DE102019120655A1 - Method and system for generating individual training sequences in a simulation environment, in particular for sports and rehabilitation purposes - Google Patents

Abstract

A method for generating individual training sequences in a simulation environment, in particular for sports and rehabilitation purposes, has the following steps: Acquisition of user data via user input and sensors; Analysis of user data and determination of calculated body and movement data; Selection of training data corresponding to the calculated body and movement data from a first data memory, the physiological effect on the user being a selection criterion; and generating a training specification; Provision of exercise modules (exercise module) in a second data memory; Combination and arrangement of the selected exercise modules according to the requirements of the training specification; Generation of a virtual environment from a simulation logic selected by the user and environment visualization and 3D modeling; Start training; a) Acquisition of user information, such as body data and movement parameters, during training; b) ACTUAL / TARGET comparison of the recorded user information with the calculated data; c) Adapting the training specification to the TARGET data; d) Adapting the combination and arrangement of the selected exercise modules according to the specifications of the adapted training specification; e) adapting the virtual environment to the changed training modules; f) Executing the adapted combination and arrangement of the selected exercise modules and jumping to step i). A system for generating individual training sequences in a simulation environment is also specified. In this way, training processes are generated based on the desired effect and the simulation is individually adapted based on the user information during the training.

Description

The present invention relates to a method and a system for generating individual training sequences in a simulation environment, in particular for sports and rehabilitation purposes. The present invention discloses a method and a system which is able to generate tailor-made, virtual training sequences in a simulation environment which can be used for training and rehabilitation purposes. These training processes in a simulation environment are generated on the basis of properties as well as training and rehabilitation goals of a user. The proposed system is able to further adapt the training during the runtime, based on the analyzed user behavior and / or user movements.

Virtual reality technology has been around for several decades. In 1968 Ivan Sutherland built the first Virtual Reality Headmounted Display (HMD), called The Sword of Democles. Since then, attempts have been made to translate this technology into consumer products. The Oculus Rift started a new Virtual Reality (VR) hype in 2010. Since then, big tech companies like Facebook have invested billions of dollars in technology and marketing to grow the market. Compared to previous attempts to market these systems, several million devices such as the Oculus Rift, Oculus Go and Samsung GearVR have already been shipped. Customers are typically tech enthusiasts and early adopters. The current market is divided into three segments: Mobile VR, Standalone VR and High End VR. Mobile VR systems like the Samsung GearVR require the user to insert a smartphone into the headset. The smartphone serves as a screen for displaying the stereoscopic image and as a computing unit for executing VR applications. Stand-alone VR systems, on the other hand, can work completely independently, as they have an integrated display and a processing unit. High end VR systems are much better quality than any other VR system. They are the only VR systems that are able to capture 6DOF of a user. This means that they not only detect the orientation of the user's head, but also the three-dimensional position of each tracker such as the headset and hand control. High end VR systems can do two things pretty well: 1. They can capture the body of the user 2. They can immerse the user in a virtual world that is generated by a computer unit and encourage the user to perform certain actions with his body . These two possibilities are the reason why these VR systems are very well suited for training and rehabilitation purposes (e.g. training or rehabilitation).

A number of companies have already created virtual reality (VR) training experiences in the form of VR simulations that move the user. These simulations have already proven to be an effective workout on calories burned. However, these simulations have no control whatsoever over what is going on with the user. In some of the simulations, the user moves at random. Others expect the user to follow a predefined choreography that is generated manually. Both types of simulations do not allow the simulation to be adjusted for the user in terms of the intensity of muscle utilization, the torque that applies to certain joints and can cause damage, the desired amount of calories burned, improvements in reaction and speed, etc. . Since the choreographic simulations are also generated manually, the effort for setting up these simulations increases massively with a higher number of different movements and increasing complexity of the simulation.

These simulations also require that the user impart knowledge about training and rehabilitation aspects in order to plan and observe the training and avoid injuries during the training. In the case of rehabilitation, the user has to know what exactly supports his progress, the simulation knows nothing about his condition.

Further efforts have been made to overcome these disadvantages.

So reveals the US 2019046859 A1 a method, a computer system and a computer program product for sports training on an augmented reality (AR) device or a virtual reality device. The system can include the acquisition of a large number of user movement data using one or more sensors. Furthermore, the system can also include a measurement of data of a user body and a gaze position based on the plurality of captured user movement data. It can further include calculating a body position difference by comparing the measured body and gaze position of the user with a body position sequence specified by experts. The system may also include determining a body position quality threshold based on the calculated Body position difference is not met. The embodiment may further include generating an exercise program based on the compared, calculated body position difference.

The WO 2004109575 A1 discloses a system and a method for specifying exercise programs in which exercise machines are controlled to correct errors for the specified exercise program by measuring the pulse rate when the exerciser is exercising with exercise machines and the exerciser is being guided to exercise according to an updated exercise program suitable for measured workouts through actual exercises. The training program allocation system measures the pulse rate of a trainee in order to specify a training program for the trainee, and comprises: pulse rate measuring means for measuring the pulse rate of the trainee; and training program update means for calculating the maximum oxygen intake amount based on the pulse rate measured by the pulse rate measuring means and updating the training program for the exercise machine based on the maximum oxygen intake amount.

The WO2007058512 A1 discloses a gait trainer that provides an exercise program and an exercise program system that uses the gait trainer. The gait trainer includes a display, pedals on which the user's feet are placed, a control unit that receives physical information about the physical condition of the user while the pedals are moving, defines an optimal exercise program based on the physical information, and the predefined exercise programs on the display. The control unit accesses a server belonging to a rehabilitation organization via a network and defines the exercise program with the aid of the rehabilitation expertise received from the server. Since the exercise program suitable for the user is provided, the exercise efficiency is maximized and possible dangerous rehabilitation situations can be prevented.

The KR 20020030755 A discloses a treadmill system with online rehabilitation diagnostic functionality. A treadmill system is provided in which the floor acts as a sensor and transmits the user's physical data to the server. The rehabilitation data of the user are answered by the online computer, whereby the physical training and the rehabilitation management of the user are systematically and continuously supported. With the data, the user can select the exercise program that is suitable for the user. The system comprises: a program connection step in which the user logs into the physical exercise program of the computer connected to the treadmill via a keyboard; an inputting step of basic data for executing the physical exercise program, wherein user weight and body fat information are inputted through the floor and console unit of the treadmill; a pulse sensor attaching step of attaching the pulse sensor to the body or the guide rod to automatically supply the user's heart rate to the computer; a step of controlling the inclination of the floor, wherein the inclination of the floor of the treadmill is automatically controlled to adapt to the virtual reality road condition determined by the physical exercise program; a sports data measuring step in which the moving speed of the user through the treadmill detected by the rotation of the treadmill and the heart rate detected by the pulse sensor are sent to the computer; and a step of responding to the rehabilitation diagnosis and the physical condition of the user from the central management server to the user's computer by analyzing the sports data of the day.

It is therefore the object of the present invention to provide an improved method and an improved system for generating individual training sequences in a simulation environment for training and rehabilitation purposes. The aim of the present invention is to provide improvements in terms of customization, control, scaling, complexity and user-friendliness of such a method and system. Furthermore, training processes should be generated based on the desired effect and the simulation should be individually adapted based on the user information during the training.

This object is achieved by a method with the features of claim 1, a system with the features of claim 12 and a computer program with the features of claim 16. Advantageous developments and embodiments of the invention are specified in the dependent claims. In this way, training processes are generated based on the desired effect and the simulation is individually adapted based on the user information during the training. This can be of particular advantage if, due to physical limitations or disabilities, the execution of the exercises does not correspond to the calculated expectations and the desired effect may not or not fully occur.

The present invention discloses a method and a system which is able to generate individual training sequences in a simulation environment for training and rehabilitation purposes. The training units are automatically adapted by the system on the basis of properties and training and rehabilitation goals of a user and can actively control the movements and thus the way in which the training processes in the simulation environment influence the body of the user. The invention enables the complexity of these training processes to be scaled in the simulation environment and takes over all planning and control activities that require domain-specific knowledge, such as training and rehabilitation knowledge, and the generation of the simulation environment.

The method according to the invention for generating individual training sequences in a simulation environment, in particular for sports and rehabilitation purposes, has the following steps: acquisition of user data via user input and / or sensors; Here, for example, all user data, such as height, weight, body dimensions, pulse, blood pressure, blood oxygen, lactate value, etc. could be recorded automatically. However, a combination of user input in a corresponding input mask and sensor detection is also possible; Analysis of user data and determination of calculated body and movement data; When calculating the movement data, it is possible to use stored data records which already contain movement data in relation to body measurements; Selection of training data corresponding to the calculated body and movement data from a first data memory, the physiological effect on the user being a selection criterion; and generating a training specification; The selection of the training data takes place automatically, for example the muscles to be used and a corresponding duration of use being made on the basis of stored stress / effect data; Providing exercise modules in a second data memory; Here, for example, specific exercises for a defined muscle group can be stored as a module; Combination and arrangement of the selected exercise modules according to the requirements of the training specification; Generation of a virtual environment, environment visualization and 3D modeling from a simulation logic selected by the user; The simulation can be, for example, a role play or, for example, a virtual trainer / avatar; The simulation can be used, for example, to increase motivation, intuitively explain exercises and create better body tension; Start training; a) Acquisition of user information, such as body data and movement parameters, during training; Here, the detection can take place, for example, via sensors attached to the body, via optical motion renders or the like; b) ACTUAL / TARGET comparison of the recorded user information with the calculated data; c) Adapting the training specification to the TARGET data; d) Adapting the combination and arrangement of the selected exercise modules according to the specifications of the adapted training specification; e) adapting the virtual environment to the changed training modules; f) Executing the adapted combination and arrangement of the selected exercise modules and jumping to step a). This creates a closed control loop that contributes to the optimization of the method, to avoiding injuries to the user and to adapting to special needs of users with physical limitations or disabilities.

The system according to the invention for generating individual training sequences in a simulation environment has at least one processor, devices for capturing user data, a first data memory with training data that corresponds to user data, a second data memory with exercise modules, virtual reality or augmented reality hardware for outputting the dynamic simulation environment, sensors and / or a movement detection device for recording user information, such as body data and movement data during training.

According to the invention, a management system can be implemented in a software system. The management system collects data about the user, based on this data it can generate individual training. The training consists of reusable components, which we call exercise modules or assets. An exercise module describes an exercise and provides training and rehabilitation-specific knowledge that relates to the exercise (e.g. how will the execution of this exercise affect the user's body?). Based on these values, the management system is able to generate individual training. Since this is handled with a software system, these training processes can be scaled in a simulation environment or the experiences can be scaled without great effort (logarithmic growth rate of the development effort, depending on the complexity of the experiences). The domain-specific knowledge is presented within the exercise modules. Thus, the user does not have to worry about any aspect of the training. The management system is also able to take into account the movements of the users during the runtime in order to further adapt the experiences. This enables the system to constantly monitor the user's movements in order to avoid injuries and ensure high quality training.

The system can be used in the context of virtual reality systems, but also augmented reality systems and all other systems that are able to enable a user to perform virtual training processes in a simulation environment.

• 1: eine Schnittstelle zur Definition des gewünschten Muskelaufwands und des resultierenden Teils der Nutzerdatendatei, einschließlich der Freiheitsgrade für bestimmte Gelenke, die in einer .json-Datei gespeichert sind;
• 2: Trainingsdaten, die aus Bewegungserfassungsergebnissen extrahiert und von externen Systemen wie Apple Watch und Fitbit gestreamt werden;
• 3: Schlüssel der Nutzerdaten und Trainingsdaten werden hinzugefügt, um die Trainingsspezifikation zu generieren;
• 4: Rückkopplungsschleife
• 5: Datenformat A beeinflusst die virtuellen Erlebnisse B, welche die Bewegungen C des Nutzers beeinflusst
• 6: Übungsmoduldatenformat;
• 7: Bewegungsdaten im .json Datenformat, die eine Bewegung mit Zeitstempeln und 3D-Positionen für eine bestimmte Verbindung beschreiben;
• 8: Wirkungsdaten in einer .json-Datei mit muskulärer Nutzung und Gelenkbereichen;
• 9: Übungsmodul-Skript, das es dem Übungsmodul ermöglicht, die Erlebnisse aktiv zu manipulieren;
• 10: Generierung von Übungsmodulen aus bestehenden Trainingsabläufe in einer Simulationsumgebung, (a) ausgeführter Code innerhalb der Erlebnisse. (b) aufgezeichnete Bewegungen während des Codes werden ausgeführt;
• 11: Der Nutzer folgt dem gerenderten dreidimensionalen Pfad;
• 12: Der Nutzer imitiert die virtuelle Person (Spiegel).
• 13: (A) Nutzer im Stehen, die Kugel nähert sich ihm. (B) Der Nutzer geht in die Hocke, um zu verhindern, dass die Kugel ihn trifft;
• 14: (A) Das zu boxende Objekt wird dem Nutzer angezeigt. (B) Der Nutzer schlägt das Objekt;
• 15: Nutzer erreicht Objekt. Das erzeugte Hindernis schränkt die Bewegungsfreiheit ein (d.h. er würde gegen das Hindernis stoßen, wenn er zum Objekt geht);
• 16: Spezifizierter Pfad und Nutzerpfad im dreidimensionalen Raum;
• 17: Funktion f, die den Abstand zwischen Zeitstempeln der angegebenen und der Nutzerbewegung darstellt. Integral der Funktion ist Metrik, wie gut der angegebene Motor ausgeführt wurde;
• 18: Verwendung der summierten Differenz der Abstände als Metrik;
• 19: Systemdiagramm;
• 20: Das Erlebnis besteht aus 3D-Modellen, Umgebungs- und Simulationslogik und Skripten. Das Übungsmodul-Skript kann auf alle davon zugreifen und so das Erlebnis generieren und manipulieren.
• 22: Beispiel für die Nutzung der Trainingsfunktion
• 24: Flussdiagramm, das die Ausführung des Übungsmodulplans beschreibt, speziell für drei Anlagenzustände (Start, Betrieb, Ende).
• 26: Flussdiagramm, das beschreibt, wie man Trainingsabläufe in einer Simulationsumgebung mit Hilfe einer Liste zur Handhabung von Übungsmodulen anpasst.
• 27: Zustandsmaschine, die eine Erlebnis beschreibt und Übungsmodule beinhaltet.
• 28: Flussdiagramm, das beschreibt, wie man Trainingsabläufe in einer Simulationsumgebung mit einem Zustandsautomaten für den Umgang mit Übungsmodulen anpasst.
• 29: (a) Speichern der Bewegung unter Verwendung von Winkeln und Längen der Glieder anstelle von dreidimensionalen Positionen, (b) Translation, (c) Rotation, (d) Skalierung, entsprechend den Spezifikationen des Nutzers, um die gesamte Bewegung anzupassen
• 30: Automatische Erstellung von Formeln
• 31: Detaillierungsgrad für den Übungsmodul-Effekt
• 32: Erstellen eines 3D-Körperscan / Körpermodells unter Verwendung der VR- und Motion Capture-Technologie.
• 33: Speichern der Bewegungen so, dass sie individuell angepasst werden können, basierend auf den Körpereigenschaften des Nutzers.
• 34: Trainingsabläufe in einer Simulationsumgebung können vordefiniert werden, indem ein oder mehrere Übungsmodule verwendet werden.
• 35: Verschiedene Bewegungsarten.
• 36: Meta Daten für Übungsmodule.
• 37: Metadaten für Übungsmodule.
• 38: Übungsarten
• 39: Personalisierung durch Parameter
• 40: Simulationsabstraktion
• 41: Hardware-Abstraktion
• 42: Hardware-Abstraktion
• 43: Vergleich von Bewegungen und Extraktion von Metadaten
• 44: Kommentierte Bewegung.
• 45: Code zur Abfrage von Bewegungsdaten.
• 46: Abfrage von Nutzer und Bewegungsdaten, um einen Würfel zu erzeugen.
• 47: Übungsmodul-Repository mit zwei Abstraktionsebenen.
• 48: Korrektheitsscore für verschiedene Übungsmodulimplementierungen
• 49: Übungsmodul Editor
• 50: Simulationstestumgebung.
• 51: Workout Customization Engine,
• 52: Workout Customization Management System (System zur Verwaltung der Trainingsanpassung)

Further measures improving the invention are shown in more detail below together with the description of several preferred exemplary embodiments of the invention with reference to the figures. It shows:

• 1 : an interface for defining the desired muscle effort and the resulting part of the user data file, including the degrees of freedom for certain joints, which are stored in a .json file;
• 2 : Exercise data extracted from motion detection results and streamed from external systems such as Apple Watch and Fitbit;
• 3 : Keys of the user data and training data are added to generate the training specification;
• 4th : Feedback loop
• 5 : Data format A influences the virtual experiences B, which influences the movements C of the user
• 6 : Exercise module data format;
• 7th : Movement data in .json data format, which describe a movement with time stamps and 3D positions for a specific connection;
• 8th : Impact data in a .json file with muscular use and joint areas;
• 9 : Exercise module script that enables the exercise module to actively manipulate the experience;
• 10 : Generation of exercise modules from existing training processes in a simulation environment, (a) executed code within the experiences. (b) recorded movements during the code are performed;
• 11 : The user follows the rendered three-dimensional path;
• 12 : The user imitates the virtual person (mirror).
• 13 : (A) User standing, the ball approaches him. (B) The user crouches to prevent the bullet from hitting them;
• 14th : (A) The object to be boxed is displayed to the user. (B) the user suggests the object;
• 15th : User reaches object. The obstacle created restricts freedom of movement (ie he would hit the obstacle if he went to the object);
• 16 : Specified path and user path in three-dimensional space;
• 17th : Function f, which represents the distance between timestamps of the specified and the user movement. The integral of the function is the metric of how well the specified engine performed;
• 18th : Use the summed difference of the distances as a metric;
• 19th : System diagram;
• 20th : The experience consists of 3D models, environment and simulation logic and scripts. The practice module script can access all of them and thus generate and manipulate the experience.
• 22nd : Example of using the training function
• 24 : Flow chart that describes the execution of the exercise module plan, especially for three system states (start, operation, end).
• 26th : Flowchart that describes how to adapt training processes in a simulation environment with the help of a list for handling training modules.
• 27 : State machine that describes an experience and contains training modules.
• 28 : Flowchart that describes how to adapt training processes in a simulation environment with a state machine for handling training modules.
• 29 : (a) Store the movement using angles and lengths of the limbs instead of three-dimensional positions, (b) translation, (c) rotation, (d) scaling, according to the user's specifications to accommodate the overall movement
• 30th : Automatic creation of formulas
• 31 : Level of detail for the practice module effect
• 32 : Creating a 3D body scan / body model using VR and motion capture technology.
• 33 : Saving the movements so that they can be customized based on the body characteristics of the user.
• 34 : Training processes in a simulation environment can be predefined by using one or more training modules.
• 35 : Different types of movement.
• 36 : Meta data for exercise modules.
• 37 : Metadata for practice modules.
• 38 : Exercise types
• 39 : Personalization through parameters
• 40 : Simulation abstraction
• 41 : Hardware abstraction
• 42 : Hardware abstraction
• 43 : Comparison of movements and extraction of metadata
• 44 : Annotated movement.
• 45 : Code for querying movement data.
• 46 : Query of user and movement data to create a cube.
• 47 : Exercise module repository with two levels of abstraction.
• 48 : Correctness score for different practice module implementations
• 49 : Editor exercise module
• 50 : Simulation test environment.
• 51 : Workout Customization Engine,
• 52 : Workout Customization Management System (system for managing training customization)

The proposed system is able to adapt virtual training processes in a simulation environment so that there is a need for relevant data telling the system how to adapt the experiences. This data includes user data and training data, which are combined to form a training specification.

User data contain all training and rehabilitation-relevant data such as body size, weight, desired training results or rehabilitation progress, etc. This data could for example be entered and specified by users, experts, doctors, etc. via special user interfaces. Alternatively or cumulatively, the system can at least partially record this data automatically. Data from training and rehabilitation platforms such as Google Fit and Apple Health can also be imported and used. In order to support the user in communicating the relevant data to the system, a survey can also be used that translates the answers, for example in the form of selected answers from a multiple-choice format, into training goals. An activity database can also store predefined training goals for specific activities such as swimming. A user can enter that he would like to train for the activity swimming, the predefined data is loaded into the interface and can be further adapted by the user, for example.

This data can, as in 1 can be saved in a .json file, but can also be saved in other ways such as .xml, database etc. Also shows 1 an interface to Definition of the desired muscle effort and the resulting part of the user data file, including the degrees of freedom for certain joints, which are stored in the .json file.

2 shows exercise data extracted from motion detection results and streamed from external systems such as Apple Watch and Fitbit. When users make movements during exercise, these can be recorded by a movement detection system such as the HTC Lighthouse, resulting in recorded movement data in a data format such as .fbx, which represents the movement of a human body. Motion analysis algorithms can be used to extract the effects of exercise such as calories burned, torque applied to specific joints, muscular usage, etc. One way to do this is to compare the body models of the associated timestamps, calculate the speed of each joint and extract the acceleration to estimate the forces applied based on the estimated weights of specific body parts. A physics engine such as Cannon.js or Unity 3D (integrated physics engine) etc. could also be used to calculate these properties. These extracted data (training data) enable the user movements and training to be evaluated. Users can receive suggestions on how to improve their performance, and the system can use this data to further adjust the training, e.g. to avoid injuries, ensure that enough calories are burned or carry out special rehabilitation exercises. Systems like Microsoft Kinect and Apple Watch could act as sensors to collect data, including heart rate. Data from these sensors can be streamed to the system according to the invention in order to generate more training data. Training data is saved in a .json file, but can also be saved in other ways such as .xml, database etc.

The training specification describes the training that must be generated in order to meet the needs of the users. The training specification is saved in a .json file, but it can also be saved in other ways like .xml, database etc. The training specification is created by combining the user data and the training data. Therefore, the amount of calories already burned lowers the amount of calories that have yet to be burned, for example. This enables the proposed system to improve quality and adaptation during training. As in 3 As shown, the generation of the training specification can be achieved, for example, by adding the keys for user data and training data to a new .json file.

As in 4th shown to be a "feedback loop" and it is possible through the combination of a system that is able to generate individual training courses in a simulation environment in a virtual environment, and a system that is able to the user and his Collect data.

In order to adapt the experience based on the training specification, the proposed system must know how to create or manipulate the experience and be able to estimate how the influence will affect the training, i.e. the system must figure out what to do do is based on the way in which the experience, i.e. the training, and thus the user, remains unaffected by any action. To make this possible, the concept of the "exercise module" is introduced. An exercise module describes an exercise and conveys training and rehabilitation-specific knowledge that relates to the exercise. Exercise modules can be used to create custom training programs.

5 shows the influence of an exercise module. The data format is marked with A and influences the virtual experience B, which influences the movements C of the user, here with an oculus crack.

Practice modules can be saved in various ways, such as .xml, folder structures, database etc. The main objectives when creating the in 6 The training module file format shown by way of example are as follows: completeness; The exercise module format is able to cover all types of exercises through body movements. Reusability; Exercise modules are reusable, ie a generated exercise module is not generated for a single experience, but can possibly be used in all training processes in a simulation environment. Creative freedom; Practice modules are available in a format that enables developers and designers to integrate practice modules into their programs or simulations as they wish. For example, a punch exercise module is not limited to a specific 3D model that encourages the user to box, but the developer / designer can integrate the exercise module into the program or game simulation with creative freedom. Expandability, The exercise module format can be expanded very easily. That means that it is possible also parameterized To cover movements, body weight exercises, exercises with objects such as weights, etc. with the help of exercise modules that can be recorded by tracking systems. The description and the image file can be used in an exercise module store to display the specified exercise and to describe what the exercise module is about.

The specification file describes the exercise, e.g. the movement that the user should perform. This can be done with standard file formats such as .fbx, but also with, for example, a .json file that describes the path in the form of a three-dimensional position of certain joints over time. 3D Bezier curves etc. can also be used to define these movements. 7th shows movement data in .json data format, which describe a movement with time stamps and 3D positions for a specific connection.

The effect file describes the training and rehabilitation-specific knowledge that relates to the exercise (e.g. how the execution of this movement will affect the user's body). Based on this data, the proposed system is able to estimate how well each individual exercise module is suitable for the training and for the user. 8th shows effect data in a .json file with muscular usage and joint areas.

The training module script can be used to integrate the training module into training processes in a simulation environment. It is responsible for encouraging the user to perform the specified exercise. This can take the form of a path renderer for the specified joints, or it can be much more complex, such as an enemy performing a certain action / attack that causes the user to move in order to prevent his or her character / attack from being damaged. the role that the user embodies in the virtual environment is met. It can be integrated by overwriting predefined functions, changing global parameters, suggesting changes to the management system, etc. Programming languages such as C #, C ++, C, Python, Java can be used, but also any other type and format of manipulating variables in a computer environment. During the simulation, a specific exercise module script class can be defined using the C # programming language, which includes a start, run and end function. This function can access global variables and thus actively change the experience and, for example, add new objects, change existing ones etc. Exercise modules that are uploaded to an exercise module store can contain predefined code in order to increase the development speed for other developers. 9 shows an exercise module script that enables the exercise module to actively manipulate the experiences.

Movements can be recorded with motion detection devices, for example in a movement detection laboratory. Additional hardware such as EMG sensors and motion analysis software can be used to determine how the recorded motion affects the human body. Motion analysis software can be used to generate three-dimensional motion. Motion analysis software can then be used to determine the impact of motion on the creation of the exercise module. As an alternative to motion analysis software, motion libraries can also be used. This makes it possible to transform already recorded and created movements into exercise modules. Exercise modules can also be obtained from existing training processes in a simulation environment. Therefore, the executed code and the user movements are tracked in parallel. Based on the way the user moves and what code is being executed, the system can extract which part of the code is responsible for moving the user in a certain way. With this knowledge, the proposed system can automatically create exercise modules using the recorded movement as a specification file, motion analysis software to generate the impact file and the corresponding code within the exercise module script. 10 shows the generation of exercise modules from existing training processes in a simulation environment, with "a" denoting the code executed within the experience and "b" denoting the code executed during the recorded movements. Artificial intelligence (AI), for example, can also be used to generate exercise modules from existing training processes in a simulation environment.

Since the user is not necessarily tied to a machine that can force him to move in a certain way, the question arises of how he can move in a certain way. Methods to achieve this include the following. The path for each specific joint such as the hand can be rendered and shown to the user in the form of a three-dimensional path within the training process in a simulation environment, as in FIG 11 shown. A virtual person can be shown that fulfills the specified movement and thus shows the user how to move. As in 12 shown to imitate the virtual person (avatar) or the mirror.

Another aspect is that the user has to prevent something within the training process in a simulation environment through passive movement. For example, he prevents the ball from hitting his character or role in the simulation by crouching, as shown in 13 is shown: In it (A) shows the user standing, with the ball approaching him. (B) The user crouches to prevent the bullet from hitting them. Alternatively, the user could actively interact with the simulation, e.g. striking an object like an enemy, as in 14th shown. 14th shows (A) the object to be boxed is displayed to the user. (B) The user hits the object.

Alternatively, the user could also be tied to a device such as a treadmill, a bicycle or the like.

One way to prevent the user from completing the task in unexpected ways is to eliminate the freedom of movement. For example, a generated obstacle can restrict the freedom of his movements and thus ensure that the user moves as expected. 15th shows a situation in which a user is reaching for an object. The obstacle created restricts the freedom of movement, ie he would hit the obstacle if he went to the object.

It is necessary to compare human movements in terms of experience, such as recognizing whether the user has fulfilled the specified movement path or not, but also for evaluation purposes, e.g. to measure the accuracy of user movements. 16 shows such a predefined movement path and the user movement in three-dimensional space. One possibility for this is to compare the position of each time stamp in three-dimensional space (specified path and user path). A function f is calculated that maps the distance between these two points for each time stamp. The integral of this function can be used to obtain a metric of how similar these two movements are. However, it is known which timestamp is the beginning of the specified path, but it is possible that it is not known where the user started to move. To solve this problem, the user's movement has to be shifted by a time stamp for several times and the metric has to be recalculated. The total metric is then the minimum value of these metrics. One way to know how often to postpone user movements is to determine when the user position was closest to the endpoint of the specified movement, and then postpone user movement one time slot per shift until the movement was at that particular point in time begins. Since the specified and the user movement can consist of several joints and thus several paths, this must be done for each path. The metric of each path comparison is added to a value that indicates how similar the user movement and the indicated movement are. 17th shows an example of a function f, which represents the distance between time stamps of the specified movement and the user movement. The integral of the function is metric and shows how well the given movement was performed.

With this method, however, movements over, for example, 1000 time windows are generally better than movements over 10000 time windows, since the errors in each time window add up. To prevent this, the resulting value can be divided by the number of time windows. The use of the summed difference of the distances as a metric is shown in 18th shown. Instead of measuring the similarity between movements in a “semi-continuous” (not continuous because the space between the time windows is interpolated), the distances can also be added. Therefore no function f or the calculation of an integral is necessary. Another way to compare movements is to compare the effects instead of the paths themselves. This can easily be done by calculating the difference for each key in the impact files of the user movement and the given movement. This method makes it possible to compare the movements in connection with the effect, but does not provide any information about the quality, the user has fulfilled the specified movement.

19th shows a system diagram for generating the virtual experiences. To do this, the proposed system must find out what the experience must look like in order to meet the training specification. As in 20th shown, the experience consists of 3D models, environment and simulation logic and scripts. The practice module script can access all of them and thus generate and manipulate the experience. Predefined elements in the proposed system are: 3D models, environment and predefined scripts with functions that can change and manipulate the experience and implement the simulation logic (prefabs). Each tutorial contains a tutorial script. Each function in the lab script (start (), running (), end ()) can be used to refer to predefined functions, 3D models and the Environment. Functions that are specialized in each exercise module can also be defined such as time delays. The exercise module script can also access all exercise module-relevant information such as the specification and the effect file, etc.

A training module can, for example, access the global prefab functionality to create a line renderer and thus render the movement specified in the specification file. A global generateCubeAtPos (3D Position) script can be used to create a cube at a specific position. By appending the global moveCubeWithConstVelocity (Velocity) script to the cube, the cube moves towards the user at a certain speed. By attaching a detectUserPunchAnDestroy () script to the cube, the simulation could also recognize whether the user boxed and destroyed the cube and thus ended the exercise module. This format enables the practice module to generate a virtual experience. With a system that is able to determine which exercise module should be carried out in which order, individual training processes can be carried out in a simulation environment for training and rehabilitation purposes or similar purposes, such as training situations for aerospace, police or military, etc. are generated. The proposed invention supports several options for generating and setting up virtual training processes in a simulation environment with this concept. This includes, for example, the use of a list structure and a state machine. These concepts could e.g. be used by simulation developers.

This concept can use a list of practice modules to generate the virtual experience. The user data and the training data are combined so that a training specification is created. The training specification is in the same data format as the specification of the exercise module effect.

An optimization algorithm now generates a combination of training modules, ie a number for each training module that determines how often the training module is performed, based on the training modules available in the experience and training specification. The available exercise modules can be saved directly in a folder, within the training processes in a simulation environment or references can be saved and the simulation will download the exercise modules accordingly. Some exercise modules can also be saved directly and a link can be saved for others. The task is to determine an exercise module combination that fulfills the training specification as well as possible, ie which exercise modules have to be performed how often so that the exercise module effects taken together are as close as possible to the training specification. To solve this problem, the system projects this problem onto a multi-dimensional packaging problem. This problem is then solved with an evolutionary algorithm. Formula 1 describes a training function that is used for the evolutionary algorithm. 22nd shows an example of using the training function. $$f={\displaystyle \sum _{j=1}^m\left|\left({\displaystyle \sum _{k=1}^n{a}_{kj}\ast x\left(a_k\right)}\right)-w_j\right|}$$

But other optimization strategies could also be used to solve this problem. The algorithms, such as the evolutionary algorithm, can also be optimized by changing the threshold values, etc.

After the system has determined a combination of training modules, the sequence of the training modules must be planned. It is possible to complete the exercise modules in random order. Another possibility is to optimize the chronological order according to the relevant data such as a number that describes how bad it is to do certain exercise modules one after the other (penalty points). Table 1 shows the penalty points for changing from exercise module n to exercise module m. Table 1 from ... to Exercise module 1 Exercise module 2 Exercise module 3 Exercise module 1 -1 -1 -3 Exercise module 2 -4 -1 -2 Exercise module 3 -2 -3 -1

These numbers can be stored in a global table, .json, .xml etc. or within the experiences. These numbers can be a global measure, but they can also be optimized for specific training processes in a simulation environment. In the case of a schedule, including the established exercise module, it is possible to determine the quality of the plan by adding up the penalty points for each exercise module and the following exercise modules. The task is to create a chronological sequence of exercise modules with the lowest possible value, ie with as few penalty points as possible.

In addition to the general simulation logic, which, for example, implements the management of points, simulation effects, etc., the generated exercise module plan must be executed. So the system performed one training module at a time, each going through all 3 steps (start, run, end). The specified code is executed accordingly and thus generates the training process in a simulation environment, ie the experience. 24 Figure 12 shows a flow chart describing the execution of the exercise module schedule, specifically for three exercise module states (start, run, end).

The entire simulation logic can change certain global variables such as runAsset, pauseAsset and endAsset in order to control the execution of the system plan. This is necessary, for example, if the simulation implements predefined simulation parts that should not be manipulated by exercise modules. According to one embodiment, external persons such as doctors or training experts could also manipulate the execution of exercise modules on-the-fly with the aid of certain parts within the simulation logic. Variables such as endAsset can, however, be set to true or false by the training module itself, but also by other entities. A movement comparison that runs within the simulation logic can, for example, end the exercise module as soon as the user has performed the specified movement. A mechanism for recognizing whether the user has hit the target could also end an exercise module that encodes a box movement. The exercise module execution can be executed in a single thread and thus run parallel to the game simulation logic, but can also be executed as an instance within the simulation logic. Table 2 shows a list of an exercise module schedule that will be executed.

The process generated an exercise module plan (asset plan) that depicts the training specifications as well as possible. The exercise modules are then carried out and thus generate the virtual environment for the training experience. Since the training specification is continuously updated as training data is generated, the training module order is also continuously updated. This leads to tailor-made training units that are also adapted and optimized during training (during the runtime). 26th shows a flowchart that describes how training processes are adapted in a simulation environment with the aid of a list for handling training modules.

The concept of a state machine makes it possible to use states and interactions to generate training processes in a simulation environment. In contrast to the list, the simulation can now be mapped to different states and thus form a much more complex simulation. 27 shows a state machine that describes an experience and includes training modules. Each node has some pre-release versions with pre-defined code, such as scene changes when entering the state, etc. and a list of training modules available for the respective state. Therefore, the available practice modules can be limited depending on the state. Even if this seems to limit the complexity of the simulation, the restriction allows different training module environments and thus a variety of different contexts to be created, which makes it possible to add complexity to the simulation. In order to manage the training modules, when entering a new condition, the same system is used as described above to create a training module plan for the respective condition, depending on the training specification and the available training modules for that condition. The state logic can implemented within the simulation logic and controls the creation and execution of the exercise module plan by limiting the available exercise modules based on the current state of the experience. The experience is adjusted for each condition, and thus the entire experience is adjusted.

In contrast to the use of a list, however, the state machine does not allow global optimization, since the user interaction influences the state of the experience and thus the optimization process. That is, the quality of the adjustment is low if the state of the experience changes frequently, since the adjustment only takes place within the framework of states.

In all embodiments, the dynamic adaptation and change of the training process is given in the simulation environment, i.e. in contrast to known video game exercises, predefined game scenes are not called up and played, but individual new scenes and sequences are created and changed on the fly based on the user data.

Methods such as state prediction can be used for this. This means that the system estimates how long the user will remain in the current state and calculates a probability for the next n states. This can be done on the basis of stored data of the user himself and / or other users and helps the proposed system to adapt the experiences on a global level by calculating several exercise module schedules, including possible future states, and choosing the schedule with the smallest expected difference decides on the training specification. 28 shows a flow diagram that describes how to adapt training processes in a simulation environment with a state machine for dealing with training modules.

In connection with the generation of exercise modules, a method was described above that makes it possible to generate exercise modules from existing training processes in a simulation environment by observing the movements of the users and the executed code in parallel. In addition to generating new training modules, this method enables the control of existing training processes in a simulation environment that were not specifically designed or used for the proposed system. This can be done by decompiling the experiences, inserting the generated exercise modules in place of the code that was used to generate the exercise modules, and compiling the experience. Therefore, the entire structure and logic of the experiences remain the same, but fixed parts that were used to generate the exercise modules can be manipulated by the proposed system in order to adapt the experiences.

Since each training process in the simulation environment or each experience is defined by the exercise modules, which flow into the training processes in the simulation environment and each exercise module is defined by a specific effect or a specific effect for the user, training processes in a simulation environment can be based on their average Effect can be classified, ie some training processes in a simulation environment can be particularly good for leg training or building endurance, etc. By comparing the training specification and the average effects of certain training processes in a simulation environment, a score can be calculated that describes how the experience fits into the training and rehabilitation goals of the user. Therefore, specific training sequences can be recommended to the user in a simulation environment that are well suited to the training and rehabilitation requirements and goals of the user.

This path can already be a good reference for the movement that the user should perform, but each user will have different body characteristics such as height and arm length. Therefore, the pure motion can be abstracted to motion of joints and connections, which can be stored in the form of angles and relationships between joints and connections instead of 3D positions. Then these movements can be translated and scaled based on the user's height and arm length to adjust their body parameters. This method can be specified with a complete 3D body scan of the user. This approach allows the movements to be defined very precisely. But for most users the original movements will be sufficient. Therefore, a body scan, etc. is not necessary for all users, even if it is a good method for increasing precision. 29 shows a corresponding procedure for the acquisition: (a) storing the movement using angles and lengths of the limbs instead of three-dimensional positions, (b) translation, (c) rotation, (d) scaling, according to the user's specifications, around the adjust entire movement. This method makes it possible to adapt every recorded movement to the body specifications of each individual user.

The specification file can also define body weight exercises, exercises with external objects or even exercises with several people.

For some trainings, such as those involving external objects, some requirements must be met, such as the presence of the required object. For example, when doing exercises with weights, the weights must be present. The SLAM method, for example, can be used to determine which objects are available. The requirement can, for example, be specified in a .json file and assign a value that is significant for certain IDs. These IDs relate to the objects that are required. The value 10 could mean that the object is absolutely necessary for the exercise, if the object is not available, the exercise cannot be carried out, so the exercise module cannot be used and integrated into the experience. The ID can also refer to a specific class of objects such as a weight, regardless of whether a 5kg weight or a 10kg weight is available.

Since the execution of an exercise always depends on external parameters such as body height, weight, arm length of the user for movements and body weight training and the weight of external objects for exercises with e.g. weights, there is a need for internal parameters. The parameter file can e.g. be a .json file and describes the parameters that are important for the systems. Based on these parameters, exact specifications for the system can be calculated. When managing the system structure, these internal parameters can be overwritten with the actual values presented by the user and the environment. The weight parameter of an exercise module can be overwritten by the actual weight of the external object that is present in the user environment.

The effect can contain fixed numbers, but also formulas that refer to the parameter file. For example, the use of the left bicep can be represented by a formula that references the weight parameter that is overwritten by the actual weight of the current external object.

The effect of exercise modules can be measured for several users. Together with their parameters, the formulas that describe the relationship between parameters and effect can be calculated and saved in the exercise module. 30th shows the automatic creation of formulas for multiple users based on the evaluation of the measured effect and the parameters.

31 shows a level of detail for the practice module effect. The practice module effect can be saved in different levels of detail. For example, it could describe an exercise module value in such a way that it is good or bad for certain purposes, describe fixed effects or effects that depend on parameters such as the body weight of the user.

For example, a three-dimensional body scan can be created using specific postures and simulation elements to filter parameters such as arm length. 32 shows the creation of a 3D body scan / body model using VR and motion capture technology. This data can be saved to improve the customization process, but it can also be applied to other things outside of the Fitspace area, e.g. it can be used by fashion companies to customize clothes. 33 shows the storage of the movements so that they can be customized based on the body characteristics of the user.

Practice modules can also contain metadata, such as a label, which they are good for, as in 36 shown. In this way, a user can determine which type of experience he prefers, such as adventure or action. The WCE can take these designations into account by calculating a penalty point on the basis of the calculated layout plan. This can be done by comparing the number of exercise modules that include the label-like action with those who do not rate how much of an action experience the experience generated is.

An alternative to the training specification is an effects system. This means that a management system has created an effect request and forwards it to the Workout Customization Engine (WCE). The request depends on the motion analysis, user data, training data, etc., as in 37 shown schematically.

As in 38 shown, there are many types of exercise module such as one to improve speed, response, burn calories through energy expenditure, improve muscle activity, etc.

In addition to using exercise modules to customize the experience, parameters can also be used. Based on the calories that a user would like to burn, for example simulation parameters such as the strength of his attacks within the simulation can be adjusted. Burning a large amount of calories would decrease strength, forcing the user to be more active in order to gain within the simulation. However, this method does not allow active, user-specific training processes to be generated in a simulation environment. The control over the experience and the training is much less than when using exercise modules. Such an adjustment based on parameters, without the need for training modules, is described in 39 shown.

Since the training processes in a simulation environment are training processes in a virtual environment, it is possible to actively manipulate what the users see. Hence, a concept called simulation abstraction is possible. Simulation abstraction means, as in 40 demonstrated that multiple users can play together even if they play different simulations. One user could play a zombie simulation with another user who is actually playing a nature simulation. Both users think that they are playing their specific simulation together. This is possible because the underlying interaction concept is the same and a management system can control the simulation. For example, the user who plays a zombie simulation could attack a zombie with a sword, the user who plays the nature simulation sees exactly what the other user is doing, but after his experience he uses a stick to destroy a trap . The users can be connected via an internet connection. This method is not limited to two users. Several users who play several training courses in a simulation environment can participate. The WCE has to optimize the experience globally, based on all training processes in the simulation environment and all users and the training specifications of the users, since all training processes in a simulation environment and users influence each other.

Similar to simulation abstraction, there is also the option of abstracting different types of hardware, as in 41 shown. For example, two users can interact within the same experience, one using an Oculus Rift at home and the other using a BlackBoxVR device in a training studio. Since these different types of hardware restrict and support user interactions in different ways, the management system must be provided by a pool of possible interactions for each hardware system in order to create a common experience in a virtual environment for several users. The method of simulation abstraction and hardware abstraction can be used at the same time. This means that multiple users can play multiple simulations with different hardware systems, but everyone thinks they will have the same experience with no hardware restrictions. These abstraction methods could be used to create much larger training flows in a simulation environment, tournaments, etc., since the hardware and experience have no bearing on who can participate.

As in 42 shown, there may be an adapter repository. Each adapter can be placed between the hardware and the software to abstract the hardware.

Since the movement of the user is evaluated and saved and the system is informed about the training and rehabilitation conditions of the user, this data can be saved in a database. 43 shows a comparison of movements and extraction of metadata. User movement is then compared with stored movements; specific properties such as small changes in direction of movement can, under certain conditions, indicate an injury that can be suggested to the user. At the end of this method, the aim is to compare the user's movement with stored movements, similar movements could indicate similar training and rehabilitation conditions, so the proposed system could recognize these conditions and help the user to find out if something is going wrong, e.g. an injury.

The hardware abstraction can also be supported by several practice module scripts, which means that there is an asset script for every type of hardware, such as Assetscript_HighEnd, Assetscript_Virzoom etc. Each exercise module contains a specification file, which can also be a standard motion file such as .fbx. Since this movement describes a part of a training session, it can, as in 44 also give comments, e.g. a START, END-Movement and END-ASSET annotation, which indicate when the movement begins, when the pure movement is carried out like a stroke and when the entire movement ends.

Data relevant to movement can be accessed using code. For example, the path of the left hand can be queried and then rendered. The queried data can include data such as movement paths, but also contain data such as total distances of certain connections etc. 45 shows a code for querying movement data. This enables these movements to be linked directly to the simulation event. Therefore, the position of the left hand at the moment the movement is performed could be queried in order to generate a cube at that position, as in FIG 46 shown so that the user can box the cube and make the move. Since these movements can be adapted using the parametric movement method already described in this document, customer-specific data can also be used to influence the simulation environment.

The lab module repository is divided into two levels of abstraction, as in 47 shown. The first level of abstraction represents the different training modules. The second is the different implementations for each training module. Developers can use level 1 to search for movements and effects and level 2 to search for simulation events and concrete implementations.

Each implementation may contain specific additional files such as 3D models etc. relevant to the execution of the specific training module implementation. To ensure the correctness of the various system implementations, a correctness or precision value could indicate how accurately the effect describes the object. This is exemplified in 48 shown. This value can be calculated by comparing the effect with the real effects measured for several users.

A tutorial editor, as in 49 shown, allows developers to render the motion specified in the tutorial and create the tutorial script using visual feedback. The tutorial script can also use global parameters such as general simulation speed, etc. to customize the experience. As a result, training sequences in a simulation environment can be adapted to the age of a user, for example, if fast movements are generally bad for the user's body. User interactions, such as the selection of a certain weapon in a simulation, etc. can also be used to influence the simulation and the training module plan. This can be implemented in the simulation logic or in certain exercise modules themselves.

A test environment that allows developers to test their simulations is in 50 shown. The simulation logic and the exercise modules are used to represent the user and the simulation environment. The developer can evaluate the quality of the simulation.

In contrast to the concept that the Workout Customization Engine (WCE) selects exercise modules and these exercise modules directly manipulate the simulation or the experience, exercise modules can also be provided in a Workout Customization Management System (WCMS). In this scenario, the practice modules propose operations and events to the management system that directly manipulate the simulation or experience.

• • Trainingsspezifikation
• • verfügbare Übungsmodule
• • weitere Daten wie Tabelle, die Strafpunktzahl für aufeinanderfolgende Übungsmodule m und n speichert (Tabelle 1)

51 shows the working principle of the Workout Customization Engine (WCE). The WCE receives the following data:

• • Training specification
• • available exercise modules
• • Further data such as a table that stores the penalty points for successive exercise modules m and n (Table 1)

The order of the training modules is determined from this data. The virtual environment results from the execution of the simulation logic and the exercise modules according to the sequence of the exercise modules. The exercise modules manipulate the virtual environment directly, e.g. by creating a 3D object. If the execution of an exercise module is to be stopped, e.g. because an issue is to be explained to the user during the simulation, the simulation logic must either intervene directly in the exercise module or in the execution of the exercise module. The simulation logic and the execution of the exercise module are therefore two different processes running in parallel that can influence each other. One could also speak of a “decentralized” solution because there is no higher-level system and both processes can manipulate and adapt the virtual environment directly.

Another solution is the Workout Customization Management System (WCMS), which is a higher-level management system. The working principle of the WCE with WCMS is in 52 shown. When executing the exercise module, the exercise modules no longer manipulate the virtual environment directly, but provide the WCMS with their functionality, i.e. make a suggestion as to what is next in the virtual environment Environment could happen. The WCMS then has the option of executing or allowing the functionality of the exercise module, pause, end the exercise module, etc. This can then be done on the basis of the simulation logic. For example, the training module that is to be carried out next could have a priority of 4, but the action that the simulation logic wants to carry out, but a priority of 10, e.g. explain something to the user during the simulation. The WCMS then decides on the basis of the priorities to first execute the action of the simulation logic, to pause the exercise module during this and to only continue to execute it after the action of the simulation logic has ended. The WCMS is therefore not an alternative to the WCE, but enables an alternative structure of the overall system. The embodiment of the invention is not limited to the preferred embodiment specified above

QUOTES INCLUDED IN THE DESCRIPTION

This list of the documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• US 2019046859 A1 [0006]
• WO 2004109575 A1 [0007]
• WO 2007058512 A1 [0008]
• KR 20020030755 A [0009]

Claims (16)

Method for generating individual training sequences in a simulation environment, in particular for sports and rehabilitation purposes, the method comprising the following steps: - Acquisition of user data via user input and sensors; - Analysis of user data and determination of calculated body and / or movement data; - Selection of training data corresponding to the calculated body and movement data from a first data memory, the physiological effect on the user being a selection criterion; and - Generating a training specification; - Provision of exercise modules (exercise module) in a second data memory; - Combination and arrangement of the selected exercise modules according to the requirements of the training specification; - Generation of a virtual environment, environment visualization and 3D modeling from a simulation logic selected by the user; - start training; the method further comprising the steps of: a) Acquisition of user information, such as body data and movement parameters, during training; b) ACTUAL / TARGET comparison of the recorded user information with the calculated data; c) Adapting the training specification to the TARGET data; d) Adapting the combination and arrangement of the selected exercise modules according to the specifications of the adapted training specification; e) adapting the virtual environment to the changed training modules; f) Executing the adapted combination and arrangement of the selected exercise modules and jumping to step a). Method for creating individual training processes in a simulation environment Claim 1 , optimizing the time sequence of the exercise module according to the relevant data such as a number that describes how bad it is in order to carry out certain exercise modules one after the other (penalty points). Method for creating individual training processes in a simulation environment Claim 1 - 2 , the effect of the training being extracted with a motion analysis algorithm and stored in the first data memory. Method for creating individual training processes in a simulation environment Claim 1 - 3 , comparing the body models of the associated timestamps, calculating the speed of each joint and extracting the acceleration therefrom and calculating the applied forces based on the weights of certain body parts. Method for creating individual training processes in a simulation environment Claim 1 - 4th , whereby an exercise module describes an exercise model and contains training- and rehabilitation-specific knowledge that relates to the exercise. Method for creating individual training processes in a simulation environment Claim 1 - 5 , wherein the method uses an environmental visualization that shows the stored training unit by visualizing the movement path or by visualizing a human model performing the training unit. Method for creating individual training processes in a simulation environment Claim 1 - 6 wherein the user or the role that the user embodies in the virtual environment is prompted to correctly execute the training modules by means of the virtual environment and objects. Method for creating individual training processes in a simulation environment Claim 1 - 7th , whereby devices for recognizing whether the user has fulfilled the specified trajectory or not are provided, whereby for this purpose the comparison of the joint positions of each time stamp is recorded in three-dimensional space and functions f are calculated therefrom which map the distance between these points for each time stamp and extract an error value from it. Method for creating individual training processes in a simulation environment Claim 1 - 8th , with an optimization algorithm generating a combination and / or chronological sequence of training modules. Method for creating individual training processes in a simulation environment Claim 1 - 9 , a state machine being provided in order to generate training sequences in a dynamic simulation environment. Method for creating individual training processes in a simulation environment Claim 1 - 10 , whereby a simulation abstraction is provided, by means of which several users can interact together, even if they are in different virtual environments and scenarios. System for generating individual training processes in a simulation environment, the system having at least one processor, devices for capturing user data, a first data memory with training data corresponding to user data, a second data memory with exercise modules, virtual reality or augmented reality hardware for output of the dynamic simulation environment, sensors and / or a movement detection device for recording user information, such as body data and movement data during training. System for creating individual training processes in a simulation environment Claim 12 the system further comprising means for connection to a network, preferably to the Internet. System for creating individual training processes in a simulation environment Claim 12 - 13 , adapter repositories are provided, each adapter being placed between the hardware and the software to abstract the hardware. System for creating individual training processes in a simulation environment Claim 12 - 14th , wherein the system is configured to perform the method of any of Claims 1 to 11 executes Computer program, comprising instructions which, when the program is loaded and executed on the system, according to one of the Claims 12 to 15th this causes the steps of the method according to one of the Claims 1 to 11 execute.

Images