EP3380280A1

EP3380280A1 - Method and device for controlling a robot during co-activity

Info

Publication number: EP3380280A1
Application number: EP16805053.2A
Authority: EP
Inventors: Christine CHEVALLEREAU; Alexis GIRIN; Philip LONG
Original assignee: Centre National de la Recherche Scientifique CNRS; Universite de Nantes; Institut de Recherche Technologique Jules Verne
Current assignee: Centre National de la Recherche Scientifique CNRS; Universite de Nantes; Institut de Recherche Technologique Jules Verne
Priority date: 2015-11-26
Filing date: 2016-11-28
Publication date: 2018-10-03
Also published as: US20190030716A1; WO2017089623A1; US11192251B2

Abstract

The invention concerns a method for controlling the operation of a robot within a system comprising said robot (100) and means for analysing the concentric environment of same comprising: i. a contact sensor; ii. a proximity sensor; iii. a vision and location sensor; characterised in that it comprises the steps consisting of: a. obtaining, for each of the axes of the robot, a maximum allowable force value; b. if the force on one of the axes of the robot is greater than the maximum value acquired in step a) (310), stopping (311) the robot in the position of same; c. obtaining a concentric monitoring space, referred to as the safe space, the area of which depends on the speed of the robot; d. monitoring the environment of the robot, by the analysis means; e. if the intrusion of an object is detected (320) in the safe space of the robot, gradually decreasing (330) the manoeuvring speed to a speed referred to as the safe speed; f. repeating step a).

Description

METHOD AND DEVICE FOR DRIVING A ROBOT INTO

ACTIVITY

The invention relates to a method and a device for controlling a robot in co-activity. The invention relates to the field of robots operating in an environment in co-activity with human operators or with other robots, or cobots. The invention is more particularly, but not exclusively, dedicated to the field of handling and assembly in the automotive, aeronautical and naval industries.

According to the prior art, a robot capable of working in co-activity, in particular with human operators, comprises several safety devices used alone or in combination. For example, such a robot comprises force sensors on its various axes and means for causing a safety stop of the robot when the force measured on one of these sensors exceeds a threshold value, for example 150 N. After a safety stop, the robot must be reset to resume normal operation.

According to another embodiment used in addition to the previous one, the robot evolves according to a so-called security speed. This safety speed is sufficiently reduced to both allow a possible operator to easily anticipate the movements of the robot and thus avoid the collision, and on the other hand, not to hurt the operator if ever such a collision happened despite everything.

These speed and force limits are defined in particular in ISO 10218 and ISO TS 15066.

These solutions of the prior art have the disadvantage of greatly reducing the productivity of the robot. In addition, acceptable stress limits are statically defined. Thus, the force limit defined to detect a collision with a static operator is not necessarily relevant if the collision occurs with a person in motion, such a collision being equally likely to have other consequences, for example a collision. fall, which generate indirect trauma. In other circumstances, a low contact force, less than 150 N, but prolonged, for example at the level of the larynx or the cage thoracic is likely to cause significant trauma. Also, although these systems of the prior art greatly reduce the productivity of the robot, they do not provide total security of operation.

The invention aims at solving the disadvantages of the prior art and for this purpose concerns a method for controlling the operation of a robot within a system comprising said robot and means for analyzing its environment, in particular concentric, comprising:

a contact sensor;

i. a proximity sensor;

iii. a vision and location sensor;

which method comprises the steps of:

at. obtain for each axis of the robot a value of maximum allowable effort;

b. if the force on one of the axes of the robot is greater than the maximum value acquired in step a) stop the robot in its position; vs. to obtain a surveillance space, called a security space, the extent of which depends on the speed of the robot;

d. monitor the environment of the robot by means of analysis;

e. if the intrusion of an object is detected in the security space of the robot, gradually reduce the speed of evolution to a so-called security speed;

f. resume in step a).

Thus, the robot continuously scans its environment and operates in safety mode, at reduced speed, only if the presence of an obstacle constituting a risk of collision occurs in this environment. The space monitored being a function of the speed of the robot, the behavior of the robot vis-à-vis the risk of collision depends on its working conditions, including its speed of execution.

The invention is advantageously implemented according to the embodiments and variants described below, which are to be considered individually or in any technically operative combination. According to an advantageous embodiment, the method of the invention comprises between steps e) and f) the steps of:

boy Wut. obtain the position of the intrusive object in the environment of the robot; h. calculate an avoidance trajectory;

i. continue moving the robot along the avoidance path calculated in h).

Thus, according to this improved embodiment, the robot continues to perform its tasks at the speed of safety as long as it is possible to avoid the obstacle and thus avoid being in the situation of step b) which leads to a stop. This embodiment is also safer for the operator by considerably reducing the risk of collision even if the attention of the operator is relaxed.

According to an improved embodiment of the previous embodiment, the method of the invention comprises between steps i) and f) the steps of:

j. if the proximity of the intrusive object is detected by the proximity sensor, calculating a distance path of said object.

Thus, if there is a risk of collision detected, detected by the proximity sensor, the robot does not just slow down but tries to escape the obstacle which increases the safety of the operator or the other robot located in the monitored environment.

Advantageously, according to this last embodiment, the method which is the subject of the invention between steps j) and f) a step consisting in:

k. if a contact with the intrusive object is detected by the contact sensor, generating a path away from the contact.

This embodiment makes it possible to prevent contact, even under reduced effort, being maintained with the object.

According to an alternative embodiment, the method which is the subject of the invention comprises, between steps g) and f) a step consisting of:

I. if the proximity of the intrusive object is detected by the proximity sensor or if contact with said intrusive object is detected by the contact sensor, place the robot in a state of gravity compensation.

This embodiment allows the operator to move the robot himself by pushing it effortlessly.

Advantageously, the orders of movement of the robot are generated by a controller delivering orders of time position in servo and; the reduction of the speed during step e) is obtained by modifying the interpolation time interval of the robot without modifying the servo frequency. According to a particular embodiment, the time position commands a (t) of the robot are delivered by the controller as a function of a theoretical trajectory s (t) according to a control time interval At corresponding to a servo frequency 1 / At, so that under nominal operating conditions a (t) = s (t); and the reduction of the speed during step e) is obtained by introducing a virtual time so that c (t + At) = s (t + k.At) with k≤l. According to an embodiment of steps j) or k) of the method which is the subject of the invention, the robot moving towards a target position by following a theoretical trajectory, the modified trajectory during these steps is obtained by bending said theoretical trajectory proportionally. a repulsion vector oriented along the detection axis of the sensor and of intensity proportional to the information delivered by said sensor. This embodiment makes it possible to alter the trajectory in real time so as to escape the collision or to limit the effects thereof, while continuing, as far as possible, the displacement towards the target point in the case of the implementation. the work of step j).

According to an embodiment of step h) of the method that is the subject of the invention, the robot moving towards a target position along a theoretical trajectory, the calculation of the avoidance trajectory of step h) comprises the generation several random theoretical positions in the robot's surveillance space, the elimination of the random positions colliding with the intrusive object, and the definition of the shortest path to reach the target position among the remaining positions. This embodiment allows the rapid generation of an avoidance trajectory. The invention is explained below according to its preferred embodiments, in no way limiting, and with reference to FIGS. 1 to 5, in which:

FIG 1 shows schematically the robot and its surveillance space in a view from above;

FIG. 2 illustrates the control principle in virtual time;

FIG. 3 represents the flowchart of an exemplary embodiment of the method which is the subject of the invention;

FIG. 4 illustrates the principle of calculating an avoidance trajectory;

and FIG. 5, schematizes the principle of calculating an escape trajectory. FIG. 1, according to an exemplary embodiment, the robot (100) which is the subject of the invention comprises means (150) for monitoring its environment, for example in the form of a vision and location sensor (150) such that a 3D camera or a laser. This sensor, attached or independent of the robot (100) monitors an area (1 10), called concentric surveillance space, and locates in this space the robot (100) and any new object (190) or operator crossing the boundary of this zone . The concentric zone of surveillance is here represented schematically and arbitrarily. In practice, it is a three-dimensional volume of shape adapted to the operation performed by the robot, and integrating said robot regardless of its articular position. The extent of the surveillance zone, in which the intrusion of an object (190) is considered as a risk of collision, is a function of the speed of movement of the robot. The higher this speed, the more extensive the surveillance zone. The robot (100) also comprises one or more proximity sensors (not shown), able to detect the presence of an object or an operator in a zone (120) constricted around the robot. The means (150) of vision and positioning perform a permanent monitoring of the environment (1 10) of the robot, while the proximity sensors deliver information only if the proximity of an object is detected. By way of non-limiting example, the proximity sensor is a light barrier or an ultrasonic sensor. The zone (120) for detecting the proximity sensors of the robot (100) is in practice a volume of any shape, depending on the technology or the combination of detection technologies. The robot (1 00) also comprises one or more contact sensors (not shown) which deliver information when an object or an operator comes into contact with the robot. By way of non-limiting example, such a contact sensor is made by measuring the control currents of the axis motors or by a force sensor. Thus the system implemented by the method that is the subject of the invention comprises several levels of detection of an intrusion, and the steps intended to protect the robot and the object of the intrusion are implemented gradually as a function of the crossing of domains monitored by these different means of analysis. Each detection level is monitored by one or a plurality of sensors. According to the method of the invention, when an object (1 90) or an operator crosses the limit (1 1 0) of the monitoring space, the speed of movement of the robot is reduced to a safety speed. Reducing the robot's traveling speed when an object crosses the boundary of the monitoring space is achieved by changing the interpolation time interval of the robot without changing the servo frequency.

Figure 2, the theoretical trajectory (200) s (t) of the robot is defined by a plurality of points (202, 203, 204, 205, 206, 207). To calculate the real trajectory, c (t), of the robot according to these points of passage, an interpolation is carried out between these points, for example by means of a spline function. From this interpolation, the computer defines the intermediate positions according to a time interval At corresponding to the control servo frequency of the robot. As an example At = 0.02 seconds and the servocontrol frequency is 50 Hz. Thus, at time t, the robot is in the position p (t) and at the time (t + At), the robot is in the position p (t + At) distant from d of the last position. The instantaneous speed of the robot between the two interpolation points is d / At. The position of the point p (t + Δt) as a function of the position p (t) is given by the desired speed of the robot as a function of the servo frequency. It is calculated from the interpolation function of the trajectory so that the actual trajectory in position, speed and acceleration of the robot, corresponds to the programmed theoretical trajectory. Thus, the robot displacement controller addresses to the axes of said robot movement commands corresponding to each position interpolation according to a fixed servo frequency. When an intrusion into the surveillance space is detected, the interpolation of the robot's displacement is performed according to a k.At interpolation interval, but the servo frequency remains the same, equal to 1 / At. Thus, starting from the position p (t + At), the next interpolation point should be at a position (21 1) as interpolated for a time At. The factor k being less than 1, the position ( 21 0) calculated interpolation is late compared to this position (21 1) theoretical. Also the robot slows down without the programming being modified. To achieve this function, the controller of the robot comprises a second clock, controllable, for the definition of the interpolation time used for the calculation of the trajectory. This mode of speed control is commonly referred to as a virtual time control and can be compared, from a didactic point of view, to the slow motion effect obtained by filming a scene at a higher image acquisition rate. than the projection frequency of the film.

Figure 3, according to an ultimate level of security, the robot control system continuously monitors the state of the contact sensors. During a test step (31 0), if a contact with the robot is detected, for example, by detecting a control current exceeding a determined threshold on one of the robot axis motors, stopping emergency (31 1) of the robot is triggered. This emergency stop stops the robot which must be reset to restart. Outside of this emergency, the robot works at its working speed, with the highest possible productivity. The environment of the robot is scanned continuously by the means of vision and location. If, during a detection step (320), an intrusion is detected as crossing the limited monitoring space, then, during a control step (330), the speed of the robot is reduced to a so-called predetermined safety speed, for example in accordance with the ISO standards 1 021 8 and ISO TS 1 5066 for the co-activity. This security speed is maintained as long as the introduced object is in the surveillance space. The reduction of speed is achieved by means of the control in virtual time, so that the program of displacement is continued but at the reduced speed allowing for example to the operator himself found in co-activity in the environment of the robot, better anticipate the movements of the robot and reduce the intensity of a possible shock. The monitoring of the continuous contact sensors and the ultimate safety mode resulting in the emergency stop of the robot remains active. According to one embodiment, the vision and positioning sensors that scan the surveillance space are able to determine the position of the intrusive object in the environment of the robot. This position is transmitted to the robot control system which calculates (340), from this information, an avoidance trajectory of the intrusive object.

4, the robot being at a point (410) of the space and moving towards a target point (420), in order to calculate the avoidance trajectory, the calculator generates a series of random points (430) in the surveillance space. Sets of points (441, 442) that collide with objects in the robot's environment, including the intrusive object, are eliminated. The calculator then determines a path (450), the shortest, passing through the remaining points and connecting the starting point (410) and the target point (420). Thus, the avoidance trajectory is calculated quickly.

Returning to FIG. 3, if during a detection step (350), a signal coming from a proximity sensor is detected, then the trajectory of the robot is modified, during a calculation step of escape (360), so as to keep the robot away from this proximity.

Figure 5, the robot is in an initial position (510) and is moving towards a target position (520). In the absence of detection by the proximity sensors, this movement is made in a direction (530) oriented from the initial position to the target position and the trajectory of the robot follows this direction. In the presence of detection by a proximity sensor (540), said sensor detects the presence of the object along a defined axis (541). This detection thus defines a vector (542), said repulsion, oriented in the direction (541) of detection of the sensor, and intensity all the more important that the detected object is close. This vector (542) is combined with the vector (530), called attraction, defining the initial trajectory of the robot, robot whose path (550) is inflected accordingly, moving it away from the intrusive object while continuing its path to the target (520).

Returning to FIG. 3, the robot operating at the reduced safety speed, if during a contact detection step (370) a contact with an intrusive object is detected, the trajectory of the robot is modified during a step (380) of removal, so as to move the robot away from this contact. The distance trajectory is calculated similarly to the escape trajectory but considering only the repulsion vector: the robot moves away from the contact by following this repulsion vector. According to one embodiment, the robot is then stopped. According to another embodiment, the robot is placed in a gravity compensation situation, which makes it possible to easily move the robot.

The above description and the exemplary embodiments show that the invention achieves the desired aim and enables the robot likely to be in co-activity with an operator to work to the maximum of its productivity, while improving the security of said operator. . The method of the invention is effective vis-à-vis the work movements of the robot but also in the context of its movement between two work stations.

Claims

Method for controlling the operation of a robot within a system comprising said robot (100) and means for analyzing its environment, in particular concentric, comprising:

i. a contact sensor;

ii. a proximity sensor (540);

iii. a sensor (150) for vision and location;

characterized in that it comprises the steps of:

at. obtain for each axis of the robot a value of maximum allowable effort;

b. if the force on one of the axes of the robot is greater than the maximum value acquired in step a) (310) stop (31 1) the robot in its position;

vs. to obtain a space (1 10) of surveillance, said security space, the extent of which is a function of the speed of the robot; d. monitor the environment of the robot by means of analysis; e. if the intrusion of an object is detected (320) in the security space of the robot, progressively reducing (330) the speed of evolution to a so-called safety speed;

f. resume in step a).

The method of claim 1, comprising between steps e) and f) the steps of:

boy Wut. obtain the position of the intrusive object (190) in the robot environment;

h. calculating (340) an avoidance trajectory (350);

i. continue moving the robot along the avoidance path calculated in h).

The method of claim 2, comprising between steps i) and f) the steps of: j. if the proximity of the intrusive object is detected (350) by the proximity sensor (540), calculating (360) a path (550) away from said object.

The method of claim 3 comprising between steps j) and f) a step of:

The method of claim 2 comprising between steps i) and f) a step of:

I. If the proximity of the intrusive object is detected by the proximity sensor or if contact with the intrusive object is detected by the contact sensor, place the robot in a state of gravity compensation.

A method as claimed in any one of claims 1 to 5, wherein the robot movement commands are generated by a controller delivering time position commands in servo and wherein the speed reduction during step e) is obtained by modifying the interpolation time interval of the robot without modifying the servo frequency.

Method according to claim 6, wherein the time position commands a (t) of the robot are delivered by the controller as a function of a theoretical trajectory s (t) according to a control time interval At corresponding to a servo frequency 1 / At, so that under nominal operating conditions a (t) = s (t); and wherein the reduction of the velocity during step e) is obtained by introducing a virtual time so that o (t + At) = s (t + k.At) with k≤l.

A process according to claim 3 or claim 4, wherein the robot moving towards a target position (520) following a theoretical trajectory, the trajectory modified during step j) or step k) is obtained by bending said theoretical trajectory in proportion to a vector (542) of repulsion, oriented along the detection axis of the sensor and intensity proportional to the information delivered by the sensor.

The method of claim 2, wherein, the robot moving to a target position (420) along a theoretical path, calculating the avoidance path (450) in step h) comprises generating multiple positions ( 430) in the concentric monitoring space of the robot, the elimination of the random positions colliding with the intrusive object, and the definition of the shortest path (450) to reach the target position (420) among the positions remaining.