WO2009155946A1

WO2009155946A1 - Adaptive robot system

Info

Publication number: WO2009155946A1
Application number: PCT/EP2008/005204
Authority: WO
Inventors: Björn MATTHIAS; Sönke KOCK; Roland Krieger
Original assignee: Abb Ag
Priority date: 2008-06-26
Filing date: 2008-06-26
Publication date: 2009-12-30

Abstract

The invention is related to a safe robot system, comprising at least one robot (16, 62), comprising at least one sensor means (22, 26, 30) for the measurement of data of at least one person (12, 66, 82), which is within or near the working range of the at least one robot (16, 62) and comprising a processing system (20) for the measured data. At least one person-related parameter is determined by the processing system (20) and a predetermined trajectory of the at least one robot (16, 62) is adaptively influenced by the at least one person-related parameter.

Description

Adaptive robot system

Description

The invention relates to a safe robot system, comprising at least one robot, comprising at least one sensor means for the measurement of data of at least one person, which is within or near the working range of the at least one robot and comprising a processing system for the measured data.

Robots or robot systems with the ability to perform cooperative tasks with human workers are increasingly known for example from the manufacturing industry but also in the area of private households. Especially within the manufacturing industry the benefit of such a robot system is as well an increase of the degree of automation, but also a limitation of the effort for automation, since those sub-tasks within the manufacturing process, which require a high effort to automate, are still done by a human worker.

An example for such a robot system, which has the ability of human-robot collaboration, is described in US 6522952, where a method for the transportation of an object in a collaborative task of a robot and a worker is disclosed.

Precondition for each robot or machine system with the ability to perform collaborative tasks with human workers is, that this system is safe and any avoidable risk for the worker or the person is avoided. Systems related to the safety of a person in or near the working range of a machine or a robot are described for example in EP1367314, where an optical safety system with at least 2 cameras is disclosed. EP1635107 discloses a method for controlling at least one safety related function of a machine.

The safety aspect as described for example within the 2 documents mentioned above is not the scope of this invention; it is rather a precondition for this invention, that the disclosed robot system for robot-human interaction is already safe. Therefore, the required safety aspects are not mentioned in the further description, unless they increase the required safety standard.

Disadvantageous within the state of the art is that the machine, the system respectively the robot system, which has a collaborative task with a person, does not consider the specific ergonomic requirements of the person, who interacts with the robot. The machine or robot has a given trajectory, described by movement path as well as by movement speed, independent of the characteristics of the interacting person. Only due to safety reasons, the movement of the machine or robot is immediately stopped, if required.

Therefore, the objective of the invention is, to provide a system, which improves the robot-human collaboration.

This problem is solved by safe system for human-robot collaboration with the features described in claim 1.

Accordingly at least one person-related parameter is determined by the processing system and a predetermined trajectory of the at least one robot is adaptively influenced by the at least one person-related parameter.

A person related parameter is any parameter describing in a quantifiable manner either a characteristic or a condition of the person, who is in a collaborative task with the robot, whereas this parameter is also of relevance for the collaborative task. A characteristic can be related to some physical aspects, for example the height of the person. In case that the person is very small, all interaction points between robot and person have to be adapted on the robot side to the small height of the person. A possible person-related parameter is furthermore a skill characteristic, for example the working speeds. If an analysis of the movement speeds of the person over some working cycles within the collaborative task shows, that the worker is a slow worker, maybe due to a handicap, the overall task speed of the robot will be adapted to the needs of the person.

Also a temporary condition of the person is a suitable parameter for an adaptive influence of the trajectory of the robot. If it is detected that the person bends down very often or stands in a ducked posture, a resulting parameter describes, that the physical condition of this person is impaired. As a result of an automated deeper analysis by the processing system, some additional breaks in-between working cycles, or a reduced movement speed or an overall stop of the collaborative task might be the consequence.

A temporary condition is also for example a direct or indirect expression of the person, that he is in trouble, for example a loud scream for help, which will lead to an immediate stop of the robot. By this way, also the degree of personal safety is increased, whereas the whole robot system must be safe also without such a function.

Hence it is possible in an advantageous way to improve the collaboration between robot and human in an ergonomic way.

In a variant of the invention the trajectory of the robot is predetermined by a robot program stored in a controller, which is related to the robot. Normally a robot is controlled by a related controller, comprising also the capability of a data processing device.

A robot program is a sequence of different commands to the robot, containing as well the movement path as the movement speed. Normally the parameters of such a program, such as coordinates or speed, are also described as variables within the program. A variable is easily influenced by a multiplication factor, for example when changing movement speed, or by an offset, for example when shifting the movement path. In a further embodiment of the invention, the system comprises at least one sensor means for the measurement of data of an object, which is handled by the at least one person and the robot in a collaborative task and those data are also provided to the processing system.

Using for example a vision based system with cameras as sensor means, an object and its position is in some cases easier to identify than a person, which carries this object, which might have more significant characteristics. Therefore, for example the movement speed of the person is easier detectable indirectly over the movement speed and movement path of an object, which is carried by the person.

In a preferred embodiment of the invention, the at least one sensor means is an optical sensor, an acoustical sensor, a heat sensor, a force sensor, an ultrasonic sensor, a radar sensor and/or an infrared sensor.

Optical sensors are for example 2 or more cameras, which are connected to a vision system, analyzing the 3D environment of the area to be supervised, e.g. the working range around the robot. The vision system might be part of the sensor system or integrated into the processing system or it is a separate component in-between sensors and processing system.

Sensors are integrated in the environment of the robot preferably on a fixed position, but also on the robot itself as a moving part. This is for example an acoustical sensor on the tip of the robot arm. Such sensors are in an advantageous manner suitable for measuring data of a person respectively of an object.

In a further variant of the system the measured data are applied to the processing system in real-time and the at least one person-related parameter is determined repeatedly times in intervals.

By this way it becomes possible, to analyze changing conditions or characteristics of the person and to provide actual person-related parameters in time intervals, for example every 0.01 ms - 50ms. The accuracy and the effect of the adaptive influence will become increased by this. In another embodiment of the invented system, the processing system is connected with a data reading device for reading person-related data from a storage medium.

This is for example an automatic reader of an ID card, which is required to be used by a person before starting the collaboration task with the robot. The use of this reader could open for example a door to an enclosed working area with the robot inside. In a very advantageous manner the effort for sensors to measure data of persons can be reduced, since some person-related data are stored on the storage medium and don't have to be measured.

It is as well possible, that only an ID-code is stored on the storage medium, whereas each ID number is related to a set of typical person-related data, like height or a personal working skill. It is also possible, to store such person-related parameters directly on the storage medium.

In a preferred variant of the invented system, the at least one person-related parameter is quantifying a characteristic or an actual condition of the respective at least one person within the working range of the robot. Such a parameter indicates for example a length, a height, a position, a distance, a movement speed, an interpreted speech or gesture or an interpreted posture of the respective at least one person within the working range of the robot.

Some of those parameters, which are not of geometrical kind, require some certain effort for interpretation, for example the interpretation of acoustical signals. In case, that the person wants to give some direct acoustical commands to the robot, for example 'higher' or 'lower', the interaction point of the robot is shifted. In case of the command 'stop' the robot performs an emergency halt.

Also the interpretation of a gesture of the person, such as a beckon of the person collaborating with the robot, has to be interpreted by a special data processing system, in this case preferably using optical signals from some cameras. As a result, the interpretation leads to a person related parameter, which is in this case for example 'attention' and causes a slowing down of the robot movement. In a variant of the invented system the adaptive influence on the predetermined trajectory of the robot causes at least in sections a change or shifting of the movement path of the trajectory.

This is for example useful, if the person collaborating with the robot is detected to be small, so that all sections of the movement path containing some interaction points with the person will be shifted down. Then the ergonomic conditions for the person are advantageously improved.

In another variant of the invented system the adaptive influence on the predetermined trajectory of the robot causes an operational halt, an operational resumption or a change of the movement speed.

This is for example useful, if a situation requiring special attention has been detected by the processing system respectively such a situation is over and the robot resumes normal operation. The interpreted speech of the collaborating person is an example for a person related parameter, which causes such an effect, like the verbal com- maods 'slow', 'stop' or 'resume'.

In a further variant of the invented system the adaptive influence on the predetermined trajectory of the robot causes a repetition and/or skipping of at least a section of the trajectory, for example if the handing over of an object from the person to the robot has failed and the handing over movement of the robot is repeated once again.

In a next variant of the invented system the adaptive influence on the predetermined trajectory of the robot causes an insertion of at least one further section of another trajectory. For example the handing over of an object from the robot to the person failed in that way, that the object falls down. If the person is detected to be still quiet and not in panic, an inserted pick-up routine for this object will be executed and then the regular trajectory will be continued.

In another variant of the invented system the adaptive influence on the predetermined trajectory of the robot causes a change of the sequence of at least two sec- tions of the trajectory. This is useful for example, if each section of the trajectory is related to a working step and the sequence of doing those steps is absolutely not of importance. If the processing system recognizes, that the person makes a lot of movements to fulfill some working steps, an automated analysis will create another sequence of those working steps, which might cause a reduced overall movement of the worker doing those working steps. Such an analysis is preferably done within the processing system, but also a separate data processing unit might be used for it.

In a preferred embodiment of the invented system the processing system is integrated into the robot controller at least in part. The number of components used is reduced by this.

In another embodiment of the invented system the actual status of the robot movement within the previous trajectory impacts the adaptive influence on the further trajectory.

As a consequence a same quantified parameter occurring at the beginning of a collaborative task causes a different influence on the robot trajectory than if it occurs at the.jend of a collaborative task. If a worker gives for example the acoustical command 'slow' at the beginning of a task, whereas approx. 120s of the task have to be done, the robot will slow down. At the end of the task, whereas approx. only 3s of the task are remaining, the same command will have no effect at all.

Further advantageous embodiments of the invention are mentioned in the dependent claims.

The invention will now be further explained by means of an exemplary embodiment and with reference to the accompanying drawings, in which:

Figure 1 shows an example for an adaptive robot system and person in work cell, Figure 2 shows a small person near robot and Figure 3 shows a tall person near robot. Fig. 1 shows an example 10 for an adaptive robot system and a collaborating first person 12 within a work cell 14. The work cell 14 is preferably an enclosed area, which is accessible for example over a lockable door. A 6-axis first robot 16 is positioned on the floor of the working cell 14, whereas the first robot 16 is connected to a robot controller 18 over a bi-directional sixth data connection 44. The robot controller 18 has also the ability of a data processing device, wherein a robot program containing a sequence of different commands is stored. Those commands of the robot program, which might be influenced by person-related parameters, contain variables, which are easily influenced by a factor or an offset. Variables are preferably used for coordinates or a reference speed. But it is also possible, that the robot controller provides the option, to have a global factor on the speed for example.

The robot controller 18 is furthermore connected with a processing system 20, which is preferably also a data processing unit like a computer. Data from the processing system 20 to the robot controller 18 are transferred over the fifth data connection 42, whereas those data contain either person-related parameters of the first person 12, such as a distance, some already interpreted speech commands like 'attention', 'stop' or 'start', or already offsets or factors like 'half speed' which have direct influence on the trajectory of the robot.

Data from the robot controller 18 to the processing system 20 are transferred over the fourth data connection 40, preferably data concerning the robot program respectively the actual status of the robot program. Those data are useful for the determining of person related parameters or also for the interpretation of some gestures.

Also for the analysis of a 3D visual data of a person within the working range of the robot it is important information to have the actual position of the robot from the robot controller, preferably in real time.

The relation, which person-related parameter causes what changes in the trajectory, is for instance either determinable within the control system 20 or within the robot controller 18 or also on a separate data processing unit, which is not shown in this figure. Such relation is for example either given by some fixed rules or is given in a predetermined decision matrix or is also realizable with methods of artificial intelligence, such as expert systems or artificial neural networks.

The processing system 20 is connected to a first 22 and second 26 optical sensor, preferably a camera, with the first 34 and second 36 data connection, whereas the data transfer is preferably in real time, which means a time delay <5ms, preferably <0.1s. Both sensors have an observation angle 24 respectively 28, which is adjusted to the working range of the robot 16. Due to the use of two optical sensors, a 3D analysis of a person within the working range is feasible. The use of more than two cameras will increase the accuracy.

The processing system 20 is also via the third data connection 38 connected to an ultrasonic sensor 30, which observes the measuring area 32 concerning objects or persons 12 within this area.

Fig. 2 shows a small second person 66, which is near and within the working range of a second robot 62. The reference height 64 is an ergonomically suitable height for an interaction point of the second person 66 with the second robot 62, for example for a handing over an object. For those segments of the trajectory of the second robot 62, which have interaction points, a suitable vertical offset has to be chosen, so that all interaction points are at the reference height 64.

Fig. 3 shows a tall third person 82 near the same robot 62. Since the third person 82 is higher than the second person 66, the optimal ergonomic height for an interaction of the third person 82 with the robot 62 is the increased height 84, which differs by the offset of 86 from the reference height 64. List of reference signs

adaptive robot system and person in work cell first person work cell first robot robot controller processing system first optical sensor observation angle-of first optical sensor second optical sensor observation angle of second optical sensor ultrasonic sensor measuring area of ultrasonic sensor first data connection second data connection third data connection forth data connection fifth data connection sixth data connection small person near robot second robot reference height second person long person near robot third person increased height height difference

Claims

1. Safe robot system, comprising at least one robot (16, 62), comprising at least one sensor means (22, 26, 30) for the measurement of data of at least one person (12, 66, 82), which is within or near the working range of the robot (16, 62), comprising a processing system (20) for the measured data, characterized in that at least one person-related parameter is determined by the processing system (20) and that a predetermined trajectory of the at least one robot (16, 62) is adaptively influenced by the at least one person-related parameter.

2. System according to claim 1 , characterized in that the trajectory of the at least one robot (16, 62) is predetermined by a robot program stored in a controller (18), which is related to the robot (16, 62).

3. System according to claim 1 or 2, characterized in that the system is comprising at least one sensor means (22, 26, 30) for the measurement of data of an object, which is handled by the at least one person (12, 66, 82) and the at least one robot (16, 62) in a collaborative task and that those data are also provided to the processing system (20).

4. System according to any of the previous claims, characterized in that the at least one sensor means (22, 26, 30) is an optical sensor, an acoustical sensor, a heat sensor, a force sensor, an ultrasonic sensor, a radar sensor and/or an infrared sensor.

5. System according to any of the previous claims, characterized in that the measured data are fed to the processing system (20) in real-time and that the at least one person-related parameter is determined repeatedly in intervals.

6. System according to any of the previous claims, characterized in that the processing system (20) is connected with a data reading device for reading person-related data from a storage medium.

7. System according to any of the previous claims, characterized in that the at least one respective person-related parameter is assigned to respective person-related data.

8. System according to any of the previous claims, characterized in that the at least one person-related parameter is quantifying a characteristic or an actual condition of the respective at least one person (12, 66, 82) within the working range of the at least one robot (16, 62).

9. System according to any of the previous claims, characterized in that the at least one person-related parameter indicates a length, a height, a position, a distance, a movement speed, an interpreted speech or gesture or an interpreted posture of the respective at least one person (12, 66, 82) within the working range of the at least one robot (16, 62).

10. System according to any of the previous claims, characterized in that the adaptive influence on the predetermined trajectory of the at least one robot (16, 62) causes at least in sections a change or shifting (86) of the movement path of the trajectory.

11. System according to any of the previous claims, characterized in that the adaptive influence on the predetermined trajectory of the robot causes an operational halt, an operational resumption or a change of the movement speed.

12. System according to any of the previous claims, characterized in that the adaptive influence on the predetermined trajectory of the at least one robot (16, 62) causes a repetition and/or skipping of at least one section of the trajectory.

13. System according to any of the previous claims, characterized in that the adaptive influence on the predetermined trajectory of the robot (16, 62) causes an insertion of at least one further section of another trajectory.

14. System according to any of the previous claims, characterized in that the adaptive influence on the predetermined trajectory of the at least one robot (16, 62) causes a change of the sequence of at least two sections of the trajectory.

15. System according to claim 2 to 14, characterized in that the processing system (20) is integrated into the robot controller (18) at least in part.

16. System according to any of the previous claims, characterized in that the actual status of the robot movement within the previous trajectory impacts the adaptive influence on the further trajectory.