect to a moving part of the robot, a memory, a microprocessor control unit configured to process data stored in the memory and to control movements of a collaborative robot, characterized in that a computer program of claim 6 is loaded into said microprocessor control unit for controlling the movements of a collaborative robot by carrying out the method of one of claims 1 to 5.