Method for accurately determining output torque and collaborative robot

Abstract

The present invention relates to a method for accurately determining an output torque, in particular an output torque of an actuator gear mechanism of a joint of a collaborative robot, by an artificial intelligence designed to output one or more output variables based on input variables, wherein the input variables of the artificial intelligence include a first angle specification corresponding to an input angular position and a second angle specification corresponding to an output angular position, and further wherein the output variables include an output torque determined by the artificial intelligence.

Description

[Technical Field]

[0001] The present invention relates generally to the technical fields of robotics, autonomous mobile robots (AMRs), particularly collaborative robots (cobots), and safety and precision engineering in workplaces and assembly systems. [Background technology]

[0002] Cobot systems are increasingly being used in facilities these days. This means that human workers increasingly share the same space as the cobot system. Such robots are specifically designed with collaborative characteristics. For example, if an object or a worker is in the way, they should first stop their movement to prevent damage and injury. Stress wave gearing (harmonic drives / harmonic gearboxes) is often used as the gearing mechanism. However, the choice of gearing mechanism is primarily a design decision.

[0003] For robust and accurate control of cobots and their joint motors, which makes them particularly suitable for use in precision environments and in work environments involving human workers, it is necessary to measure and / or know the delivery / output torque as accurately as possible.

[0004] Torque sensors are often used for this purpose. However, they are often expensive and often very heavy, hindering the flexibility of the robot during use and increasing energy consumption. Furthermore, torque sensors only provide one measurement value, which is not checked further. This can lead to uncontrolled and dangerous situations, especially if the torque sensor experiences a technical defect or emits a corrupted signal.

[0005] EP 2231369 B1 discloses a robot and a method for monitoring moments on such a robot. The robot includes at least two joints and parts that can move relative to each other by at least one joint. At least one torque sensor is located on at least one moving part. Two sensors and a redundant evaluation unit are provided for redundant torque detection. A device is provided for switching off the robot or triggering a safety state if the same torque measurements recorded by the at least two sensors deviate from each other outside a specified tolerance range. The two sensors, in the form of a full bridge with strain gauges, are located in the gear mechanism of the robot so that they detect the same torque. Two computer units, in the form of integrated circuits with different designs, are connected to a microcontroller in a transmitting unit, and a first check of the measured torque value is performed. However, this solution has the disadvantage that two sensors are required to detect torque, which is expensive and has a negative impact on the robot's dynamics and weight.

[0006] Another state-of-the-art solution is indirect determination via the motor current flow, which is inexpensive but very inaccurate and therefore no longer meets the current quality standards for high-performance cobot systems. Summary of the Invention [Problem to be solved by the invention]

[0007] It is therefore an object of the present invention to further develop collaborative robot systems to work accurately and safely, especially in shared working environments with humans, while ensuring high precision and flexibility of the robot and minimizing costs as much as possible without compromising the quality of the product and the results achieved. [Means for solving the problem]

[0008] The present invention provides a method according to claim 1. The method therefore relates to a method for accurately determining an output torque, in particular an output torque of an actuator gear mechanism of a joint of a collaborative robot, by means of an artificial intelligence designed to output one or more output variables based on input variables, wherein the input variables of the artificial intelligence comprise a first angle specification corresponding to an input angular position and a second angle specification corresponding to an output angular position, and wherein the output variables comprise an output torque determined by the artificial intelligence.

[0009] The artificial intelligence determines the torque output with a certain precision, which is higher than using traditional methods. As a result, this precise measurement allows the cobot control and motor control to function and operate more accurately.

[0010] The determination does not rely on direct measurement of torque, for example, by a torque sensor, which is often expensive and heavy. If such a torque sensor is nevertheless used, the present invention has the additional advantage that redundant determination of torque further increases accuracy. The direct output of the torque sensor is an excellent way to train AI. Furthermore, the solution according to the present invention can determine torque independently of the torque sensor, which is particularly advantageous, for example, in the event of a torque sensor failure. This avoids dead time and faulty and / or uncontrolled (and therefore often dangerous) cobot behavior.

[0011] The present invention further provides a robot, in particular a cobot (or a structural unit of a cobot), as set forth in claim 4. Thus, a cobot or cobot component is provided, comprising: a robot device, in particular a robot arm and / or assembly system, having one or more joints that can be moved by a motor, in particular a servomotor; at least one gear mechanism, in particular a reduction gear mechanism; one or more input angular position sensors, in particular input encoders, for determining an input angular position of the gear mechanism on the input side; one or more output angular position sensors, in particular output encoders, for determining an output angular position of the gear mechanism on the output side; and an electronic control system designed to take into account a value of the gear mechanism's output torque during operation of the cobot, provided by artificial intelligence, based on data including the input angular positions of the input angular position sensors and the output angular positions of the corresponding output angular position sensors. It should be emphasized that the concept disclosed herein is not limited to cobots and can be used with any type of robot.

[0012] The computations can be performed locally and / or non-locally, and can also be distributed, in particular in the context of cloud computing, for example. Cobots according to the invention, which are usually local, benefit from the invention through improved and more accurate behavior with improved operational safety.

[0013] Furthermore, the present invention provides an artificial intelligence system and a suitable training method.Accordingly, according to the present invention, there is provided an artificial intelligence for accurately determining an output torque, in particular an output torque of an actuator gear mechanism of a joint of a collaborative robot, which is designed to output one or more output variables based on input variables, wherein the input variables of the artificial intelligence include a first angle specification corresponding to an input angular position and a second angle specification corresponding to an output angular position, and the output variables include an output torque determined by the artificial intelligence.

[0014] AI can be trained continuously while in operation, which ensures a very high level of training and also accounts for changes that occur over time (e.g., changing environmental conditions or wear and tear).

[0015] Preferably, the training takes into account direct torque measurements as target values ​​(eg, measurements from a torque sensor).

[0016] Unsupervised learning has proven particularly suitable. In particular, good results have been achieved using multilayer perceptrons, especially recurrent multilayer perceptrons, as neural networks. The recurrent nature ensures sufficient feedback.

[0017] The AI ​​can be placed in a kind of control and / or feedback loop, effectively complementing unsupervised learning with torque measurements as target values.

[0018] The present invention significantly improves the dynamic usability of cobots in changing conditions: for example, high accuracy after replacing spare parts on a cobot (or, for example, a modified cobot) is guaranteed by the high flexibility of AI.

[0019] For example, the AI ​​may consider angular velocity, speed, acceleration, and temperature as additional features (input variables) to properly account for their effects, although the invention is by no means limited thereto.

[0020] Through AI, the present invention makes effects such as friction, wear, material fatigue, reversal margins, reactions, other nonlinear effects, manufacturing tolerances, and / or individual variations in manufacturing available for efficient evaluation and consideration in cobot control. These effects are often difficult to simulate, and even more rarely have access to direct and adequate mathematical descriptions. Artificial intelligence helps overcome these problems.

[0021] The present invention is in no way limited to the effects mentioned herein (in the sense of an exhaustive list). On the contrary, a particular advantage of using artificial intelligence in this case is that there is no need to manually model each individual effect, but rather all occurring effects, and all effects of the appropriate magnitude, are simultaneously reproduced directly by the AI.

[0022] Through accurate and optionally redundant determination of output torque, the present invention makes a lasting contribution to technological advances in the field of collaborative robots.

[0023] Preferred embodiments of the present disclosure are described below with reference to the following figures: [Brief explanation of the drawings]

[0024] [Figure 1] 1 shows a robot joint with input and output angular position encoders. [Figure 2] 1 shows a robot joint with an output torque sensor and an output angular position encoder. [Figure 3] FIG. 1 shows a schematic diagram of a drive train for a robotic joint with output torque sensors and angular position encoders on both sides in accordance with an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0025] FIG. 1 shows a robot joint 100 with input and output angular position encoders.

[0026] In this case, the input angle encoder or angular position encoder 1 is provided by an absolute encoder 1 .

[0027] Also provided is an output angle encoder or angular position encoder 2 by means of an incremental rotary encoder 2 .

[0028] This cobot can be utilized in the present invention due to its dual assembly with angular position encoders.

[0029] FIG. 2 shows a robot joint 100 with an output torque sensor 210 and an output angular position encoder 202 .

[0030] A stress wave gear mechanism, a wave gear mechanism, or a sliding wedge gear mechanism 203 is used to reduce the drive train of the robot joint.

[0031] FIG. 3 is a schematic diagram of a drive train for a robotic joint with output torque sensors and angular position encoders on each side in accordance with one embodiment of the present invention.

[0032] An electric motor 330 drives the rotation of the shaft 320. A voltage converter 340 provides the necessary energy, while a motor electronic control system 350 controls the motor 330. The shaft 320 terminates at the output side in a reducer 360. A gear mechanism, for example, provides the gear reduction.

[0033] A rotary encoder or angular position sensor 301 is located on the input side (input angular position sensor 301). Another rotary encoder or angular position sensor 302 is located on the output side (output angular position sensor 302). According to the invention, the torque to be output at the output side of the joint is determined by artificial intelligence from the measurements (and / or the difference between the measurements) of the angular position sensors 301, 302.

[0034] The torque can also be determined directly by a torque sensor on the output side. In particular, a redundant torque determination is possible, which is particularly accurate. Torque sensor failures are also prevented. Furthermore, the torque measured directly by the torque sensor can be used to train AI in an excellent way.

[0035] For example, in continuous motion, the AI ​​is further trained using a torque sensor.

[0036] After that, you can remove the torque sensor, for example temporarily, but it can also stay there: if it fails at some point, training will stop and only the best and most extensively trained AI (based on the output values ​​of the two encoders) will take over torque decisions.

[0037] The present invention has the further advantage that the AI ​​dynamically takes into account environmental effects (e.g., temperature), as well as signs of wear, material fatigue, component manufacturing tolerances, etc., particularly through continuous training.

[0038] The present invention therefore contributes to improved cobot control by being able to operate with much more accurately determined torque output.

[0039] The technical effect of the present invention therefore arises when torque determination is performed redundantly (by rotary encoder + AI and directly by torque sensor) as well as when the torque sensor fails, in which case downtimes and malfunctions of the cobot are particularly avoided.

[0040] This brings economic benefits.

[0041] The embodiments outlined herein can be further developed by many details of the invention, in particular those already mentioned above.

[0042] While some aspects have been described in the context of a device, it will be apparent that these aspects also constitute a description of a corresponding method, where a block or device corresponds to a method step or a function of a method step. Similarly, aspects described in the context of a method step also represent a description of a corresponding block or element, or feature of a corresponding device.

[0043] Exemplary embodiments of the present invention can be implemented in a computer system. The computer system may be a local computing device (e.g., a personal computer, laptop, tablet computer, mobile phone, or control device embedded in a robot) having one or more processors and one or more storage devices, or a distributed computing system (e.g., a cloud computing system having one or more processors or one or more storage devices distributed across various locations, e.g., local clients and / or one or more remote server farms and / or data centers). The computer system may comprise any circuit or combination of circuits. In an exemplary embodiment, the computer system may comprise one or more processors, which may be of any type. As used herein, a processor may refer to any type of computing circuit, such as, but not limited to, a microprocessor, a microcontroller, a complex instruction set computer (CISC), a reduced instruction set computer (RISC), a very long instruction word (VLIW) microprocessor, a graphics processor, a digital signal processor (DSP), a multi-core processor, a field programmable gate array (FPGA), or any other type of processor or processing circuit. Other types of circuitry that a computer system may include may be custom circuitry, application specific integrated circuits (ASICs), or the like, such as one or more circuits (e.g., communications circuitry) for use in wireless devices such as mobile phones, tablet computers, laptop computers, two-way radios, and similar electronic systems. A computer system may also include one or more storage devices, which may include one or more storage elements suitable for a particular application, such as main memory in the form of random access memory (RAM), one or more hard disks, and / or one or more drives that handle removable media such as CDs, flash memory cards, DVDs, and the like.The computer system may also include a display device, one or more speakers, and a keyboard and / or controller, which may include a mouse, trackball, touch screen, voice recognition device, or any other device that allows a system user to input information to and receive information from the computer system.

[0044] Some or all of the method steps may be performed by (or using) hardware devices such as a processor, microprocessor, programmable computer, or electronic circuitry, and in some exemplary embodiments, one or more of the significant method steps may be performed by such devices.

[0045] Depending on certain implementation requirements, exemplary embodiments of the present invention may be implemented using hardware or software. Implementations may also be performed using non-volatile storage media such as floppy disks, DVDs, Blu-ray discs, CDs, ROMs, PROMs and EPROMs, EEPROMs, or digital storage media such as FLASH memory on which electronically readable control signals are stored that interact (or can interact) with a programmable computer system to perform the respective methods. Thus, the digital storage medium may be computer-readable.

[0046] Some exemplary embodiments of the present invention comprise a data carrier with electronically readable control signals that can interact with a programmable computer system to perform one of the methods described herein.

[0047] In general, exemplary embodiments of the present invention may be implemented as a computer program product having program code that is effective to perform one of the methods when the computer program product is run on a computer. The program code may, for example, be stored on a machine-readable medium.

[0048] A further exemplary embodiment comprises the computer program for performing one of the methods described herein, stored on a machine-readable medium.

[0049] In other words, an example embodiment of the present invention is, therefore, a computer program with a program code for performing one of the methods described herein when the computer program runs on a computer.

[0050] Therefore, a further exemplary embodiment of the present invention is a storage medium (or data carrier or computer-readable medium) comprising a computer program stored thereon for performing one of the methods described herein when executed by a processor. The data carrier, digital storage medium or recording medium is typically tangible and / or non-transitory. A further exemplary embodiment of the present invention is a device as described herein, comprising a processor and a storage medium.

[0051] A further exemplary embodiment of the present invention is, therefore, a data stream or a sequence of signals representing the computer program for performing one of the methods described herein, for example the data stream or the sequence of signals being adapted to be transmitted via a data communication connection, for example via the Internet.

[0052] Another exemplary embodiment comprises a processing means, for example a computer, or a programmable logic device, configured to or adapted to perform any of the methods described herein.

[0053] A further exemplary embodiment comprises a computer having installed thereon the computer program for performing one of the methods described herein.

[0054] Another exemplary embodiment according to the present invention comprises a device or system configured to transmit (e.g., electronically or optically) a computer program for performing one of the methods described herein to a receiver. The receiver may be, for example, a computer, a mobile device, a storage device, etc. The device or system may comprise, for example, a file server for transmitting the computer program to the receiver.

[0055] In some exemplary embodiments, a programmable logic device (e.g., a field programmable gate array, FPGA) may be used to perform some or all of the functions of the methods described herein. In some exemplary embodiments, a field programmable gate array may cooperate with a microprocessor to perform any of the methods described herein. In general, it is preferred that the methods be performed by a respective hardware device.

[0056] Exemplary embodiments may be based on artificial intelligence, particularly the use of machine learning models or algorithms. Machine learning may refer to algorithms and statistical models that a computer system can use to perform a specific task without explicit instructions, rather than relying on models and inference. Machine learning, for example, may use transformations of data that can be derived from an analysis of historical data and / or training data, instead of rule-based transformations of data. For example, the content of images or other data, such as sensor data or measurement data, may be analyzed using a machine learning model or algorithm. To enable a machine learning model to analyze the content of images, the machine learning model may be trained using training images as input and training content information as output. By training a machine learning model with a large number of training images and / or training sequences (e.g., words or sentences) and associated training content information (e.g., labels or annotations), the machine learning model “learns” to recognize image content, such that image content not included in the training data can be recognized using the machine learning model. The same principle can be used for other types of sensor data. By training a machine learning model using training sensor data and a desired output, the machine learning model “learns” a transformation between sensor data and output, which can be used to provide an output based on non-training sensor data provided to the machine learning model. The provided data (e.g., sensor data, metadata, and / or image data) can be pre-processed to obtain feature vectors that are used as inputs for machine learning models.

[0057] A machine learning model can be trained using training input data. The above example uses a training method called supervised learning. In supervised learning, a machine learning model is trained using multiple training sample values, each of which can include multiple input data values ​​and multiple desired output values, i.e., each training sample value is associated with a desired output value. By specifying both the training sample values ​​and the desired output value, the machine learning model "learns" which output value should be provided based on input sample values ​​similar to the sample values ​​provided as part of training. In addition to supervised learning, semi-supervised learning can also be used. In semi-supervised learning, some of the training sample values ​​lack the desired output value. Supervised learning can be based on supervised learning algorithms (e.g., classification algorithms, regression algorithms, or similarity learning algorithms). Classification algorithms can be used when the output is restricted to a limited set of values ​​(categorical variables), i.e., when the input is classified as one value from a limited set of values. Regression algorithms can be used when the output indicates an arbitrary numerical value (within a range). Similarity learning algorithms can be similar to both classification and regression algorithms, but are based on learning from examples using a similarity function that measures how similar or related two objects are. In addition to supervised or semi-supervised learning, unsupervised learning can be used to train machine learning models. In unsupervised learning, input data (only) may be provided, and unsupervised learning algorithms can be used to find structure in the input data (e.g., by grouping or clustering the input data and finding commonalities in the data). Clustering is the assignment of input data containing multiple input values ​​into subsets (clusters) such that input values ​​within the same cluster are similar according to one or more (predetermined) similarity criteria, but dissimilar to input values ​​included in other clusters.

[0058] Reinforcement learning is a third group of machine learning algorithms. In other words, reinforcement learning can be used to train machine learning models. In reinforcement learning, one or more software actors (i.e., "software agents") are trained to perform actions in an environment. A reward is calculated based on the actions performed. Reinforcement learning is based on training one or more software agents to select actions that increase the cumulative reward, resulting in the software agents becoming better at the tasks given to them (as evidenced by an increasing reward).

[0059] Furthermore, some techniques can be applied to some machine learning algorithms. For example, feature learning can be used. In other words, a machine learning model may be trained at least in part using feature learning, and / or a machine learning algorithm may comprise a feature learning component. Feature learning algorithms, called representation learning algorithms, can preserve information in their input but transform it to make it useful, often as a preprocessing step before performing classification or prediction. Feature learning can be based on, for example, principal component analysis or cluster analysis.

[0060] In some examples, anomaly detection (i.e., outlier detection) may be used, which aims to identify input values ​​that are significantly different from the majority of the input and training data and thus raise suspicion. In other words, a machine learning model may be at least partially trained using anomaly detection, and / or a machine learning algorithm may comprise an anomaly detection component.

[0061] In some examples, a machine learning algorithm may use a decision tree as a predictive model. In other words, the machine learning model may be based on a decision tree. In a decision tree, observations about an object (e.g., a set of input values) may be represented by branches of the decision tree, and output values ​​corresponding to the object may be represented by leaves of the decision tree. A decision tree may support both discrete and continuous values ​​as output values. When discrete values ​​are used, the decision tree may be referred to as a classification tree. When continuous values ​​are used, the decision tree may be referred to as a regression tree.

[0062] Association rules are another technique that can be used in machine learning algorithms. In other words, a machine learning model can be based on one or more association rules. Association rules are created by identifying relationships between variables in a large data set. A machine learning algorithm can identify and / or use one or more ratio rules that represent knowledge derived from the data. The rules can be used, for example, to store, manipulate, or apply the knowledge.

[0063] Machine learning algorithms are typically based on machine learning models. In other words, the term "machine learning algorithm" may refer to a set of instructions that can be used to create, train, or use a machine learning model. The term "machine learning model" may refer to a data structure and / or set of rules that represent learned knowledge (e.g., based on training performed by a machine learning algorithm). In example embodiments, use of a machine learning algorithm may refer to use of an underlying machine learning model (or multiple underlying machine learning models). Use of a machine learning model may mean that the data structure / set of rules that make up the machine learning model are trained by a machine learning algorithm.

[0064] For example, the machine learning model can be an artificial neural network (ANN). ANNs are systems inspired by biological neural networks, such as those found in the retina or brain. ANNs comprise multiple interconnected nodes and multiple connections between the nodes, called edges. Typically, there are three types of nodes: input nodes that receive input values, hidden nodes that are (only) connected to other nodes, and output nodes that provide output values. Every node can correspond to an artificial neuron. Every edge can transmit information from one node to another. The output of a node can be defined as a (nonlinear) function of its inputs (e.g., the sum of its inputs). The input of a node can be used in a function based on the "weights" of the edges or nodes that provide the input. The weights of the nodes and / or edges can be adjusted as part of the learning process. In other words, training an artificial neural network can involve adjusting the weights of the nodes and / or edges of the artificial neural network, i.e., to achieve a desired output for a particular input.

[0065] Alternatively, the machine learning model may be a support vector machine, a random forest model, or a gradient boosting model. A support vector machine (i.e., a support vector network) is a supervised learning model with an associated learning algorithm that can be used to analyze data (e.g., in classification or regression analysis). A support vector machine can be trained by providing inputs with multiple training input values ​​that belong to one of two categories. A support vector machine can be trained to assign new input values ​​to one of two categories. Alternatively, the machine learning model may be a Bayesian network, which is a probabilistic directed acyclic graphical model. A Bayesian network can use a directed acyclic graph to represent a set of random variables and their conditional dependencies. Alternatively, the machine learning model or its training can be based on a genetic algorithm, which is a search algorithm and heuristic technique that mimics the process of natural selection. [Explanation of symbols]

[0066] 1 Angular position encoder / absolute encoder (input side) 2 Angular position encoder / incremental rotary encoder (output side) 3 Gear mechanism 20 Drive shaft 100 Robot Joints 202 Angular position encoder (output side) 203 Gear Mechanism 210 Torque sensor (output side) 301 Angle position encoder (input side) 302 Angular position encoder (output side) 310 Torque sensor (output side) 320 Drive shaft 330 Motor / Electric Motor 340 Voltage Converter / Voltage Source 350 Electronic Control System / Motor Unit 360 reducer

Claims

1. 1. A method for accurately determining an output torque, in particular an output torque of an actuator gear mechanism (203) of a joint (100) of a collaborative robot by means of artificial intelligence, in particular a machine learning model or algorithm, designed to output one or more output variables based on input variables, comprising: The input variables of the artificial intelligence are: a first angle specification corresponding to an input side angular position; a second angle specification corresponding to the output side angular position; Including, Further, the output variables include the output torque determined by the artificial intelligence.

2. 2. The method of claim 1, wherein the first angular specification is provided on the input side to the gear mechanism by an angular position sensor (301), in particular an encoder, in particular an absolute encoder, and the second angular specification is provided on the output side to the gear mechanism by an angular position sensor (302), in particular an encoder, in particular an incremental rotary encoder.

3. The method of claim 1 or 2, wherein the gear mechanism (203) comprises a stress wave gear mechanism, a wave gear mechanism and / or a sliding wedge gear mechanism.

4. a robotic device, in particular a robotic arm and / or assembly system, having one or more joints (100) that can be moved by a motor (330), in particular a servo motor; at least one gear mechanism (203), in particular a reduction gear mechanism; one or more input angular position sensors (301), in particular input encoders, for determining an input angular position relative to the gear mechanism at the input side; one or more output angular position sensors (302), in particular output encoders, for determining an output angular position relative to the gear mechanism on the output side; The value of the gear mechanism-side torque delivered during the operation of the cobot is provided by artificial intelligence, particularly a machine learning model or a machine learning algorithm, The input side angular position of the input side angular position sensor (301), the output angular position of the corresponding output angular position sensor (302); an electronic control system designed to take into account data including: A robot, in particular a cobot or cobot component, comprising:

5. The system further comprises a torque sensor (310), in particular an output torque sensor (310), wherein the output torque provided by the artificial intelligence, in particular a machine learning model or algorithm, and the torque detected by the torque sensor complement each other; especially, Redundancy in torque value determination is used to increase the accuracy of the value, and / or 5. The cobot or cobot component of claim 4, wherein the output torque provided by artificial intelligence is used in case the torque sensor (310) fails or provides an erroneous and / or corrupted signal.

6. 6. A cobot or cobot component according to claim 4 or 5, comprising at least one stress wave gear mechanism, wave gear mechanism and / or sliding wedge gear mechanism.

7. 7. A method for operating a cobot according to any one of claims 4 to 6, comprising a step of taking into account a delivery torque determined by artificial intelligence according to a method according to any one of claims 1 to 3.

8. A computer, computer system or computer network configured to carry out the method of any one of claims 1 to 3 or 7, locally or non-locally.

9. 1. An artificial intelligence, in particular a machine learning model or algorithm, for accurately determining an output torque, in particular an output torque of an actuator gear mechanism (203) of a joint (100) of a collaborative robot, the model or algorithm being designed to output one or more output variables based on input variables, The input variables of the artificial intelligence are: a first angle specification corresponding to an input side angular position; a second angle specification corresponding to the output side angular position; Including, Further, the output variables include the output torque determined by the artificial intelligence.

10. Artificial intelligence according to claim 9, specifically designed for unsupervised learning involving recurrent multi-layer perceptrons.

11. The artificial intelligence according to claim 9 or 10, wherein the input variables further comprise angular velocity, speed, acceleration and / or temperature.

12. The following datasets, namely: A data set containing data on the input and output angular positions of the gear mechanism (203).

12. A method for training an artificial intelligence according to any one of claims 9 to 11, in particular by unsupervised learning, to accurately determine the delivery torque using as input data

13. 13. The method according to claim 12, wherein a measured value of the output torque, in particular measured by a torque sensor (310), is further used as a setpoint, in particular as part of a feedback control loop in which the artificial intelligence is arranged.

14. 14. The method of claim 13, wherein the training is performed continuously, but the training is stopped and / or interrupted if it is determined that the value of the output torque measured by the torque sensor (310) is impaired or the corresponding signal is discontinued.

15. 15. A computer program comprising instructions that cause a computer to carry out a method according to any one of claims 1 to 4, 7 or 12 to 14 when the program is run by the computer.

16. taking into account, and for the purpose of taking into account, the effects of one or more of friction, wear, material fatigue, reversal margins, reactions, other non-linear effects, manufacturing tolerances, and / or individual variations in manufacturing; and / or for the purpose of modelling the dependence of environmental and state variables, in particular one or more of angular velocity, velocity, acceleration, temperature, by said artificial intelligence, Use of a cobot according to any one of claims 4 to 6 and / or a computer program according to claim 14 for determining an output torque.

Images