Method and system arrangement for the hardware-efficient controlling of a device

Abstract

The present invention is directed to a method which enables a device to be configured particularly efficiently and control commands to be selected particularly efficiently. The present invention allows for the selection options to have to be assigned only once in the two-dimensional storage structure, and thus an efficient method is provided which consumes less processing power and memory. The advantage is that parameters do not have to be repeatedly read out or selected and redundantly stored. As a result, errors are avoided and far less memory is required than in the prior art. The invention is also directed to a correspondingly designed system arrangement, which is designed to carry out the method. A computer program is also proposed, which provides commands suitable for implementing the method and operating the system arrangement.

Description

[0001] Method and system arrangement for hardware-efficient control of a device

[0002] The present invention is directed to a method which makes it possible to configure a device particularly efficiently or to select control commands particularly efficiently. The present invention ensures that the selection options only have to be assigned once in the two-dimensional memory structure, thus creating an efficient method which consumes less processor power and memory. This has the advantage that parameters do not have to be read out or selected repeatedly and do not have to be stored redundantly. This avoids errors and the memory requirement is significantly reduced compared to the prior art. The invention is also directed to a correspondingly configured system arrangement which is configured to carry out the method. Furthermore, a computer program is proposed which contains commands which are suitable for implementing the method orto operate the system arrangement.

[0003] In an increasingly digitalized world where data volumes are growing exponentially, storage efficiency has become a critical factor. Optimal use of storage resources not only plays a role in ensuring smooth operations but also has a significant impact on the environment.

[0004] With the exponential growth of digital data in recent years, the demand for storage space has also increased inexorably. Companies, government agencies, and private individuals generate large amounts of data every day, whether through social media, cloud services, IoT devices, or other digital applications. This enormous data growth presents a challenge, as providing sufficient storage space is not only expensive but also energy-intensive.

[0005] Traditional storage systems are often not optimally designed for efficient resource utilization. This leads to wasted energy and physical space, which in turn leads to higher energy consumption and CO2 emissions. Server farms, data centers, and cloud infrastructures require significant amounts of electricity to operate and cool the equipment. Inefficient storage design unnecessarily increases the energy requirements of these facilities, thus contributing to environmental pollution.

[0006] Improving storage efficiency offers a variety of benefits for businesses and the environment. Reducing the amount of storage required can reduce costs and streamline operations. Efficient storage systems enable faster data processing and retrieval, resulting in improved performance and user experience. At the same time, environmental impact is minimized by using less energy and requiring less physical space.

[0007] Various state-of-the-art measures are known to improve storage efficiency. One possibility is to use innovative storage technologies that enable higher data density per storage space. Examples include solid-state drives (SSDs) and holographic storage. Furthermore, data management is of great importance. Effective data compression, deduplication, and archiving can avoid redundant data and optimize storage space utilization.

[0008] Efficient data structures are an essential component of software development because they form the basis for organizing and accessing data. An efficient data structure is characterized by its ability to perform operations such as inserting, deleting, and searching data quickly and with a small memory footprint. Examples of efficient data structures include lists, arrays, stacks, queues, trees, and hash tables. Each of these data structures offers specific advantages and is suitable for different use cases. By choosing the right data structure, developers can optimize the performance of their programs and significantly reduce execution time. In addition, algorithms based on efficient data structures also play a crucial role in overcoming challenges in processing large amounts of data or in optimizing search and sorting operations.Overall, efficient data structures contribute significantly to improving the performance and scalability of software solutions.

[0009] Selecting an efficient storage structure can be difficult because various factors must be considered. Every application has specific data access requirements. Some applications require fast insertion and deletion, while others must be optimized primarily for read operations. Therefore, choosing the right data structure depends heavily on the application's requirements. The size and type of data to be processed influence the choice of storage structure. Some data structures are better suited for small amounts of data, while others are better suited for large data sets or specific data types.

[0010] The performance and efficiency of a data structure depend on the operations performed. Some operations can be executed faster in certain data structures than in others. Analyzing the runtime complexity and memory requirements of the operations is important for making the most efficient choice.

[0011] Conventional methods require complex control of machines and devices. Complex devices offer a multitude of input functions, each of which is specifically configured. This creates a large number of selection options, some of which must be manually assigned, then further processed, and saved. The problem here is that the machines sometimes deliver incorrect results if incorrectly configured. A single configuration error can render the end result unusable, making subsequent corrections very time-consuming and expensive.

[0012] Another disadvantage of the state of the art is that some of the inputs have to be made by human users and are often overwhelmed by the high number of inputs. In the state of the art, matrices with configuration parameters must be specified in such a way that, for an nxn matrix, n 2 Parameters must be set. Not only is this input time-consuming and error-prone, but the memory requirements and processor load for further processing also increase with the number of input options. In addition, the corresponding parameters typically have to be transmitted via an over-the-air interface, thus requiring a corresponding amount of bandwidth. In mobile networks, however, fluctuations often occur, and especially with real-time requirements, such large numbers of parameters can no longer be adequately transmitted.

[0013] It is therefore an object of the present invention to propose a method or a system arrangement that enables input parameters to be efficiently queried and stored. Furthermore, it is an object to generally create a hardware-efficient device or system arrangement that can execute the method. The method should also be configured to operate the system arrangement. Furthermore, a computer program product is to be proposed that has control commands that execute the method or operate the system arrangement.

[0014] The problem is solved by the features of patent claim 1. Further advantageous embodiments are specified in the subclaims.

[0015] Accordingly, a method for hardware-efficient control of a device is proposed, comprising

[0016] - providing a symmetric, two-dimensional memory structure with pairwise selection options, wherein the selection options are arranged continuously in both dimensions and the selection options are present as either qualitative or quantitative parameters;

[0017] - reading of at least one input sensor for the assignment of

[0018] Parameter values ​​to the parameters, where the parameters each correspond to the pairwise

[0019] Selection options are only assigned once; applying a weighting stored in a persistent data store to the parameter values ​​and normalizing the weighted parameter values; and controlling the device using the weighted and normalized parameter values, whereby groups of parameter values ​​are assigned to control commands.

[0020] The proposed method is hardware-efficient in several ways, as both the execution and storage of the two-dimensional memory structure are efficient. This is because the parameters are assigned only once in pairs. This results in smaller data sets compared to the state of the art, which consume less memory and thus generally fewer hardware resources. Processing the collected data is also more efficient. This overcomes the disadvantage of the state of the art, namely that a two-dimensional memory structure, i.e., a matrix, must always be completely filled with values. However, in this case, pairs of selection options are available, so it is not necessary to query the selection options twice and save the results.

[0021] This allows the device / hardware to be controlled efficiently and the technical effect is already present when reading the input sensor and the underlying system arrangement.

[0022] In a preparatory process step, a symmetrical, two-dimensional storage structure is provided. The storage structure can generally be a matrix or an nxn storage structure. This means that the two dimensions of the storage structure are of equal rank, i.e., equal size. This creates pairs of selection options. Thus, it is possible for the options A, B, and C to be arranged horizontally, i.e., along the x-axis, in the two-dimensional storage structure, and the same options A, B, and C to be arranged vertically. This creates a two-dimensional storage structure with n = 3. The selection options are thus arranged continuously in both dimensions, whereby continuous also implies that they can be arranged inversely. For example, the sequence of options A, B, and C can run horizontally, and C, B, and A vertically.The selection options have parameters that can be either qualitative or quantitative. Qualitative parameters can, for example, specify that something is bad, medium, or good. This creates a value range that is defined according to a predefined rule. In the example mentioned, there can be a uniform distribution and the parameter values ​​can be equidistant. For example, the quality can be divided such that the option "good" is assigned a value of 0-0.33. A value of "medium" can be assigned the value range 0.34 to 0.66. The qualitative parameter "bad" can be assigned the value range 0.67-1. This makes it possible to process qualitative parameters automatically. Quantitative parameters are, for example, numerical values ​​that are specified accordingly.

[0023] To describe the two-dimensional memory structure, at least one input sensor is read to assign parameter values ​​to the parameters. An input is therefore expected that assigns parameter values ​​to the parameters of the two-dimensional memory structure. A parameter can, for example, be a variable that must be assigned according to a data type. The parameter can be a floating-point number and the corresponding value can be, for example, 0.5. For this purpose, at least one input sensor is provided, which writes data once or continuously, which is then assigned to the parameters as parameter values. The input sensor can also be designed virtually, so that a loop is implemented that expects a corresponding input.

[0024] The parameters are only assigned once in the paired selection options. This has the advantage that not the entire data structure needs to be written, but less than half. The total number of data records to be written is therefore n squared half minus n half, i.e. n 2 / 2 - n / 2. This avoids having to describe all the values ​​in the matrix, and compared to conventional input methods, fewer than half the input parameters need to be stored. This creates hardware efficiency in several ways. The savings in memory requirements arise from the fact that the two-dimensional memory structure is symmetrical, and thus the data is written redundantly. For example, the value in the first row, second column is the reciprocal value of the value in the second row, first column. This would essentially result in a mirrored matrix with duplicate values, which is avoided according to the invention. It is also possible for corresponding values ​​to be inverse / reciprocal of each other. For example, the value 5 can be in the first row, second column, and the value 0.2 can be in the second row, first column.However, this can also be calculated in a simple way, so it's not necessary to query and save both parameter values. If information is available that corresponding pairs of values ​​are not identical but inverse, this can be calculated with little technical effort.

[0025] As soon as the parameter values ​​are available, a weighting stored in a persistent data store is applied to the parameter values, and the weighted parameter values ​​are then normalized. This means that a weighting is stored in advance in a data store, which indicates which parameters are how important for the final result. For example, safety-critical parameters can be defined and then given a higher weighting. The weighting can be saved in a preparatory process step and can be provided via an air interface or stored locally. In order to process the corresponding values, normalization occurs so that all values ​​are kept within a specific value range. If, for example, an input of 0-1 is possible, the values ​​cannot be normalized or are multiplied by a normalization factor of 1.If a definition range or value range of 0-10 applies to another parameter, the corresponding parameter can be normalized such that the parameter value is also mapped to a range of 0-1. For example, if a value of seven is provided in this process step, it can be normalized accordingly to the value 0.7. This normalization step creates parameter values ​​that can then be compared with one another. This also makes it possible to process qualitative assessments in such a way that they can be processed according to the remaining parameter values. The end result could, for example, be individual factors that are multiplied together to produce a single parameter value that is created depending on the multiple parameter values ​​provided. However, the individual normalized parameter values ​​can also be taken into account in subsequent processing steps.

[0026] The machine or device, which can also be an actuator, is then controlled using the weighted and standardized parameter values. The required parameter values ​​are now available, and based on these parameter values, a device configuration can be created. Groups of parameter values ​​are assigned to control commands. For example, it can be detected that a vehicle component is showing a certain level of wear. A decision must therefore be made as to whether or not this component should be replaced. In this exemplary embodiment, it must also be determined which component should now be used. The component is examined using at least one input sensor, and the read-out parameter values ​​are entered into the two-dimensional memory structure. This provides an inventory, and the current state of the component / system is described in the memory structure.

[0027] It is also possible that, for example, there are two sensors that each provide the same value (e.g. CD and DC) (=efficient) - but require a new calculation if one of the values ​​C or D does not deviate reciprocally.

[0028] One example is controlling an actuator that, for example, controls an aircraft's rudder via the output parameter. In such a case, it may be possible to perform multiple sensor queries and, if necessary, even query the entire matrix.

[0029] Using predefined groups of parameter values, a control command can now be selected which, for example, states that the corresponding component must be removed and another component inserted. Now that various parameters have been assigned, the target requirements become clear, and the control command can therefore also state that the component should be removed and replaced with another component of type X. The parameters can also refer to multiple components, and multiple control commands are issued. This can mean that a first component can remain in a vehicle, while a second component must be replaced. In this way, a plurality of parameter values ​​is extracted using multiple or just a single input sensor. This then helps in deciding how a device should be controlled and what exactly it should do.For example, the proposed device could be a robotic arm that executes commands from input sensors. Several sets of parameter values ​​are specified, which lead to one or more control commands.

[0030] This allows multiple parameter values ​​to be stored, and a control command can be assigned to this group of parameter values, which can optionally be available in multiple versions. A control command can specify that a component must be replaced with a first component, or it can specify that this component must be replaced with a second component. Thus, a complex decision is made automatically that controls the device, and the appropriate control commands are identified and then executed based on the recorded parameter values.

[0031] According to one aspect of the present invention, the input sensor provides a maximum upper limit of inputs which corresponds to half the controlled number of selection options minus half the number of selection options. This has the advantage of increasing hardware efficiency and, in particular, the feature allows the parameters of the paired selection options to be assigned only once, eliminating the need to describe the entire matrix. Since a diagonal is formed in the two-dimensional memory structure in that the individual options do not need to be specified relative to one another, i.e., identical options such as option A and option A do not need to be assigned, and the matrix would otherwise be mirrored, this has the advantage that fewer than half of the matrix entries need to be described.According to a further aspect of the present invention, the method is performed iteratively, and when the parameters change, the machine is re-controlled. This has the advantage that the device can react to changes in the parameter values, and the device is controlled accordingly. Furthermore, not all values ​​need to be re-evaluated; rather, it is sufficient if only a fraction of the parameters is changed. This saves computational effort because only a single parameter needs to be changed, and the entire matrix does not have to be rewritten. A new calculation can then be performed.

[0032] According to a further aspect of the present invention, the weighting has a positive and / or negative amount. This has the advantage that individual parameters can be inverted and thus easier to evaluate. The stored weighting can thus indicate which values ​​are better or worse. This makes it possible to invert semantic entries, and thus to identify the control command easier. For example, a selection criterion can be to query what is particularly bad or what is particularly good. These values ​​should be inverse to one another, thus ensuring consistent input. For example, for a first parameter, a particularly high value can be particularly good, and for a second parameter, a particularly high value can be very bad.In order to be able to evaluate this as a whole, one of the parameters can be inverted and thus these individual parameters can be multiplied together to produce a meaningful and machine-automated total amount.

[0033] According to a further aspect of the present invention, qualitative or quantitative parameters are normalized to a maximum value of 1. This has the advantage that all parameter values ​​are made comparable, and in particular, qualitative parameters are also mapped to a comparable value. Here, the value 1 is merely an example value and indicates a particularly positive value, as this can be processed particularly efficiently. In general, it would also be possible to normalize values ​​to other value ranges, for example, to a value between 1 and 1,000. However, according to the invention, it has been empirically shown that the cited example is particularly efficient.

[0034] According to a further aspect of the present invention, the two-dimensional memory structure is stored as a matrix and / or as vectors. This has the advantage of creating particularly efficient memory structures, and thus the two-dimensional data structure can also be broken down into individual vectors of different lengths. Furthermore, it is also possible to provide a single vector that specifies a plurality of values ​​per row. This also makes it possible to perform mathematically simple calculations, and again, the advantageous technical effect is that the proposed method is particularly hardware-efficient.

[0035] According to a further aspect of the present invention, a curve for determining at least one parameter value is read out using the input sensor. This has the advantage that the input sensor can randomly select curve support points, thus resulting in particularly efficient input, which also allows values ​​to be interpolated or extrapolated. This is mapped to the curve, which is then selected and read out by the input sensor. Conversely, the input can also be made from the curve into the matrix.

[0036] According to a further aspect of the present invention, multiple two-dimensional memory structures are provided, and these are normalized to provide a single two-dimensional memory structure. This has the advantage that various parameters can be queried, and then either individual evaluations can be performed, or the memory structures can be superimposed, figuratively speaking, and the values ​​can be queried according to an average. Thus, an average can be calculated for each position in the two-dimensional memory structures.

[0037] According to a further aspect of the present invention, the method is carried out in a distributed computer network and / or locally. This has the advantage that the individual method steps can be executed on different computer systems and, for example, data can also be exchanged via a network. For example, the persistent data storage can be available locally, and the two-dimensional storage structure can be provided via a computer network. This creates technical advantages in transmission and results in greater efficiency.

[0038] According to another aspect, control commands comprise actions and / or action parameters. This has the advantage that not only the actions are specified, but also configurations of the actions can be created. For example, an action may involve replacing a certain component, and the action parameter can be used to specify which new component is to be installed. Furthermore, the action parameters can also describe the action itself, which, for example, must be executed at a specific time.

[0039] According to a further aspect of the present invention, individual method steps are performed iteratively to identify deviations. This has the advantage that the parameters are queried multiple times, allowing incorrect entries to be identified, or the average value from various queries can be calculated and used subsequently.

[0040] According to a further aspect of the present invention, the machine or device is operated in a virtualized manner. This has the advantage that the machine is implemented as a software component, or the device as multiple software components, and thus the method can also be implemented completely virtually, i.e., on a computer. The matrix can be used per criterion, in addition to a criteria matrix.

[0041] The object is also achieved by a system arrangement for hardware-efficient control of a device, comprising an interface unit configured to provide a symmetrical, two-dimensional memory structure with paired selection options, wherein the selection options are continuously arranged in both dimensions and the selection options are present as either qualitative or quantitative parameters; at least one input sensor for assigning parameter values ​​to the parameters, wherein the parameters of each of the paired selection options are only assigned once; a weighting unit configured to apply a weighting stored in a persistent data memory to the parameter values ​​and to normalize the weighted parameter values; and a control unit configured to control the device by means of the weighted and normalized parameter values, wherein groups of parameter values ​​are assigned control commands.

[0042] The problem is also solved by a computer program product with control commands which implement the proposed method or operate the proposed device.

[0043] According to the invention, it is particularly advantageous that the method can be used to operate the proposed system arrangement, devices, and units. Furthermore, the proposed devices and units are suitable for implementing the method according to the invention. Thus, each device implements structural features suitable for executing the corresponding method. However, the structural features can also be configured as method steps. The proposed method also provides steps for implementing the function of the structural features. Furthermore, physical components can also be provided virtually or in a virtualized form.

[0044] Further advantages, features and details of the invention will become apparent from the following description, in which aspects of the invention are described in detail with reference to the drawings. The features mentioned in the claims and in the description can each be essential to the invention individually or in any combination. Likewise, the features mentioned above and those further explained here can each be used individually or in groups in any combination. Parts or components with similar functions or that are identical are sometimes provided with the same reference numerals. The terms “left”, “right”, “top” and “bottom” used in the description of the exemplary embodiments refer to the drawings in an orientation with normally legible figure designations or normally legible reference numerals.The embodiments shown and described are not intended to be exhaustive, but rather are exemplary in nature to illustrate the invention. The detailed description is intended to inform those skilled in the art; therefore, known circuits, structures, and methods are not shown or explained in detail in order not to obscure the understanding of the present description. The figures show:

[0045] Figure 1: a schematic embodiment diagram of a method for hardware-efficient control of a device according to one aspect of the present invention;

[0046] Figure 2: a two-dimensional memory structure which, according to a further aspect of the present invention, is described by means of a curve and vice versa from the curve to the input; and

[0047] Figure 3: a symmetric two-dimensional memory structure according to another aspect of the present invention.

[0048] Figure 1 describes, in a schematic flow diagram, a method for hardware-efficient control of a device, comprising providing 100 a symmetrical, two-dimensional memory structure with paired selection options, wherein the selection options are arranged continuously in both dimensions and the selection options are present as either qualitative or quantitative parameters; reading 101 at least one input sensor for assigning parameter values ​​to the parameters, wherein the parameters of each of the paired selection options are only assigned once; applying 102 a weighting stored in a persistent data memory to the parameter values ​​and normalizing the weighted parameter values; and controlling 103 the device using the weighted and normalized parameter values, wherein groups of parameter values ​​are assigned control commands.

[0049] Figure 2 shows a matrix in which options A, B, C, D, and E are entered in the top columns and the side rows. These are entered consecutively. As shown here, a diagonal is formed because options with the same column and row number do not need to be compared. For example, if it is stated which option is preferred and with which value, it makes no sense to compare option A with the same option A, and therefore no value can be entered in the diagonal, or the value 1 is entered by default. Furthermore, it can be seen here that comparing, for example, option C with option B would produce the same value as comparing option B with option C. It is therefore not necessary to carry out all of these comparisons, and it is therefore possible, figuratively speaking, to fill in only the upper part of the matrix, i.e. above the diagonal.The diagonal would simply mirror the parameter values ​​or mirror them inversely. Thus, it is not necessary to fill the entire matrix; instead, the invention provides that the parameters of each paired selection option are only filled once.

[0050] A paired selection option is, for example, row B with column C or row C with column B. Here, equal values ​​or inverse values ​​arise, so this only needs to be completed once. In one example, it is intended that a decision must be made as to which option is preferred and to what extent this is the case. For example, if option B is compared with option C, a value can be specified that indicates the extent to which option B is preferred over option C. In the opposite case, this would produce an inverse value, so the counter question does not need to be asked. Thus, not all fields of the matrix need to be filled in, and it follows that only the values ​​in pairs above or below the diagonal need to be filled in. Since each paired selection option only needs to be queried once, the number of queries is n squared half minus n half, i.e. n 2 / 2 - n / 2, which is less than half of the total number of fields of the memory structure, and thus the proposed method is particularly efficient.

[0051] The arrows shown in Figure 2 can also be bidirectional according to one aspect of the present invention.

[0052] Figure 3 shows a similarly designed memory structure, where the selection options are also listed consecutively, but the order is inverted. This also results in a diagonal, but opposite to the diagonal in Figure 2. The result, however, is the same, and the technical effect is achieved in a similar way.

Claims

Patent claims 1 . A method for hardware-efficient control of a device, comprising: - providing (100) a symmetrical, two-dimensional memory structure with pairwise selection options, wherein the selection options are arranged continuously in both dimensions and the selection options are present as either qualitative or quantitative parameters; - reading (101) at least one input sensor for assigning parameter values ​​to the parameters, wherein the parameters of each of the paired selection options are assigned only once; - applying (102) a weighting stored in a persistent data store to the parameter values ​​and normalizing the weighted parameter values; and - controlling (103) the device by means of the weighted and standardized parameter values, wherein groups of parameter values ​​are assigned to control commands.

2. The method according to claim 1, characterized in that the input sensor provides a maximum upper limit of inputs which corresponds to half the squared number of selection options minus half the number of selection options.

3. Method according to claim 1 or 2, characterized in that the method is carried out iteratively and when the parameters change, the device is controlled again.

4. Method according to one of the preceding claims, characterized in that the weighting has a positive and / or negative amount.

5. Method according to one of the preceding claims, characterized in that qualitative and / or quantitative parameters are standardized to a maximum value of 1.

6. Method according to one of the preceding claims, characterized in that the two-dimensional memory structure is stored as a matrix and / or as vectors.

7. Method according to one of the preceding claims, characterized in that a curve for determining at least one parameter value is read out by means of the input sensor.

8. Method according to one of the preceding claims, characterized in that a plurality of two-dimensional memory structures are provided and these are normalized in order to provide a single two-dimensional memory structure.

9. Method according to one of the preceding claims, characterized in that the method is carried out in a distributed computer network and / or locally.

10. Method according to one of the preceding claims, characterized in that control commands comprise actions and / or action parameters. 11 . Method according to one of the preceding claims, characterized in that individual method steps are carried out iteratively to identify deviations.

12. Method according to one of the preceding claims, characterized in that the device is operated in a virtualized manner.

13. System arrangement for hardware-efficient control of a device, comprising: - an interface unit configured to provide (100) a symmetrical, two-dimensional memory structure with pairwise selection options, wherein the selection options are arranged continuously in both dimensions and the selection options are present as either qualitative or quantitative parameters; - at least one input sensor for assigning (101) parameter values ​​to the parameters, wherein the parameters of each of the paired selection options are assigned only once; - a weighting unit configured to apply (102) a weighting stored in a persistent data memory to the parameter values ​​and to normalize the weighted parameter values; and - a control unit configured to control (103) the device by means of the weighted and standardized parameter values, wherein groups of parameter values ​​are assigned to control commands.

14. Computer program product, comprising commands that are used during the execution of the program by at least one computer causing it to carry out the steps of the method according to one of claims 1 to 12.

15. A computer-readable storage medium comprising instructions which, when executed by at least one computer, cause the computer to carry out the steps of the method according to any one of claims 1 to 12.

Images