System and method for monitoring a hazardous area of ​​a robot

Description

[0001] The present invention relates to a system for monitoring a danger zone of a robot according to the preamble of claim 1 and a method for monitoring a danger zone of a robot according to the preamble of claim 9.

[0002] The invention relates to the field of robot motion planning under runtime or real-time conditions. Path planning algorithms and kinematic algorithms are used to optimize robot trajectories with regard to time, energy consumption, wear, acceleration, speed, or momentum transfer to human body parts.

[0003] In particular, a distinction should be made between two scenarios. If a human is not present in a robot application, path planning should be carried out without safety technology criteria, i.e., without protecting the human. In this case, purely economic optimization criteria can apply. However, this changes fundamentally when a human is present. This human should be reliably detected by industrial peripheral sensors.

[0004] The technical specification ISO / TS 15066:2016 provides guidance for collaborative robot operation, where a robot system and humans share the same workspace. In such operations, the integrity of safety-related control systems is of particular importance, especially when controlling process parameters such as speed and force.

[0005] The ISO 10218-1 and ISO 10218-2 standards on the safety of industrial robots form a basis and provide guidance on the operating methods of collaborative robots.

[0006] DE 10 2004 043514 A1 discloses a device for controlling at least one safety-relevant function of a machine, comprising a machine control system for controlling the movements of the machine, at least one sensor for detecting an object within a monitoring area, and an evaluation unit for defining a danger zone and for triggering the safety-relevant function when the detected object enters the danger zone. To define the danger zone, the evaluation unit is coupled to the machine control system, and the evaluation unit is designed to derive the parameters required for defining the danger zone from the control signals used by the machine control system to control the movement of the machine. A corresponding method is also described.

[0007] DE 10 2016 007519 A1 discloses a method for monitoring a system with at least one robot, in particular multiple robots. The method comprises the steps of monitoring a distance range of at least one robot of the system and transferring this robot to a safe state based on detecting the position of at least one person within this distance range.

[0008] EP 3 099 452 B1 discloses a safety device for workers in the working area of ​​moving automatic machines, in particular industrial robots, wherein the safety device has a mobile detection device, which is to be carried by the worker, for locating the worker and an evaluation and control device, wherein the devices are in a preferably wireless communication connection and the evaluation and control device also has an interface for the input of data about the current danger zone of the automatic machine, in particular the industrial robot.

[0009] DE 10 2013 110 901 A1 discloses a technology for planning and implementing human-robot collaboration processes. A planning process involves creating a layout of a workspace in which a human and a manipulator interact within a collaboration area. Human-robot collaboration zones are identified, and, if necessary, body zones or parts at risk of collision are specified. Boundary geometries are recorded for the manipulator and / or a manipulator assembly, and a motion plan is recorded for at least one reference point. Using an automated evaluation module, a permissible process speed is determined for tracking the motion plan. If adhered to, only a permissible impairment of the human can be expected in the event of a collision between manipulator and human. A speed profile is created based on the determined permissible speeds.

[0010] One object of the invention is to provide an improved system for monitoring a robot's danger zone, comprising at least one sensor with at least one spatial monitoring area. A further object of the invention is to expand a solution space for productive and collaborative robot behavior.

[0011] The object is achieved according to claim 1 by a system for monitoring a danger zone of a robot, with at least one sensor with at least one spatial monitoring zone for monitoring the danger zone, and a control and evaluation unit, and a robot controller for controlling the movements of at least one dangerous part of the robot, wherein the robot controller and the control and evaluation unit are electronically connected to one another by means of at least one interface, wherein the sensor is designed to cyclically send at least 3D data of the monitoring zone to the control and evaluation unit, wherein the sensor and the control and evaluation unit are further designed to generate at least one spatial protection zone in the monitoring zone, wherein the control and evaluation unit is designed,To locate persons in the monitoring area of ​​the sensor based on the 3D data and to determine their distance from the dangerous part of the robot, wherein the control and evaluation unit is designed to arrange the spatial protection area such that the spatial protection area completely surrounds and encloses the dangerous part of the robot and a surface of the protection area forms an outer safety boundary, wherein the control and evaluation unit is designed to determine the position of the safety boundary depending on a distance, a direction of movement and / or a speed of movement of the person to the dangerous part of the robot, wherein the robot controller is designed to move the dangerous part of the robot freely within the protection area,wherein the robot controller and / or the control and evaluation unit is designed to take into account the volume of the person or the volume of individual extremities and to add additional 3D buffer zones based on the volume of the person or the volume of individual extremities, wherein the volume of the person or the volume of individual extremities is taken into account with the additional 3D buffer zones in the kinematics of the hazardous part of the robot, wherein these 3D buffer zones additionally enclose the protection area.

[0012] The object is further achieved according to claim 9 by a method for monitoring a danger zone of a robot, comprising at least one sensor with at least one spatial monitoring zone for monitoring the danger zone, and a control and evaluation unit, and a robot controller for controlling the movements of at least one dangerous part of the robot, wherein the robot controller and the control and evaluation unit are electronically connected to one another by means of at least one interface, wherein the sensor cyclically sends at least 3D data of the monitoring zone to the control and evaluation unit, wherein the sensor and the control and evaluation unit generate at least one spatial protection zone in the monitoring zone, wherein the control and evaluation unit locates persons in the monitoring zone of the sensor based on the 3D data and determines their distance from the dangerous part of the robot,wherein the control and evaluation unit arranges the spatial protection zone such that the spatial protection zone completely surrounds and encloses the hazardous part of the robot and a surface of the protection zone forms an outer safety boundary, wherein the control and evaluation unit determines the position of the safety boundary depending on a distance, a direction of movement and / or a speed of movement of the person to the hazardous part of the robot, wherein the robot controller moves the hazardous part of the robot freely within the protection zone, wherein the robot controller and / or the control and evaluation unit is designed to take into account the volume of the person or the volume of individual extremities and to add additional 3D buffer zones based on the volume of the person or the volume of individual extremities,where the volume of the person or the volume of individual extremities is taken into account with the additional 3D buffer zones in the kinematics of the dangerous part of the robot, and where these 3D buffer zones additionally enclose the protection area.

[0013] The control and evaluation unit is designed to locate objects or persons in the sensor's field of vision using the sensor's 3D data and to determine their distance from the dangerous moving part of the robot.

[0014] One or more sensors, especially 3D sensors, are preferably safety-certified. These capture 3D data of the monitored area in real time and with synchronization information and deliver the data to the control and evaluation unit.

[0015] A causal and dynamic adaptation of boundary conditions for robot motion planning can be achieved as follows: The periphery of an application is equipped with at least one sensor, for example, an industrial safety sensor. This sensor detects the presence of a person. If a person is detected, the boundary conditions for path planning are changed, for example, in a path planning program, if necessary.

[0016] According to the invention, the robot's workflows are maintained for as long as possible even in the presence of a person, enabling safe collaboration. The workflows are causally and dynamically adapted depending on whether and where a person is near the robot and has been detected by the sensor or peripheral sensors. Different strategies exist for this. They differ fundamentally in the spatial and temporal domains.

[0017] In the spatial domain, the first step is to eliminate crushing and shearing scenarios between the person and the moving, hazardous part of the robot, as these pose the greatest risk of injury. Finally, the goal is to prevent collisions with the robot or its hazardous parts.

[0018] In the time domain, all moving parts or dangerous parts of the robot can be slowed down or stopped to give the person or worker more time to move away or escape from a potentially dangerous situation.

[0019] If the danger no longer exists, it is intended, for example, that the robot's movement will restart.

[0020] Crushing injuries caused by moving parts of the robot can be avoided in the following way. For example, the protection zone or the volume within which possible trajectories can be calculated can be cleverly restricted. If, for example, a person is present, an additional 3D buffer zone can be added to the existing protection zone. This 3D buffer zone further encloses the protection zone. This 3D buffer zone helps prevent contact between the hazardous part of the robot and the person, since there is still enough space within the protection zone to avoid crushing injuries, regardless of all possible trajectories.

[0021] The 3D buffer zones added when a person is present can, for example, be linked to anthropometric data—e.g., the thickness of a hand, the thickness of an arm, the thickness of a leg, etc. Additional buffer zones can also be taken into account, which may result from the typical movement speed of human limbs. There are numerous standards for such anthropometric data, which also distinguish between people's age and ethnic origin.

[0022] Shearing caused by the robot's moving parts can be avoided in the following ways, for example: The potential shearing of human limbs in robot joints can be prevented. For example, certain joint angles of the robot can be restricted to prevent shearing. In particular, acute angles should not be permitted. It is intended that each robot axis can be assigned different restrictions.

[0023] For example, deceleration or stopping of the hazardous moving parts of the robot is provided. Finally, the relative speed between the person and the robot can be causally controlled. For example, measures from standards such as ISO 10218-2 or ISO / TS 15066:2016 are taken into account using the so-called speed and separation monitoring procedure. Unlike the procedures mentioned above, the restriction here is not in the spatial domain (e.g., the trajectory), but in the time domain, since deceleration or stopping takes place according to the trajectory and the movement is resumed depending on the situation, provided the hazard potential no longer exists.

[0024] For example, a hierarchy of hazard reduction measures is provided. For example, it is planned to use a control system or a "rule engine" that processes rules that determine under which causal conditions which of the hazard reduction measures described above can be used. If, for example, a person is still far away from the robot, less far-reaching methods such as the mean reflected mass (MRM concept) can be used initially. If the person comes closer, dangerous shear angles can be avoided, and the speed of the dangerous parts can be reduced. The protection area or volume for possible situational path planning is hardly restricted. Only when the person is in the immediate vicinity of the robot can crushing be avoided and the robot's freedom of movement restricted. An assessment can also be made as to whether, for example,There is enough time for certain mechanisms, such as "obtuse joint angles," to occur within the anticipated human approach time. If this is impossible, the robot must be stopped.

[0025] The measures described above allow the robot to maintain movement for significantly longer than would be possible if the robot were stopped or slowed down while following its path. This means that production can be maintained even in the presence of a person until a solution space or a protection zone for new robot paths allows movement. This also allows the robot to perform evasive maneuvers in the presence of the person. This also ensures ergonomic collaboration between people and robot. The frequency of hazardous conditions is also reduced. This, in turn, has a positive impact on the performance level of the entire safety application. If the person is not present, the full volume or the entire monitoring area can be utilized for path planning; the same applies to other restrictions.This allows optimization based purely on economic criteria. The measures described above can enable a new type of safety-oriented interaction between humans and robots.

[0026] According to the invention, the robot controller is designed to adapt the kinematics of the dangerous part of the robot in the protection area spatially and / or temporally.

[0027] In a further development of the invention, the control and evaluation unit is designed to cause at least the dangerous part of the robot to carry out an evasive movement if the distance of the person to the dangerous part of the robot falls below predefined distance values.

[0028] According to the further development of the invention, the control and evaluation unit is designed to redirect the dangerous part of the machine in order to maintain the productivity of the work process or the robot.

[0029] For example, the control and evaluation unit is designed to cause at least the dangerous part of the robot to avoid ramming into the person.

[0030] Avoiding collisions with the robot's moving parts can be achieved as follows, for example. The publication "Mean Reflected Mass: A Physically Interpretable Metric for Safety Assessment and Posture Optimization in Human-Robot Interaction" by STEINECKER et al., "Mean Reflected Mass: A Physically Interpretable Metric for Safety Assessment and Posture Optimization in Human-Robot Interaction," in: 2022 International Conference on Robotics and Automation (ICRA), IEEE, 2022, pp. 11209-11215, introduces and describes the robot's reflected mass. Along with the contact geometry and the relative velocity between the person and the robot, the reflected mass is one of the parameters with the greatest influence on the severity of human injuries in a collision. The reflected mass depends on the robot configuration and can be optimized, especially for kinematically redundant robots. The mean reflected mass (MRM) metric is independent of the contact / movement direction and enables the evaluation and optimization of the robot posture with regard to safety.In contrast to existing metrics, this one is physically interpretable, meaning it can be related to biomechanical injury data for a realistic and model-independent safety analysis.

[0031] The plan is now to develop algorithms for calculating the mean reflected mass and to apply them causally and context-specifically when a person is detected by the sensor. In the absence of the person, these algorithms are not activated, as this typically reduces productivity or cycle time in the application. What's interesting about this approach is that the concept of the mean reflected mass allows for a significantly better decision than the worst case scenario.

[0032] In a further development of the invention, the control and evaluation unit is designed to identify individual body parts or extremities of the person and to determine their distance, direction of movement and / or speed of movement to the dangerous part of the robot.

[0033] For example, the body parts or extremities at risk are identified. Based on the application context, it is possible to determine which body parts could be affected if a person is present. Accordingly, causal specifications or causal if-then specifications are made for the motion planning algorithm. In particular, if a calibrated sensor with a measuring spatial monitoring area is used, a measurement and estimation of the human limbs or body parts can also be performed. This measurement can also be used to verify the plausibility of the input parameters for the path planning calculation.

[0034] In a further development of the invention, the robot controller and / or the control and evaluation unit is designed to take into account biomechanical properties of the person or individual extremities in the kinematics of the dangerous part of the robot.

[0035] In particular, a distinction is made between permanent or static contacts and temporary contacts.

[0036] With permanent contacts, there is a risk of crushing in the event of a collision between a person and the robot, as a human body part can become trapped between the robot and the application surface.

[0037] With temporary or transient contact, the risk is not as high, as the person can still avoid the contact, for example, by simply being pushed to the side. Accordingly, limitations are less restrictive here.

[0038] Biomechanical limits according to DIN ISO / TS 15066 are defined to prevent biomechanical stresses caused by robot movement that have the potential to cause minor injury to an operator in the event of contact between the operator and the robot.

[0039] Pressure values ​​derived from conservative estimates of pain sensations obtained in studies can be used to determine transient pressure and force limits. The transferred energy resulting from hypothetical robot-human contact can then be modeled, assuming a completely inelastic contact between the robot and the operator and taking into account the robot's payload capacity and factors related to the operator's body part in contact. Once the transferred energy has been determined, recommendations can be developed for the speed limit related to robot movement in the collaboration space. This is done to keep the transferred energy below the threshold of minor injury to humans in the event of contact between the robot and the operator in the collaboration space.

[0040] Biomechanical limit values ​​according to DIN ISO / TS 15066, Table A.2 are defined as follows: Body region Specific body area Quasi-static contact Transient contact Maximum permissible pressure a< ps N / cm 2< Maximum permissible force b< N Factor for the maximum permissible pressure c< P T Factor for the maximum permissible force c< F T skull and forehead d< 1 Center of forehead 130 130 Not applicable Not applicable 2 temple 110 Not applicable Face d< 3 masticatory muscle 110 65 Not applicable Not applicable Neck 4 neck muscle 140 150 2 2 5 Seventh cervical vertebra 210 2 Back and shoulders 6 shoulder joint 160 210 2 2 7 Fifth lumbar vertebra 210 2 2 rib cage 8 sternum 120 140 2 2 9 pectoral muscle 170 2 Belly 10 abdominal muscles 140 110 2 2 pool 11 pelvic bones 210 180 2 2 Upper arms and elbow joints 12 deltoid muscle 190 150 2 2 13 humerus 220 2 Forearms and wrists 14 spoke 190 160 2 2 15 forearm muscle 180 2 16 arm nerve 180 2 Hands and fingers 17 Index fingertip D 300 140 2 2 18 Index fingertip ND 270 2 19 Index finger joint D 280 2 20 Index finger end joint ND 220 2 21 ball of the thumb 200 140 2 2 22 Palm D 260 2 23 Palm ND 260 2 24 Back of hand D 200 2 25 Back of hand ND 190 2 thighs and knees 26 thigh muscle 250 220 2 2 27 kneecap 220 2 lower leg 28 mid-shin 220 130 2 2 29 calf muscle 210 2

[0041] In a further development of the invention, the robot controller and / or the control and evaluation unit is designed to take into account biomechanical properties of the person or individual extremities and, based on the biomechanical properties of the person or individual extremities, to calculate permissible speeds of the dangerous part of the robot and to take them into account in the kinematics of the dangerous part of the robot.

[0042] According to the training, limit values ​​from biomechanical properties can be converted into permissible maximum robot speeds, provided that the body part, the contact scenario and the so-called effective robot mass are known.

[0043] Examples of speed limits calculated on the basis of the body model for transient contact according to DIN ISO / TS 15066, Table A.5 are defined as follows: Body region Speed ​​limit depending on the effective mass of the robot, based on the maximum pressure value at an area of ​​1 cm 2 < mm / s 1 2 5 10 15 20 Hand / Finger 2 400 2 200 2 000 2 000 2 000 1 900 forearm 2 200 1 800 1 500 1 400 1 400 1 300 upper arm 2 400 1 900 1 500 1 400 1 300 1 300 Belly 2 900 2 100 1 400 1 000 870 780 pool 2 700 1 900 1 300 930 800 720 Thigh 2 000 1 400 920 670 560 500 lower leg 1 700 1 200 800 580 490 440 Shoulder 1 700 1 200 790 590 500 450 rib cage 1 500 1 100 700 520 440 400

[0044] In a further development of the invention, the sensor is a time-of-flight sensor, a laser scanner with multiple scan planes, a time-of-flight camera, a stereo camera, an FMCW lidar sensor, an event camera, a radar sensor, an ultra-wideband radio sensor or an infrared camera.

[0045] Such sensors are suitable for effectively monitoring a spatial surveillance area.

[0046] Time-of-flight measurement systems enable distance measurement by determining the time difference between the emission of light and the return of the light reflected from the measuring object.

[0047] For example, the light-time-of-flight sensor operates according to a direct time-of-flight (dTOF) method, whereby short light pulses or groups of light pulses are emitted and the time until a remission or reflection of the light pulses is received by an object is measured. The light signals are formed by light pulses.

[0048] However, other time-of-flight methods are also possible, for example the phase method, according to which transmitted light is amplitude-modulated and a phase shift between transmitted and received light is determined, whereby the phase shift is also a measure of the time-of-flight (indirect time-of-flight method, iTOF).

[0049] Furthermore, a CW (Continuous Wave) method or the synonymous continuous wave method can be used, whereby a temporally constant light signal is used.

[0050] In this method, for example, the single photon events are distributed into two counters via a gating signal and a phase is calculated from the ratio of the counter readings.

[0051] A 3D camera, for example, monitors the surveillance area using a large number of recorded distance values. A 3D camera has the advantage that a volume-like protected area can be easily monitored.

[0052] A stereo camera, for example, monitors the surveillance area using a large number of recorded distance values. The distance values ​​are determined based on the two cameras of the stereo camera, which are mounted at a base distance from each other. A stereo camera also has the advantage of being able to monitor a volume-like protected area.

[0053] A time-of-flight camera determines distance values ​​based on the measured time of light, which is determined by an image sensor. A time-of-flight camera also has the advantage of being able to monitor a volume-like protected area.

[0054] For example, the sensor is designed as a frequency-modulated continuous wave (FMCW) LiDAR sensor.

[0055] In contrast to a LiDAR sensor based on time-of-flight measurement of laser pulses, an FMCW LiDAR sensor does not emit pulsed light beams into the monitored area, but rather continuous light beams. During a measurement, i.e., a discrete-time scanning of a measuring point in the monitored area, these beams exhibit a predetermined frequency modulation, i.e., a temporal change in the wavelength of the transmitted light. The measurement frequency is typically in the range of 10 to 30 Hz. The frequency modulation can, for example, be implemented as periodic up- and down-modulation. Compared to the emitted transmitted light, the transmitted light reflected from measuring points in the monitored area exhibits a time offset corresponding to the light travel time. This time offset depends on the distance of the measuring point from the sensor and is accompanied by a frequency shift due to the frequency modulation.In the FMCW LiDAR sensor, emitted and reflected light are coherently superimposed, allowing the distance of the measurement point from the sensor to be determined from the superposition signal. Compared to pulsed or amplitude-modulated incoherent LiDAR measurement principles, the coherent superposition measurement principle offers, among other advantages, increased immunity to ambient light from other optical sensors / sensor systems or the sun, for example. Compared to radar sensors with wavelengths in the millimeter range, the spatial resolution is improved, making it possible to measure a person's geometric characteristics.

[0056] If a measurement point moves toward or away from the sensor with a radial velocity, the reflected transmitted light also exhibits a Doppler shift. An FMCW LiDAR sensor can detect this change in the transmitted light frequency and use it to determine the distance and radial velocity of a measurement point in a single measurement, i.e., a single scan of a measurement point. Whereas a LiDAR sensor based on time-of-flight measurement of laser pulses requires at least two measurements, i.e., two temporally spaced scans of the same measurement point, to determine the radial velocity.

[0057] Event cameras output an asynchronous stream of events triggered by changes in the lighting situation. The pixels of an event camera respond independently to occurring brightness changes. Each pixel stores a reference brightness value and continuously compares it with the current brightness value. If the brightness difference exceeds a threshold, the pixel resets its reference value and generates an event: a discrete packet containing the pixel address and timestamp. Events can also contain the polarity (increase or decrease) of a brightness change or an instantaneous measurement of the illuminance.

[0058] The radar sensors, for example, form spatial monitoring zones for monitoring the protected area. The protected areas can have almost any geometry. For example, the protected areas are conical or club-shaped for spatial protected areas, starting from the radar sensor housing. For example, the opening angle of a protected area is + / -60°. Smaller or larger opening angles are also possible. However, with a sensor with more than one receiving antenna and / or transmitting antenna, rectangular or cuboid-shaped protected areas can also be formed.

[0059] For example, the radar sensor or each radar sensor with the receiving antenna emits radar waves in the frequency range from 40 GHz to 125 GHz. The frequency band of the radar sensor may be smaller than the specified frequency range.

[0060] For example, the sensor is an ultra-wideband radio sensor. The ultra-wideband radio sensor forms a radio location system, in particular an ultra-wideband radio location system, with the frequency used being in the range of 3.1 GHz to 10.6 GHz, and the maximum transmission power per radio station being 0.5 mW.

[0061] An absolute bandwidth of an ultra-wideband radiolocation system is at least 500 MHz or a relative bandwidth is at least 20% of the central frequency.

[0062] The range of such a radio location system is, for example, 0 to 50 m. The short duration of the radio pulses is used for location.

[0063] The radio tracking system therefore only emits low-energy radio waves. The system is highly flexible and interference-free.

[0064] Sensors used in safety technology must be particularly reliable and intrinsically safe and therefore meet high safety requirements, for example the EN13849 standard for machinery safety and the EN61496 device standard for electro-sensitive protective devices (ESPE).

[0065] To comply with these safety standards, a number of measures must be taken, such as secure electronic evaluation through redundant and / or diverse electronics or various functional monitoring systems, especially monitoring for contamination of optical components, including a front screen. A safety laser scanner complying with such standards is known, for example, from DE 43 40 756 A1.

[0066] The term "functionally safe" is to be understood in the sense of the aforementioned or comparable standards; i.e., measures have been taken to control errors up to a specified safety level. The safe sensor and / or at least one non-safe sensor also generate non-safe data, such as raw data, point clouds, or the like. "Non-safe" is the opposite of "safe," referring to non-safe devices, transmission paths, evaluations, and the like, and therefore the aforementioned fail-safe requirements are not met.

[0067] In a further development of the invention, the robot controller is designed to evaluate a 3D model of the environment and, based on the 3D model, to move the dangerous part of the robot within the protection area.

[0068] According to the training, for example, a CAD model of the application is available, which clearly shows which volume or geometric protection zone can be used for path planning. Of course, the aim is to prevent the robot from moving its hazardous moving part into the application's infrastructure, such as desks, conveyor belts, walls, etc. Accordingly, corresponding infrastructure is considered prohibited areas that must be avoided by the hazardous part of the robot.

[0069] According to the invention, the robot controller and / or the control and evaluation unit is designed to take into account the volume of the person or the volume of individual extremities and to add additional 3D buffer zones based on the volume of the person or the volume of individual extremities, wherein the volume of the person or the volume of individual extremities with the additional 3D buffer zones are taken into account in the kinematics of the dangerous part of the robot (3).

[0070] The 3D buffer zones added when a person is present can, for example, be linked to anthropometric data—e.g., the thickness of a hand, the thickness of an arm, the thickness of a leg, etc. Additional buffer zones can also be taken into account, which may result from the typical movement speed of human limbs. There are numerous standards for such anthropometric data, which also distinguish between people's age and ethnic origin.

[0071] In a further development of the invention, the control and evaluation unit is designed to compare the received 3D data of the surveillance area with known position data of the environment and to check for agreement.

[0072] According to the training, for example, a static environment is trained. This makes it easier to detect and track dynamically moving objects such as people.

[0073] For example, the robot is a mobile robot or a stationary robot.

[0074] The robot can, for example, be a multi-axis robot, such as an assembly robot in a production line.

[0075] The invention will be explained below with reference to further advantages and features, using exemplary embodiments, with reference to the accompanying drawings. The figures of the drawing show: Figure 1 a system for monitoring a danger zone of a robot with at least one sensor;

[0076] In the following figures, identical parts are provided with identical reference numerals.

[0077] Figure 1shows a system 1 for monitoring a danger zone 2 of a robot 3, with at least one sensor 4 with at least one spatial monitoring zone 5 for monitoring the danger zone 2, and a control and evaluation unit 6, and a robot controller 11 for controlling the movements of at least one dangerous part 9 of the robot 3, wherein the robot controller 11 and the control and evaluation unit 6 are electronically connected to one another by means of at least one interface 10, wherein the sensor 4 is designed to cyclically send at least 3D data of the monitoring zone 5 to the control and evaluation unit 6, wherein the sensor 4 and the control and evaluation unit 6 are further designed to generate at least one spatial protection zone 7 in the monitoring zone 5, wherein the control and evaluation unit 6 is designed,To locate persons 8 in the monitoring area 5 of the sensor 4 based on the 3D data and to determine their distance from the dangerous part 9 of the robot 3, wherein the control and evaluation unit 6 is designed to arrange the spatial protection area 7 such that the spatial protection area 7 completely surrounds and encloses the dangerous part 9 of the robot 3 and a surface 12 of the protection area 7 forms an outer safety boundary 13, wherein the position of the safety boundary 13 can be determined depending on a distance, a direction of movement and / or a speed of movement of the person 8 to the dangerous part 9 of the robot 3, wherein the robot controller 11 is designed to move the dangerous part 9 of the robot 3 freely within the protection area 7.

[0078] For example, the sensor 4 is a time-of-flight sensor, a laser scanner with multiple scan planes, a time-of-flight camera, a stereo camera, an FMCW lidar sensor, a radar sensor, an ultra-wideband radio sensor, or an infrared camera. Such sensors 4 are suitable for effectively monitoring a spatial surveillance area 5.

[0079] For example, robot 3 is a mobile robot or a stationary robot. Robot 3 can be a multi-axis robot, for example, an assembly robot in a production line.

[0080] The control and evaluation unit 6 is designed to localize objects or persons 8 in the field of view of the sensor 4 based on the 3D data of the sensor 4 and to determine their distance from the dangerous moving part 9 of the robot 3.

[0081] This sensor 4 detects the presence of a person 8 or a human. If a person 8 is detected, the boundary conditions for the path planning are changed, for example, in a path planning program, if necessary.

[0082] According to the invention, the workflows of robot 3 are maintained for as long as possible even in the presence of a person 8, enabling safe collaboration. The workflows are causally and dynamically adapted depending on whether and where a person 8 is in the vicinity of robot 3 and has been detected by sensor 4 or the peripheral sensors. Different strategies exist for this. They differ fundamentally in the spatial and temporal domains.

[0083] In the spatial domain, the first step is to eliminate crushing and shearing scenarios between the person 8 and the moving hazardous part 9 of the robot 3, as these pose the greatest risk of injury. Finally, the goal is to avoid ramming the robot 3 or the hazardous parts 9 of the robot 3.

[0084] In the time domain, all moving parts or dangerous parts of the robot 3 can be slowed down or stopped in order to give the person 8 or the worker more time to move away or free themselves from a potentially dangerous situation.

[0085] Avoiding crushing injuries caused by moving parts of the robot 3 can be achieved, for example, as follows. For example, it is intended to cleverly restrict the protective zone 7 or the volume in which possible trajectories may be calculated. If, for example, a person 8 is present, an additional 3D buffer zone can be added to the existing protective zone 7. This 3D buffer zone further encloses the protective zone 7. This 3D buffer zone helps to ensure that contact or touching between the dangerous part 9 of the robot 3 and the person 8 does not occur at all, since for all possible trajectories within the protective zone 7 there is still enough space to avoid crushing injuries.

[0086] The 3D buffer zones that are added when a person is present can, for example, be linked to anthropometric data - e.g., the thickness of a hand, the thickness of an arm, the thickness of a leg, etc. Additional buffer zones that can result from the typical movement speed of human limbs can also be taken into account.

[0087] Shearing caused by the moving parts 9 of robot 3 can be avoided, for example, as follows. The potential shearing of human limbs in robot joints can be prevented. For example, certain joint angles of robot 3 can be restricted to prevent shearing. In particular, acute angles should not be permitted. It is intended that each robot axis can be assigned different restrictions.

[0088] For example, a slowing down or stopping of the hazardous moving parts 9 of robot 3 is provided. Finally, the relative speed between person 8 and robot 3 can be causally controlled. Unlike the aforementioned methods, the restriction here is not in the spatial domain (e.g., the trajectory), but in the time domain, since the deceleration or stopping is carried out in accordance with the trajectory, and the movement is resumed situationally, provided the hazard potential no longer exists.

[0089] For example, a hierarchy of hazard mitigation measures is provided. For example, it is planned to use a control system or a "rule engine" that processes rules that determine under which causal conditions which of the hazard mitigation measures described above can be used. If, for example, a person 8 is still far away from robot 3, less far-reaching methods such as the mean reflected mass (MRM concept) can be applied initially. If person 8 comes closer, dangerous shear angles can be avoided, for example, and the speed of the hazardous parts 9 can be reduced. The protection zone 7 or volume for possible situational path planning is barely restricted for this purpose. Only when person 8 is in the immediate vicinity of robot 3 is crushing finally avoided and robot 3's freedom of movement restricted.An assessment can also be made as to whether, for example, there is enough time for certain mechanisms, such as "blunt joint angles," to be possible within the anticipated approach time of person 8. If this is not possible, for example, robot 3 must be stopped.

[0090] Thanks to the measures described above, the movement of robot 3 can be maintained for significantly longer than would be possible if robot 3 were stopped or if robot 3 were slowed down in accordance with its path. This allows production to be maintained even in the presence of a person 8 until a solution space or a protection zone 7 for new robot paths allows movement. This allows robot 3 to perform evasive maneuvers in the presence of person 8. This also ensures ergonomic cooperation between people 8 and robot 3.

[0091] According to the invention, the robot controller 11 is designed to adapt the kinematics of the dangerous part 9 of the robot 3 in the protection zone 7 spatially and / or temporally.

[0092] For example, the control and evaluation unit 6 is designed to cause at least the dangerous part 9 of the robot 3 to perform an evasive movement if the distance of the person 8 to the dangerous part 9 of the robot 3 falls below predefined distance values.

[0093] For example, the control and evaluation unit 6 is designed to identify individual body parts or extremities 15 of the person 8 and to determine their distance, direction of movement and / or speed of movement to the dangerous part 9 of the robot 3.

[0094] For example, the endangered body parts or extremities 15 are identified. From the application context, it can be deduced, for example, which body parts could be affected by the presence of a person 8. Accordingly, causal specifications or causal if-then specifications are made for the motion planning algorithm. In particular, if a calibrated sensor 4 with a measuring spatial monitoring area 5 is used, a measurement and estimation of the human limbs or body parts can also be performed. This measurement can also be used to verify the plausibility of the input parameters for the path planning calculation.

[0095] For example, the robot controller 11 and / or the control and evaluation unit 6 is designed to take into account biomechanical properties of the person 8 or individual extremities 15 in the kinematics of the dangerous part 9 of the robot 3.

[0096] In particular, a distinction is made between permanent or static contacts and temporary contacts.

[0097] In the case of permanent contacts, there is a risk of crushing in the event of a collision between person 8 and robot 3, as a human body part can be trapped between robot 3 and the application surface.

[0098] With temporary or transient contacts, the risk is not as high, as Person 8 can still avoid the contact, for example, if they are simply pushed to the side. Accordingly, limitations are less restrictive here.

[0099] For example, the robot controller 11 and / or the control and evaluation unit 6 is designed to take into account biomechanical properties of the person 8 or individual extremities 15 and, based on the biomechanical properties of the person 8 or individual extremities 15, to calculate permissible speeds of the dangerous part 9 of the robot 3 and to take them into account in the kinematics of the dangerous part 9 of the robot 3.

[0100] According to the training, limit values ​​from biomechanical properties can be converted into permissible maximum robot speeds, provided that the body part, the contact scenario and the so-called effective robot mass are known.

[0101] For example, the robot controller 11 is designed to evaluate a 3D model of the environment and, based on the 3D model, to move the dangerous part 9 of the robot 3 within the protection zone 7.

[0102] For example, a CAD model of the application is available, which clearly indicates which volume or geometric protection zone 7 can be used for path planning. Naturally, the aim is to prevent the robot 3 from moving its hazardous moving part 9 into the application's infrastructure, such as desks, conveyor belts, walls, etc. Accordingly, corresponding infrastructures are considered prohibited areas that must be avoided by the hazardous moving part 9 of the robot 3.

[0103] For example, the control and evaluation unit 6 is designed to compare the received 3D data of the monitoring area 5 with known position data of the environment and to check for consistency.

[0104] For example, a static environment is learned. This makes it easier to detect and track dynamically moving objects such as people. Reference symbol:

[0105] 1 System 2 Danger zone 3 Robot 4 Sensor 5 Spatial monitoring area 6 Control and evaluation unit 7 Protection zone 8 Persons 9 Dangerous part of the robot 10 Interface 11 Robot controller 12 Surface of the protection zone 13 Safety boundary 15 Extremities

Claims

1. A system (1) for monitoring a hazard zone (2) of a robot (3), having at least one sensor (4) having at least one spatial monitored zone (5) for monitoring the hazard zone (2); and a control and evaluation unit (6); and a robot controller (11) for controlling the movements of at least one hazardous part (9) of the robot (3), wherein the robot controller (11) and the control and evaluation unit (6) are electronically connected to one another by means of at least one interface (10); wherein the sensor (4) is configured to cyclically transmit at least 3D data of the monitored zone (5) to the control and evaluation unit (6); wherein the sensor (4) and the control and evaluation unit (6) are further configured to generate at least one spatial protected zone (7) in the monitored zone (5); wherein the control and evaluation unit (6) is configured to localize persons (8) in the monitored zone (5) of the sensor (4) with reference to the 3D data and to determine their distance from the hazardous part (9) of the robot (3), wherein the control and evaluation unit (6) is configured to arrange the spatial protected zone (7) such that the spatial protected zone (7) completely surrounds and includes the hazardous part (9) of the robot (3) and a surface (12) of the protected zone (7) forms an outer safety boundary (13), with the control and evaluation unit (6) being configured to fix the location of the safety boundary (13) in dependence on a distance, on a direction of movement and / or on a movement speed of the person (8) with respect to the hazardous part (9) of the robot (3), and wherein the robot controller (11) is configured to freely move the hazardous part (9) of the robot (3) within the protected zone (7), characterized in that the robot controller (11) and / or on the control and evaluation unit (6) is / are configured to take account of the volume of the person (8) or the volume of individual extremities (15) and to add additional 3D buffer zones based on the volume of the person (8) or the volume of individual extremities (15), with the volume of the person (8) or the volume of individual extremities (15) having the additional 3D buffer zones being considered in the kinematics of the hazardous part (9) of the robot (3),with said 3D buffer zone additionally surrounding the protected zone.

2. A system (1) in accordance with claim 1, characterized in that the control and evaluation unit (6) is configured to cause at least the hazardous part (9) of the robot (3) to perform an evasive movement if the distance of the person (8) from the hazardous part (9) of the robot (3) falls below predefined distance values.

3. A system (1) in accordance with any one of the preceding claims, characterized in that the control and evaluation unit (6) is configured to identify individual extremities (15) of the person (8) and to determine their distance, direction of movement, and / or speed movement with respect to the hazardous part (9) of the robot (3).

4. A system (1) in accordance with any one of the preceding claims, characterized in that the robot controller (11) and / or the control and evaluation unit (6) is / are configured to take account of biomechanical properties of the person (8) or of individual extremities (15) in the kinematics of the hazardous part (9) of the robot (3).

5. A system (1) in accordance with any one of the preceding claims, characterized in that the robot controller (11) and / or the control and evaluation unit (6) is / are configured to take account of biomechanical properties of the person (8) or of individual extremities (15) and to calculate permitted speeds of the hazardous part (9) of the robot (3) based on the biomechanical properties of the person (8) or of individual extremities (15) and to take them into account in the kinematics of the hazardous part (9) of the robot (3).

6. A system (1) in accordance with any one of the preceding claims, characterized in that the sensor (4) is a time of flight sensor, a laser scanner having a plurality of scan planes, a time of flight camera, a stereo camera, an FMCW LiDAR sensor, a radar sensor, an ultra-wideband radio sensor, or an infrared camera.

7. A system (1) in accordance with any one of the preceding claims, characterized in that the robot controller (11) is configured to evaluate a 3D model of the environment and to additionally move the hazardous part (9) of the robot (3) within the protected zone (7) starting from the 3D model.

8. A system (1) in accordance with any one of the preceding claims, characterized in that the control and evaluation unit (6) is configured to compare the received 3D data of the monitored zone (5) with known position data of the environment and to check them for agreement.

9. A method (1) of monitoring a hazard zone (2) of a robot (3), having at least one sensor (4) having at least one spatial monitored zone (5) for monitoring the hazard zone (2); and a control and evaluation unit (6); and a robot controller (11) for controlling the movements of at least one hazardous part (9) of the robot (3), wherein the robot controller (11) and the control and evaluation unit (6) are electronically connected to one another by means of at least one interface (10); wherein the sensor (4) cyclically transmits at least 3D data of the monitored zone (5) to the control and evaluation unit (6); wherein the sensor (4) and the control and evaluation unit (6) generate at least one spatial protected zone (7) in the monitored zone (5); wherein the control and evaluation unit (6) localizes persons (8) in the monitored zone (5) of the sensor (4) with reference to the 3D data and determines their distance from the hazardous part (9) of the robot (3), wherein the control and evaluation unit (6) arranges the spatial protected zone (7) such that the spatial protected zone (7) completely surrounds and includes the hazardous part (9) of the robot (3) and a surface of the protected zone (12) forms an outer safety boundary (13), with the control and evaluation unit (6) fixing the location of the safety boundary (13) being fixable in dependence on a distance, on a direction of movement and / or on a movement speed of the person (8) with respect to the hazardous part (9) of the robot (3), and with the robot controller (11) freely moving the hazardous part (9) of the robot (3) within the protected zone (7), characterized in that the robot controller (11) and / or on the control and evaluation unit (6) is / are configured to take account of the volume of the person (8) or the volume of individual extremities (15) and to add additional 3D buffer zones based on the volume of the person (8) or the volume of individual extremities (15), with the volume of the person (8) or the volume of individual extremities (15) having the additional 3D buffer zones being considered in the kinematics of the hazardous part (9) of the robot (3) and with said 3D buffer zone additionally surrounding the protected zone.