Description

The invention relates to a method for the control of a robot while the robot is executing a motion program.

It is known that industrial robots are used for example in permanently repeated production processes such as the assembly of parts in a series production. Such an industrial robot is normally characterized by a robot arm and five to seven degrees of freedom in movement in total so that normally each position within the working range of the robot can be reached by the end of the robot arm in every orientation in the case of at least six degrees of freedom in movement. The working range of a typical industrial robot amounts to 1 - 4 meters around the associated rotary robot base. Of course, also other manipulators can be considered as robots.

The robot movement is determined by a motion program, comprising typically the data of a given movement path, the associated individual movement speed along the movement path, and also the movement direction. The robot movement is highly accurate and reproducible with a tolerance of for example less than 1 mm along the whole movement path. The motion program is normally stored on a data processing device such as a computer. Additionally, an interface between data processing device and robot is required, comprising for example amplifiers to generate electrical control signals for the motor drives of the axes of the robot. In some cases those functionalities are combined in a single robot controller.

For some tasks in industrial production robots are not suitable, especially in the case of unpredictably strict tolerance requirements on parts to be assembled. In such cases a human worker is still unavoidable, since his ability to recognize individual tolerances of workpieces and to adapt the motion of parts to be assembled and also of the assembly tools is still unmatched.

Prior art, for example DE-102007008238 A1 , describes systems for human-robot collaboration to perform a handling or an assembly task. The advantage of such systems is that the human worker is assisted by a robot doing for example the movement of a heavy part, whereas the human worker or operator fully concentrates on the more complex aspects. The human worker manually guides the robot to the desired position in the desired orientation, without himself lifting the heavy part to be assembled.

Disadvantageous within the state of the art is that safety regulations require limits on speed and position of the robot manipulator in the presence of a human worker. Speed limitations also imply compromises in cycle times for the achievement of the overall manufacturing task.

Therefore this invention aims at the increase of the speed of human-robot cooperation tasks while simultaneously respecting the requirements of safety regulations.

This problem is solved by afore mentioned method for robot control while the robot is executing a motion program comprising at least three motion program sequences, which is characterized at least by the following steps:

• a first fixed motion program sequence, comprising at least data of motion path, motion speed and motion direction, is executed by the robot in autonomous motion state, • subsequently and according to the motion control program, the robot arm is moved into a transfer zone, which is located within the working range of the robot, and the robot is switched into a person-safe waiting motion state,

• an operator is entering the transfer zone,

• the robot is executing in a safe human-robot cooperation an adaptively influ- enceable motion program sequence, which is controlled at least in part by the operator,

• the robot arm is temporarily moved back to a transfer point within the transfer zone and the operator is leaving the transfer zone to a safe position,

• temporarily, the robot arm is moved along an autonomously calculated path from the transfer point to a start point,

• a second fixed motion program sequence is executed by the robot from the start point in autonomous motion state

For a given assembly or other production task, the potential for such an improvement can be derived from identifying which steps in the procedure are most efficiently handled by a robot in autonomous motion state and which are best handled by a human operator working in cooperation with the robot. To release this potential, one must divide the production procedure into different sequences.

In an autonomous motion state the robot acts independently of direct input from the operator and can operate at high speeds, assuming that no humans are present in the relevant robot working space. The autonomous motion state is referring to autonomous motion by the robot and not referring to the motion of the robot under direct program control from the controller.

During execution of an adaptively influenceable program sequence the robot motion must be limited to a level that is safe for the presence of and manual guiding by a human operator. This can mean, for example, that speed limit on the robot motion must be enforced, for example an upper limit of 250 mm/s. It can also mean that an enabling switch must be pressed by one or by each of the operator's hands for the duration of the entire adaptively influenceable program sequence. Also, it can mean that the robot motion is restricted to a supervised stand-still at zero speed, when the operator is not activating any intended input devices. In special cases, it can mean that the robot speed is limited to very low speeds such as 30 mm/s or less.

By splitting the robot movement program into adaptively influenceable sequences requiring an operator on one side and fixed sequences not requiring an operator on the other side, the cycle time of the overall movement program might be reduced significantly. Normally, only the final assembly movement of the robot, for example fitting together parts, requires a human robot cooperation, whereas the major part of the movement program, such as gripping a part from a supplying device and moving it close to the other part, do not require human-robot cooperation when executing.

Since a robot movement within a human-robot cooperation task is normally limited to a speed such as 250 mm/s due to safety reasons, the cycle time for performing such a cooperation task is significantly longer than the cycle time of a comparable robot task in autonomous motion state, where the robot moves for example with a maximum speed of more than 1000 mm/s. Also other safety procedures such as analyzing the common working range of robot and operator and different zones thereof or stopping the program execution until a certain zone becomes released by the leaving operator are increasing the required cycle time.

Therefore, the robot control according to the invention is based on the idea, to surround the essentially required program sequence for human-robot cooperation by program sequences in autonomous motion state, wherein the robot works at higher movement speed and no safety aspects of the operator in a common working range have to be respected.

The safety of the worker is assured in that a clear changeover between autonomous motion state and adaptively influenced motion state of the robot is realized. Therefore, a common transfer zone is foreseen providing full safety for the operator. Robot and operator might be within this zone at the same time to change over the move- ment task. The operator is not allowed to enter this zone unless the robot is in a per ^¬ son-safe state, otherwise an emergency stop will be executed.

On the other side, the robot will not start to execute a movement program sequence in autonomous motion state, unless the operator has left the transfer zone. Of course it is possible that the robot starts executing a motion program sequence, which is basically foreseen to be executed in autonomous motion state, with significantly slower movement speed in a person-safe motion state while the operator is still in the transfer zone, but in this case the operator must already have signalized his intent to change over the control to the robot.

In principle, the working area can be divided into three parts: a transfer zone for the changeover, a working zone for the human-robot cooperation and a no-go zone for the human operator, which is only intended for robot movements in autonomous motion state. Of course, dependent on the given environmental conditions and application requirements, such zones may overlap, be identical or non-existent. The transfer-zone may be for example identical with the whole working zone for human-robot cooperation. It is also conceivable that a no-go zone is not provided at all. It is also possible that numerous adjacent or partly overlapping working- and/or transfer zones are provided so that the safety level for the worker increases with decreasing distance between the human worker and the robot.

Also, varying zones dependent on the executed movement program or dependent upon the type of work piece to be handled are within the scope of the invention.

Hence, it is possible to reduce the overall cycle time of a human-robot cooperation task in an advantageous manner.

In a variant of the invention the adaptively influenceable motion program sequence is not executed before generating a release signal, for example by manually pushing a button of a control device. The safety of the system will be further increased by this additional feedback of the operator. According to another embodiment of the invention the second fixed motion program sequence is not executed before generating a release signal, for example by manually pushing a button of a control device. Also this additional feedback will further increase the safety of the system.

In a preferred embodiment of the invention the release signal is autonomousally generated based on measurement data of sensors, which are supervising at least the transfer zone or the limits thereof, such as a light barrier, a camera, an acoustical sensor, an infrared sensor and/or ultrasonic sensor.

Hence it is possible, for example, to visually supervise the presence of a human worker in the transfer zone after finishing the cooperation task. When the transfer zone has been detected to be empty after the worker has left, a release signal will be given autonomously so that the next program sequence in autonomous motion state can be started. Of course, a supervision of the transfer zone or other adjacent zones may also be realized based on the measurements of other sensors, such as microphones for detecting for example some spoken commands of the operator, infrared sensors measuring for example the body temperature of the worker or ultrasonic sensors measuring for example a distance between worker and robot. Preferably, the analysis of the measurement data of the sensors is done on a data processing device such as a computer.

Using such an autonomously generated release signal, the procedure according to the invention is simplified on one side, whereas on the other side the safety of the operator is still at the required high level. The safety of the human operator is furthermore increased when an acoustical and/or optical signal is given after the robot has been switched into the person-safe waiting motion state.

In a further variant of the invention, the movement speed of the robot is limited to a certain maximum speed while executing the adaptively influenceable motion program sequence. Preferably the speed limit does not exceed 250 mm/s, for example 200 mm/s.

According to another variant of the invention, the robot is switched into an intrinsically safe operation motion state before starting the execution of the adaptively influence- able motion program sequence. Such an intrinsically safe operation motion state provides once again an increased level of safety for the operator compared to the person-safe wait motion state, since it is physically excluded that the robot can hit the operator in any dangerous way. Optionally, the robot can be switched back to the normal person-safe motion state after the worker takes over the control of the adaptively influenceable robot program sequence.

In a preferred embodiment of the invention the robot movement is adaptively influenced at least concerning the movement path, the movement speed or the movement direction while executing the motion program sequence. It is possible as well that the path of the robot movement is fixed and the worker can move the robot only along this so-called virtual rail. It is also conceivable that the robot movement is influenced in all three spatial directions. In this case, the robot waits for example for a force impact on a force sensor and the robot arm becomes autonomously moved in the direction of the force impact.

In a preferred embodiment of the invention the processed measurement data of at least one sensor provided in the proximity to the robot or mounted on the robot, such as a force sensor as described before, an acoustical sensor, a current sensor and/or an optical sensor are the basis for the adaptive influence on the motion program sequence. Such types of sensors are well-suited for detecting especially the intention of the operator either based on his movements, his voice and/or his posture for example. The temporarily required interpretations for example of the voice and/or the gestures of the operator are preferably done autonomously using a data processing device such a computer. In some cases a software using artificial intelligence such as artificial neural networks is advantageous. In a certain variant of the invention the motion program sequence is adaptively influenced by the control signals of a separate control device such as a switch, a joystick, a control panel and/or a teaching device. In this case the effort for the interpretation of the intention of the operator using the control devices is advantageously low.

Preferably the motion program comprises the data of at least one gripping movement of a gripper tool attached to the robot. This simplifies the handling of a part.

According to a preferred embodiment of the invention, the motion program is stored in a robot controller which is directly related to the robot. A robot controller comprises normally a data processing device as well as electrical amplifiers for the motor drives of the axes of the related robot. Storing the movement program in the already provided data processing device of the robot controller will reduce the hardware effort to practice the method according to the invention.

An additional comparable effect is gained by the integration - at least in part - of the processing and/or analyzing task of the measurement data into the robot controller. In this case the measurement data of the sensors which supervise the transfer zone will be transmitted at least in part to the robot controller respectively its data processing device and will be analyzed there for example by a suitable software program.

According to a preferred embodiment of the invention, the motion program comprises more than two fixed and more than one adaptively influenceable motion program sequence, whereas in a further embodiment a selection and/or a sequence of several motion program sequences of the motion program may be altered temporarily.

Hence it becomes possible to use the method according to the invention also in motion programs of higher complexity. The case-based sequential execution of several program sequences of all types in any order are within the scope of the invention. This includes also the start and/or end of the motion program with an adaptively influenceable program sequence or a selection of different and part-dependent sequences to gain a higher flexibility. 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 a hardware embodiment to practice the method according to the present invention and Figure 2 shows an example for a movement path of the robot according to the method of the present invention

Fig. 1 shows an example 10 for a hardware embodiment to practice the method according to the present invention. A robot 14 is located within a first working booth 22. At the end of the robot arm 16 a gripper tool 18 is mounted, which holds a first work- piece 24. A human operator 12 has entered the booth 22 through the door 47 and impacts a force on a not shown sensor on the robot arm 16 so that he controls the robot movement by this force in a human robot cooperation task. In his other hand the operator holds a control device 48 connected to the robot controller 34, wherewith he is also able to control the robot movement. This has to be seen preferably as an exemplary alternative to the aforementioned control by force which is normally not used at the same time.

The first transfer zone 20 is supervised by a first 36 and a second 38 camera mounted at the upper booth. Also other sensors such as light barriers are suitable for supervision. The data signals of the cameras 36, 38 are transmitted with a first 40 and a second 42 connection line, for example a data transmission cable, to the robot controller 34. The robot controller 34 comprises a data processing device with suitable software for analyzing the picture data of the cameras 36, 38. Result of such an analysis is for example the information, that a person 12 is within the transfer zone 20 or not, it can also comprise the information, that a person 12 will probably enter the transfer zone 20 in a short time such as 0.5 s. Such predicted information can be based for example on the analysis of the movement of a person 12 which is outside the transfer zone 20 but nevertheless within a supervised area.

In any case any presence of a human operator 12 in the no-go zone, which is defined in the Fig. 1 as the working booth 22 except the transfer zone 20, has to be detected.

Of course it is possible to provide a dedicated data processing device separate from the robot controller for the purpose of processing and/or analyzing measurement data of sensors 36, 38. Also the alternative and/or additional use of different sensors for supervising the transfer zone and or adjacent zones is possible. This may be achieved by a combination of access gates, light curtains, pressure-sensitive mats, laser scanners, vision systems or other suitable means fulfilling the safety standards governing such detection.

A third connection line 44 connects the robot controller with the robot 14. The third connection line 44 comprises for example as well the power supply for the drives of the different axes from the robot 14. But also lines for a control signal for example from a force sensor mounted on the robot arm 16 to the robot controller 34 or lines for control signals from the robot controller 34 to the gripper tool 18 are foreseen.

The human operator and the end of the robot arm 16 are present within the first transfer zone 20. The first workpiece 24 is foreseen to be assembled with the third workpiece 28 in human-robot cooperation task. The third workpiece 28 is provided on a second transporting device 32, such as a conveyor. A second workpiece 26 which can be of the same type as the first workpiece 24 is provided on a first transporting device 30 at the other end of the booth 22. It is also thinkable that a larger number of such workpieces are provided in a common supply box for example.

The human operator 12 can signal the ready-status of the human-robot cooperation task by pushing the switching button 46 of a control device, which is located outside the booth 22. Alternatively, the means for signalling the ready-status of the robot 14 for beginning a program sequence in autonomous motion state can be of visual, acoustic or other suitable means perceivable by the operator 22.

Fig. 2 shows an example for a movement path of the robot according to the method of the present invention. The movement path shown is related to the tool center point (TCP) of a not shown robot at the end of its robot arm. The hardware embodiment shown in Fig. 1 corresponds in principle with the not shown hardware environment of Fig. 2. The working area of the not shown robot is limited by a second booth 78, whereas a second transfer zone is marked with the reference sign 76.

The movement path according to the movement program comprises several movement path sequences (62, 64, 66, 68, 70, 72). The movement executed by the robot starts at the first point A 52. It is assumed that the robot has gripped a workpiece. According to a first fixed robot motion program sequence, which is determined in movement path, movement speed and movement direction, the TCP moves along this path 62. A human operator is not present in the second transfer zone 76, so that the robot moves with full speed according to the first fixed motion program sequence, for example with 1200 mm/s, also when passing into the second transfer zone 76. The robot movement ends at the second point B 54.

In the case that a person would be present in the second transfer zone 76, the robot would have either never started the associated movement sequence, or would have executed a stop before passing into the transfer zone, or at least would have slowed down to an absolute person-safe speed of a few mm/s before passing into the second transfer zone 76. If a person is detected in the no-go zone within the second booth 78 and outside the second transfer zone 76, an emergency stop of the robot would be the consequence.

Now the robot is in a safe waiting position at the second point B 54. Waiting position can either mean that the robot does not execute any movement or that the robot executes a very slow and absolute safe movement, for example along a circular path. Now a human operator enters the second transfer zone 76 and takes over the control of the following adaptively influenceable motion program sequence 64, 66. He impacts for example a force on the end of the robot arm which is detected by appropriate sensors. The signals of the sensors are autonomously analyzed and the drives of the robot controlled in such a way that the robot arm moves in the direction of the applied force. Thus, the operator is able to precisely control the movement of the robot arm to the third point C 56. The robot is in a safe motion state so that for example the maximum movement speed does not exceed a certain limit of, for example, 250 mm/s.

While reaching the third point C 56, the workpieces to be assembled are moved into the desired position. Now the operator releases the workpiece by opening the gripper tool and the second part of the adaptively influenced robot program sequence starts. The operator moves the robot arm along the movement path 66 to the fourth point D 58. The movement path itself is not fixed by the program sequence. The operator is free to force any path within the given limits. First 64 and second 66 parts of the adaptively influenced robot program sequence can of course be assumed as a single program sequence.

It is also thinkable that the operator controls the human-robot cooperation by some verbal commands such as 'left', 'up', 'down', 'stop' or 'release' or with a portable control device comprising for example a joystick.

The fourth point D 58 is not necessarily the same point as the second point B 54, it is rather any point within the transfer zone 76 at which the operator wants to changeover the control of the robot program back to the robot.

The TCP of the robot is temporarily moved from the fourth point D 58 to the next determined point, at which the next motion sequence starts. In the example this point is the same as the second point B 54, from where the adaptively influenced robot sequence has started. Of course it can by any other point. The belonging movement path 68 is preferably autonomously generated by the robot respectively its control. In the easiest case, it is a linear movement path 68 in between both points D 58 and B 54. Dependent on the safety requirements of a human operator which still might be within the transfer zone, the robot movement might not be executed at all or at least in a safe motion state with very limited movement speed, such as a few mm/s. Normal procedure is that the operator has left the transfer zone and the robot executes the fixed program sequence without any restrictions.

In the next movement sequence, the robot TCP moves back to the first point A along the path 70, which is in this case identical with path 62, except for the movement direction. Now the robot moves along the path 72 to the fifth point E 60, where for example a new workpiece is retrieved, and moves back to the first point A 52 prepared to execute the whole robot movement program once again.

Of course, the sum of the fixed movement path 70, 72 and 74 can be assumed as one single motion program sequence, and also the sum of part 1 64 and the part 2 66 of the adaptively influenceable movement sequence can be assumed as one movement program sequence. The most important criterion for beginning a new movement sequence is that either the human operator takes over the control from the robot or gives it back to the robot.

List of reference signs

example for a hardware embodiment to practice the method according to the present invention operator robot robot arm gripper tool first transfer zone first working booth first workpiece second workpiece third workpiece first transporting device second transporting device robot controller first camera second camera first connection line second connection line third connection line switching button of a control device door control device example for a movement path of the robot according to the method of the present invention first point A second point B third point C fourth point D fifth point E movement path according to first fixed motion program sequence movement path part 1 according to adaptively influenceable motion program sequence movement path part 2 according to adaptively influenceable motion program sequence autonomously calculated movement path movement path according second fixed motion program sequence movement path according third fixed motion program sequence movement path according fourth fixed motion program sequence second transfer zone second first working booth