DE102019207144A1 - Method for recognizing an operator of a work machine - Google Patents

Abstract

The invention relates to a method for recognizing an operator of a work machine (1, 2) by means of a pattern recognition (101) of a personal operating pattern (100), with features (BM) being extracted from the operating pattern (100) of the current operator and these features ( BM) can be compared in a classification (103) with reference features (RM) of operating patterns (100) for previously trained operators in order to identify the current operator by assigning (104) to one of the previously trained operators (104).

Description

The present invention relates to a method for recognizing an operator of a work machine by means of pattern recognition of personal operating patterns. The invention also relates to an electronic control device which is set up to carry out the method according to the invention.

State of the art

Nowadays, work machines such as B. construction machinery, agricultural machinery, etc., a complex control by an operator. This must have a large number of operating elements, such as B. buttons, pedals, levers, control devices, touch screens, etc. operate, often simultaneously. When operating the respective working machine, each operator proceeds according to his own operating pattern in order to operate the operating elements. In addition to the experience of the operator, this person-related operating pattern is primarily shaped by his personal characteristics and preferences. As an example, an excavator's bucket can be controlled differently when lifting bulk material, by moving the bucket first in one direction and then the other in one case, moving the bucket in the reverse order in another case, and in still another case the excavator bucket is moved in both directions at the same time.

In the case of work machines in particular, it is important to only operate them with a valid operator qualification, for example a driver's license or participation in a seminar, as well as with sufficient skills. In practice, not every person who operates a work machine is also approved for it.

Methods used nowadays provide the operator with the aid of a registration means, e.g. B. a chip card to recognize. In addition, methods for recognizing people using a camera have also been developed. On the basis of this personal identification, personal access authorization for the working machine is then granted or withdrawn.

Independently of this, pattern recognition is known with which patterns can be recognized and compared. In the specialist book “H. Niemann: Classification of Patterns Springer, Berlin 1983 "the pattern identification is described in detail. The pattern recognition can be roughly summarized as follows: At the beginning, data and / or signals are recorded and patterns are derived from them. These patterns can be expressed mathematically as feature vectors and matrices. Features are extracted from the patterns, often after preprocessing, and it depends on the application which features are extracted. The features are typically transferred by means of transformation (for example discrete Fourier transformation) into a feature space in which the patterns are represented by points. Optionally, a feature reduction can then take place. Finally, the features are assigned to different classes by a classifier - often an artificial neural network.

Disclosure of the invention

A method for recognizing an operator of a work machine is proposed. Construction machinery, agricultural machinery, etc. are to be regarded as examples of work machines. Each operator operates the machine according to his personal operating pattern. That is, the operator controls controls of the work machine or parts of the work machine, such as. B. Tools, in the same or similar situations in the same or similar manner. Examples of the operating elements are buttons, pedals, levers, control devices, touchscreens, etc., further examples being given below. If the control of the operating elements as well as their overlapping, chronological sequence and / or combination are recorded regularly, the characteristics of the operator can be determined and the operating pattern can be related directly to the current operator. In addition to the experience of the operator, the personal operating pattern is primarily shaped by his personal characteristics and preferences. Accordingly, the operating pattern can be assigned directly to the operator and the personal operating pattern of each operator differs from others.

The operating pattern can be recorded or obtained by evaluating the control of the operating elements by the current operator. For this purpose, signals that are sent from the operating elements to an electronic control unit can be used to record the operating pattern. Alternatively or additionally, a behavior of the work machine or a system of the work machine, such as. B. a tool can be detected. For this purpose, signals from sensors can be evaluated. Inertial sensors, for example an inertial measurement unit (IMU), distance sensors, position sensors, inclination sensors, speed sensors, etc., may be mentioned as an example of such sensors. The signals evaluated for the detection of the operating pattern can go through preprocessing in which undesired and / or irrelevant signal parts, e.g. B. noise, can be reduced or removed. In the preprocessing z. B. an averaging Comparison with a threshold value and / or a normalization.

In this method, a known pattern recognition of the personal operating patterns takes place. In the pattern identification, features for the operation of the current operator are extracted from the recorded personal operating pattern, in particular from the signals mentioned. In the case of a classification as part of the pattern recognition, these characteristics of the current operator are compared with reference patterns for previously trained operators. The classification is carried out by a classifier, which is preferably an artificial neural network, and is preferably based on a machine learning process using random samples. During the classification, the features are converted into a number using automatically generated and nested weights, which indicates the probability that the current operator corresponds to one of the trained operators. If the probability for the previously trained operator is sufficiently high and is therefore above a threshold value, then the current operator is assigned to the previously trained operator and is thus recognized as that operator. If the probability does not reach a sufficient size for any of the previously trained operators, therefore it is below the threshold value for each of the previously trained operators, the current operator is recognized as a new operator who does not match any of the previously trained operators.

These reference patterns are recorded in a learning phase in which the current operator is known, i. H. otherwise identified or recognized by the process. The characteristics of the operator are learned from a training data set. This training data set includes the different characteristics of the operator and expresses them. In a development or application phase, the reference features are obtained from the training data set in a training step. This can be done locally in a control unit of the work machine. In order to be able to clearly assign a feature vector to the operator, the classifier must be trained. As a result, a large number of sample examples are required that represent the different characteristics of the operator under different environmental conditions, ideally under all environmental conditions. So that the characteristics of the operator can be determined using machine learning, a large number of other operators are provided for comparison. All operators who are intended for the respective work machine are preferably taught in beforehand. This then results in a rigid structure for pattern recognition and preprocessing based on fixed filter masks and weights.

This offers the advantage that the current operator can be identified solely on the basis of his actions. The detection takes place independently of cameras or registration means, such as B. chip cards or PIN, which, as explained in detail later, such registration means can be used for preselection. On the other hand, the operator is actually recognized and not just specific features or properties. Compared to methods which only recognize the ability and / or experience of the anonymous operator in question, this method offers the advantage that attributes that are already assigned to the previously trained operator can be adopted for the current operator.

It should be noted that the operating patterns for the same operator differ for different activities. So z. B. moves the excavator bucket in a different way when loading bulk material than when loading a solid block or digging a pit. In the following, a sequence of work steps of the work machine and in particular its tools, with which a specific task, for example digging the pit, is carried out, is regarded as a work sequence, wherein a work sequence can also consist of a single work step. According to one aspect, the operating pattern can be obtained for specific work sequences of the work machine. Accordingly, different operating patterns can also be obtained for different work sequences and these can be used independently of one another for pattern recognition in order to assign the current operator to the previously trained operator and thus to recognize him. In other words, the operating pattern represents the way in which the operator works in a specific work sequence. During such work sequences, switches, buttons, pedals, joysticks, touchscreens, etc. are mainly used as operating elements, depending on the work machine. In this case, only one of the operating patterns for a work sequence, but preferably several operating patterns for different work sequences, can be used in the pattern recognition and in the subsequent assignment.

According to a further aspect, the operating pattern can be obtained in a driving mode of the work machine. The operating pattern when driving is very different from the operating pattern when executing a work sequence. In particular, other controls, such as. B. Steering wheels, acceleration and brake pedals, clutches, etc. are controlled so that the characteristics of the operator are expanded to include additional attributes. In other words, the operating pattern represents a driving behavior of the operator during the driving mode. The driving mode, especially on one Road, can be recognized by one or more of the following factors: The position of a switch for the driving mode, a menu and / or software setting, the speed of the machine, possibly the decoupling of a rear-wheel drive, a position of tools or other machine parts, for example Example in the folded position, and / or based on the position of the machine that z. B. is determined via a navigation system. Only the operating pattern in driving mode can be used for pattern recognition. This is particularly advantageous if the work machine is being operated by a different person during the approach / departure than during the work. Preferably, both operating patterns in the driving mode and operating patterns for the work sequences described above are used for pattern recognition.

• Es kann eine Trajektorie der Arbeitsmaschine und/oder zumindest eines Werkzeugs der Arbeitsmaschine berücksichtigt werden. Insbesondere auch der zeitliche Verlauf der Trajektorie. Beispielsweise bedient ein Bediener, vor allem wenn dieser unerfahren ist, ein Werkzeug, das sich in mehrere Richtungen, z. B. mittels eines Joysticks, bedienen lässt, meist derart, dass es sich zuerst entlang einer ersten Richtung (z. B. vor und zurück) bewegt und im Anschluss entlang einer zweiten Richtung (z. B. Drehung nach rechts und links) und ggf. wieder im Anschluss entlang einer dritten Richtung (z. B. hoch und runter) bewegt, d. h. quasi eindimensional entlang der Achsen. Ein anderer Bediener, vor allem wenn dieser erfahrener ist, überlagert die Bewegungen in zwei oder ggf. drei Richtungen. Zudem kann die Gleichförmigkeit der Trajektorie berücksichtigt werden.

In the following, criteria are shown that can be taken into account in pattern recognition. Individual or multiple criteria can be taken into account at the same time, as well as other criteria that are not listed here:

• A trajectory of the work machine and / or at least one tool of the work machine can be taken into account. In particular, the course of the trajectory over time. For example, an operator, especially if he is inexperienced, operates a tool that moves in several directions, e.g. B. by means of a joystick, can be operated, usually in such a way that it first moves along a first direction (e.g. back and forth) and then along a second direction (e.g. turning to the right and left) and possibly then moved again along a third direction (e.g. up and down), ie quasi one-dimensionally along the axes. Another operator, especially if he is more experienced, superimposes the movements in two or possibly three directions. In addition, the uniformity of the trajectory can be taken into account.

The range of motion of the work machine and / or of the at least one tool can be taken into account. A movement space can be understood as the sum of the trajectories of the work machine and / or of the at least one tool with the external dimensions - hence the contour - of the work machine and / or of the at least one tool. In order to determine the range of motion of the work machine, the outer points of the contour of the work machine including the structures and / or the at least one tool are viewed spatially (three-dimensional) over a period of time, over a work step or over a work sequence. The position of the machine can be determined for this purpose. This can be used to avoid a collision with the environment, i. H. with people, the infrastructure, other machines and / or objects. Advantageously, the work machine can be given a limited range of motion within which the work machine is allowed to move. The limited range of motion can preferably also be predetermined in advance.

In order to determine the movement space of the at least one tool, the outer points of the contour of only the tool are viewed spatially superimposed over a period of time, over a work step or over a work sequence. This can be used to find a collision-free and efficient trajectory of the tool when guiding the tool. Likewise, a limited range of motion for the at least one tool can advantageously be specified. In addition, it can be determined at what time and at what location the tool actively comes into contact with a workpiece or material. As an example, in the case of an excavator with a bucket, it can be determined when and where a cutting edge of the bucket is actively cutting material from the ground and when and where this material is dumped by tilting the bucket. The removed material and thus the dimensions of the excavated pit can then also be determined from this.

The guidance of the at least one tool can be taken into account. It can be determined for each tool or in combination how many iterations and / or corrections and / or intermediate stops and / or detours etc. were required in order to move the tool to the predetermined target position. For this purpose, a counter can preferably be provided which increases with the iterations or corrections. Using the trajectory already mentioned, detours, intermediate stops, reversals, repetitions of movements and / or circumnavigations of the target position can be taken into account.

The efficiency of the work sequence can be taken into account. This also includes how often the operator has to repeat a work step before the work is completed. As a specific example in the case of an excavator, it can be recorded how fully the excavator shovel was loaded and how often the loading process must be repeated before the material to be loaded has been completely picked up.

The precision of a work sequence can be taken into account. It is checked to what extent the operator has adhered to the specified target data for the work sequence or the workpiece to be processed. The target data can be specified, for example, using BIM (Building Information Modeling) and implemented using 3D localization data (GNSS). It it is then measured whether these target data have been exceeded and / or undercut. For this purpose, measurements can be carried out on the one hand, e.g. B. by means of a camera and / or LIDAR, on the other hand, the sum of the movement spaces of the tool can be evaluated. As an example, target data for a pit to be excavated are specified by the BIM and made available to the excavator operator using the 3D localization data. The dimensions of the pit are typically specified as target data, but the evenness of the pit can also be included in the assessment. After the operator has excavated the pit and / or already during the excavation, it is measured whether the target data have been complied with and, if not, how large the deviation is.

It can be taken into account how far the operator takes the machine to the stability limits and whether he exceeds them. The stability limit can e.g. B. be a permissible inclination of the work machine on uneven ground.

It can be taken into account how high the workload of the at least one tool is and whether a permissible maximum load is exceeded and for how long.

The presence of objects and / or people in the work area can be taken into account. The operator controls the work machine and the at least one tool with greater caution and therefore differently when objects within the work area are in the way and / or when people enter or stay in the work area. The object and / or the person can preferably be recognized by the system of the work machine itself or from outside through machine vision. In addition, the trajectories already mentioned can be used to assess whether the work machine and / or the at least one tool was on a collision course with the object or the person or whether critical movements, such as e.g. B. has performed a braking operation or evasive movements.

The dynamics, the speed and / or the acceleration, also as a deceleration when braking, can be taken into account for the work machine and / or the at least one tool. On the one hand, the time required for a work sequence can be recorded; this can also be combined with the efficiency mentioned above. On the other hand, unbraked movements of the tools against (mechanical) stops can be recorded. An operator who knows the inertia and the delay of the tool (also for different loads) can estimate the (future) movement of the tool and counteract it accordingly at an early stage (anticipatory / predictive control). This predictive control can also be captured. In the example of an excavator, the excavator arm and / or the load can overshoot, which is avoided by the operator.

The starting procedure for the driven machine can be taken into account. The start procedure is different for each operator, be it in the order or the time for individual steps or between steps.

According to one aspect, it can be provided that, after the current operator has been assigned to the trained operator, the reference features of the corresponding trained operator are updated by the current features that have been extracted from the operating pattern of the current operator. As a result, changes in the operating pattern for the respective driver can be adopted in the reference features. The changes can result from the fact that the operator is more familiar with the work machine.

The identification of the operator can advantageously be used to manage personal access authorizations for operating the work machine and / or at least one tool. Each trained operator is assigned his or her access rights. The access authorizations can be restricted to certain working machines or even further to certain functions and / or setting options of the working machine. For example, a trained operator can only have access authorization to drive a work machine, but not to carry out a work step with it. The access authorization can mainly depend on the qualifications of the operator. The operator has a qualification such. B. obtained a driver's license or attended a seminar. It is also possible to take into account the experience of the operator when granting access.

If the current operator is assigned to one of the trained operators, the current operator is also granted the appropriate access authorization for the assigned trained operator. This also means that certain functions and / or certain setting options, right through to certain operating elements, are blocked for the current operator if the assigned, trained operator does not have the appropriate access authorization. As a result, each operator is granted or prevented access authorization to operate the machine and / or the at least one tool in accordance with his qualifications and, if necessary, his experience. The machine can also be stopped or only to certain work areas or Operations are restricted. With the access authorization, only individual functions, work processes and / or setting options can be released or blocked. It is also possible to simply adjust the speed, power, forces, etc. of the working machine and throttle them if necessary. If the work machines are also networked, other operators / drivers in other machines / vehicles and / or a control center can be notified and, if necessary, warned. The access authorization can also be granted or prevented for certain work sequences or in certain modes, for example the driving mode, independently of one another, whereby, as already described above, the operating patterns for the working sequences and, if necessary, for the driving mode are recorded and evaluated separately. For example, an operator can be granted access authorization to drive the work machine, but not to execute work sequences.

This type of access authorization, which is based on the recognition of the operator based on his actions, reduces misuse and misuse. On the one hand, people who are not intended to operate the machine cannot control it because of the access authorization. On the one hand, such persons can be workers who want to operate the machine despite a lack of qualification, which often leads to incorrect use. On the other hand, this also includes people who want to steal or otherwise misuse the machine. The main advantage here is that with this method, compared to conventional methods for access authorization, it is much more difficult, if not impossible, for these people to pretend to be a person with appropriate access authorization. A key or registration means, such as e.g. B. a chip card can be passed on or stolen. Processes for person recognition by means of a camera require a functioning and correctly adjusted camera; it is also possible to bypass this type of person recognition.

On the other hand, the constitution of a person can be observed and in particular a change in the constitution can be recorded. People behave differently due to tiredness, stress and excessive demands as well as under the influence of alcohol or other drugs. Such a changed constitution of the operator is also reflected in his operating pattern. Even if the current operator was previously taught in, the features of his operating pattern differ from the reference features due to the factors mentioned above and the current operator is not clearly recognized. Depending on how much the constitution of the operator has changed, appropriate measures, as already described above, can be applied and z. B. the work machine are stopped or only individual functions or work processes are blocked or speed, power, forces, etc. of the work machine are reduced.

However, in order to enable the current operator to operate the work machine in order to be able to record an operating pattern at all, it can be provided according to one aspect to only prevent the access authorization after the current operator has been assigned or cannot be assigned. The work machine is therefore approved in advance. In this case it is preferred to choose the number of criteria (see above) in the classification as small as possible. Ideally, the classification is only reduced to criteria of the start procedure.

According to one aspect, the trained operators can be preselected during pattern recognition. This reduces the number of potential trained operators. The characteristics in the classification are then only compared with reference characteristics for the trained operators selected in the preselection. This simplifies the classification and can be carried out more quickly.

The preselection is preferably carried out by a registration means, e.g. B. a chip card. The current operator registers with the registration means on the work machine. This preferably reduces the number of trained operators who can be used for the comparison to one. In this case it only needs to be determined whether the current operator corresponds to the trained operator who has registered or not. The registration means is accordingly used for a first release and the method offers verification when the registration means are used for personal identification.

Through the preselection, in particular through the registration means, the phase in which the working machine is generally released can be reduced or even omitted. The access authorizations assigned to the registered operator can be taken into account here.

The features which are extracted from the operating pattern of the current operator are advantageously stored on a tachograph assigned to the current operator. This means that the characteristics are safely stored as data and correctly assigned.

The computer program is set up to carry out each step of the method, in particular if it is carried out on a computing device or control device. It enables the implementation of the method in a conventional electronic control unit without having to make structural changes to it. For this purpose it is stored on the machine-readable storage medium.

By uploading the computer program to a conventional electronic control device, the electronic control device is obtained which is set up to carry out the method for recognizing an operator of a work machine by means of pattern recognition of personal operating patterns.

Figure list

• 1 zeigt eine Innenansicht eines ersten Ausführungsbeispiels einer Arbeitsmaschine mit diversen Bedienelementen.
• 2 zeigt eine Innenansicht eines zweiten Ausführungsbeispiels einer Arbeitsmaschine mit diversen Bedienelementen.
• 3 zeigt ein Ablaufdiagramm einer ersten Ausführungsform des erfindungsgemäßen Verfahrens zur Erkennung eines Bedieners.
• 4 zeigt ein Ablaufdiagramm einer zweiten Ausführungsform des erfindungsgemäßen Verfahrens zur Erkennung eines Bedieners und einer Zugriffsberechtigu ng.
• 5 zeigt ein Ablaufdiagramm einer dritten Ausführungsform des erfindungsgemäßen Verfahrens zur Erkennung eines Bedieners und einer Zugriffsberechtigu ng.

Exemplary embodiments of the invention are shown in the drawings and explained in more detail in the description below.

• 1 shows an interior view of a first exemplary embodiment of a work machine with various operating elements.
• 2 shows an interior view of a second exemplary embodiment of a work machine with various operating elements.
• 3 shows a flow chart of a first embodiment of the inventive method for recognizing an operator.
• 4th shows a flow chart of a second embodiment of the inventive method for recognizing an operator and an access authorization.
• 5 shows a flowchart of a third embodiment of the method according to the invention for recognizing an operator and an access authorization.

Embodiments of the invention

1 Figure 11 shows an interior view of a cab of a typical agricultural machine 1 , specifically a combine harvester, as a first exemplary embodiment of a work machine. There are various control elements 10-15 shown, which are operated by an operator (not shown here) to the agricultural machine 1 and / or a tool 20th (visible here from the cab) of the agricultural machine 1 to operate and thus to control. The typical operating elements here are switches 10 , Buttons 11 (especially knobs), a joystick 12 (Lever with additional controls), a touchscreen 13 , a steering wheel 14th and pedals 15th shown. Especially the steering wheel 14th and the pedals 15th the operator uses a drive mode in which he is during the approach to the work area and the departure, whereas the joystick 12 mainly used to carry out work sequences. The touchscreen 13 serves both as a display and for setting. With the switches 10 and buttons 11 various functions can be controlled.

2 Figure 11 shows an interior view of a cab of a typical construction machine 2 , specifically a shovel loader, as a second embodiment of the work machine. Here, too, there are various controls 10-14 shown, which are operated by an operator to the construction machine 2 and / or a tool (not visible here from the cabin) of the construction machine 2 to operate and thus to control. There are switches here too 10 , Buttons 11 , a joystick 12 and a steering wheel 14th shown.

It should be noted here that the operating elements described can be provided for all types of work machines and are usually provided depending on the manufacturer and equipment of the work machine. The 1 and 2 are primarily intended to provide an overview of the various operating elements that can be provided in a work machine and that are operated by the operator.

In 3 a flowchart of an embodiment of the method according to the invention is shown, in which the recognition of the current operator is shown. At the beginning there is an operating pattern 100 from sensor signals S _Se and from operating signals S _B , which are from the operating elements 10-15 are transmitted to an electronic control unit, determined. The operating signals S _B reflect, in particular, an activation, a change and / or an actuation by the current operator. Will the operating pattern 100 recorded in a driving mode of the implement, operating signals S _{B of} the steering wheel are primarily 14th and the pedals 15th as well as relevant switches, if applicable 10 , Buttons 11 and the touchscreen 13 detected. Will the operating pattern 100 Recorded during a work sequence, then operating signals S _{B of} the joystick are primarily 12 , as well as relevant switches 10 , Buttons 11 and the touchscreen 13 detected. In the course of a work sequence, the implement can also be driven, so that the operating signals S _{B of} the steering wheel are also driven 14th and the pedals 15th can be taken into account. The division described above is intended only as an example, but not exclusively. Relevant sensor signals S _Se can be evaluated in the same way.

• - Eine Trajektorie der Arbeitsmaschine und/oder zumindest eines Werkzeugs der Arbeitsmaschine;
• - der Bewegungsraum der Arbeitsmaschine und/oder des zumindest einen Werkzeugs der Arbeitsmaschine;
• - die Zielführung des zumindest einen Werkzeugs;
• - die Effizienz bei einer Arbeitssequenz;
• - die Präzision bei einer Arbeitssequenz;
• - die Stabilität der Arbeitsmaschine;
• - die Arbeitslast des zumindest einen Werkzeugs;
• - das Vorhandensein von Objekten und/oder Personen im Arbeitsbereich;
• - die Dynamik, Geschwindigkeit und/oder Beschleunigung der Arbeitsmaschine und/oder des zumindest einen Werkzeugs; und/oder
• - die Startprozedur für die Arbeitsmaschine.

From the operating pattern 100 of the current operator are known per se and described in the introduction 101 Features BM of the current operator extracted. In the Pattern recognition 101 the following criteria are considered individually or in combination:

• - A trajectory of the work machine and / or at least one tool of the work machine;
• the movement space of the work machine and / or of the at least one tool of the work machine;
• - the guidance of the at least one tool;
• - the efficiency of a work sequence;
• - the precision in a work sequence;
• - the stability of the work machine;
• the workload of the at least one tool;
• - the presence of objects and / or people in the work area;
• the dynamics, speed and / or acceleration of the work machine and / or of the at least one tool; and or
• - the starting procedure for the working machine.

Operators who are to be used were the work machines in advance 1 , 2 to use in a database 102 learned (not shown). A pattern recognition can also be used for this, the operators being known. The features extracted in the process are assigned to the known operators by an artificial neural network and are later used as reference features RM. The BM characteristics of the current user can be stored on a tachograph (not shown) which is assigned to the current operator,

The characteristics BM of the current operator are in a classification 103 by an artificial neural network with the reference features RM, which are assigned to the previously trained operators. On the basis of this comparison, either an assignment is made 105 of the current operator to the previously trained operator, whereby the current operator was recognized, or the current operator cannot be assigned to any previously trained operator 105 . Even then, the current operator cannot be assigned 105 if their constitution has changed, for example due to tiredness, stress, excessive demands or under the influence of alcohol or other drugs. Then the operating pattern changes 100 so that the features BM no longer match the reference features RM.

A valid assignment is made 105 of the current operator to the previously trained operator, reference features RM for this operator are updated with the currently determined features BM 106 .

The 4th and 5 each show a flowchart for two exemplary embodiments of the method according to the invention, each of which has an access authorization 201 , 301 depending on the assignment 105 of the current operator is managed by the previously trained operator. In 4th becomes the working machine 1 , 2 at the beginning in one state 200 controlled in which each operator controls the working machine 1 , 2 and the tools 20th can control across the board. This allows the operating pattern 100 of the current operator are recorded, preferably already during the start procedure of the machine 1 , 2 . This is followed by the recognition of the user, as in connection with 3 already described, executed. In the database 102 is an access authorization in addition to the reference features RM 201 saved for the previously trained operators and assigned to them. The qualification, e.g. B. a driver's license or a seminar attended, as well as experience with access authorization 201 be saved. The access authorization 201 can be used to operate the machine 1 , 2 and / or for the tool 20th as well as for certain functions and / or certain setting options. In addition, the access authorization 201 for the above-mentioned work sequences and the driving mode. A valid assignment is made 104 of the current operator to the trained operator, the current operator is given the access authorization 201 assigned to this trained operator. Can the current operator 201 are not assigned, he does not receive any access authorization 201 or only to a limited extent. Finally the waitress 202 the working machine 1 , 2 and / or the tool 20th as well as certain functions and / or setting options according to the access authorization 201 enabled or blocked.

In 5 Registration takes place at the beginning 300 of the current operator with a chip card or other means of registration. By registering 300 a pre-selection of the trained operator is made. If the registered operator has not been trained beforehand, the process is ended and an error message / signal is output if necessary. By registering 300 is the operation 302 the working machine 1 , 2 and / or the tool 20th right from the start according to the access authorization 201 enabled or disabled for the trained operator who has registered. The access authorization is as already described in the database 102 saved and assigned to the trained operator. When classifying 103 the characteristics of the current user are only compared with reference characteristics RM for the trained operator who has registered. This can accelerate this process significantly. The identification of the current operator is then used for verification 303 the assignment of the current operator to the trained operator who has registered. If a different person is recognized as the current operator than the trained operator who has registered, the machine can be operated afterwards 1 , 2 and / or the tool 20th stopped 304 will.

Claims (14)

Method for recognizing an operator of a work machine (1, 2) by means of pattern recognition (101) of a person-related operating pattern (100), with features (BM) being extracted from the operating pattern (100) of the current operator and these features (BM) in a classification (103) can be compared with reference features (RM) of operating patterns (100) for previously trained operators in order to identify the current operator by assigning (104) to one of the previously trained operators. Procedure according to Claim 1 , characterized in that the operating pattern (100) of the current operator and / or the previously trained operator based on the actuation of operating elements (10, 11, 12, 13, 14, 15) of the machine (1, 2) and / or by means of analysis a behavior of the work machine (1, 2) or a system of the work machine (1, 2) is obtained. Method according to one of the preceding claims, characterized in that the operating pattern (100) is obtained for certain work sequences of the work machine (1, 2). Method according to one of the preceding claims, characterized in that the operating pattern (100) is obtained in a driving mode of the work machine (1, 2). Method according to one of the preceding claims, characterized in that at least one of the following criteria is taken into account in the pattern recognition (101): A trajectory of the work machine (1, 2) and / or at least one tool (20) of the work machine (1, 2) ; - A movement space of the work machine (1, 2) and / or of the at least one tool (20) of the work machine (1, 2); - A route guidance of the at least one tool (20); - an e efficiency in a work sequence of the work machine (1, 2); - a precision in a work sequence of the work machine (1, 2); - A stability of the working machine (1, 2); - a workload of the at least one tool (20); - A presence of objects and / or people in a work area of the work machine (1, 2); - a dynamics, speed and / or acceleration of the machine (1, 2) and / or of the at least one tool (20); and / or - a starting procedure for the working machine (1, 2). Method according to one of the preceding claims, characterized in that, after the current operator has been assigned (104) to the trained operator, the reference features (RM) of the corresponding trained operator are replaced by the current features (BM) that are derived from the operating pattern (100) of the current operator must be updated (106). Method according to one of the preceding claims, characterized in that personal access authorizations (201, 301) are granted or prevented depending on the assignment (104, 105) of the current operator to the trained operator and that the operation (202, 302) of the machine ( 1, 2) and / or at least one tool (20) of the work machine (1, 2) is restricted according to the personal access authorization (201, 301) depending on the assignment (104, 105) of the current operator to the trained operator. Procedure according to Claim 7 , characterized in that the access authorization (201, 301) is only prevented after the current operator has been assigned (104) or cannot be assigned (105). Method according to one of the preceding claims, characterized in that the trained operators are preselected for the pattern recognition (101) and the features (BM) in a classification (103) are only compared with reference features (RM) for the trained operators selected in the preselection will. Procedure according to Claim 9 , characterized in that the preselection is made by a registration means with which the current operator registers (300). Method according to one of the preceding claims, characterized in that the features (BM) which are extracted from the operating pattern of the current operator are stored on a tachograph assigned to the current operator. Computer program which is set up, each step of the method according to one of the Claims 1 to 11 perform. Machine-readable storage medium on which a computer program is based Claim 12 is stored. Electronic control device which is set up to use a method according to one of the Claims 1 to 11 to recognize an operator of a work machine (1, 2) by means of a pattern recognition (101).

Images