Vehicle monitoring method and assembly

Abstract

The present disclosure generally relates to a vehicle monitoring method and assembly. Image data at least relating to a vehicle is acquired by at least a plurality of cameras in a workshop into which a user of the vehicle cannot enter. And evaluating the at least acquired image data by the control device based on at least one algorithm including generative artificial intelligence. At least the evaluated image data is output by at least one output device.

Description

Technical Field

[0001] This disclosure generally relates to vehicle monitoring methods and assemblies. Background Technology

[0002] Vehicles are often located in environments inaccessible to vehicle owners. For example, when maintenance or repair procedures are performed on a vehicle, the owner typically cannot enter the workshop. At this time, the owner has no opportunity to inspect their vehicle and / or the workflow during the maintenance or repair procedure. On the other hand, vehicle owners today generally demand transparency; for example, they want to know if their vehicle is being handled carefully. Furthermore, vehicle owners often want to know what maintenance or repair procedures have been performed on their vehicle and how they were performed.

[0003] Existing methods, such as those known from US 11411960 B2, US 11358613 B2, US 2019 / 0272508 A1, and US2023 / 0377047 A1, reveal systems where a camera can acquire image data of a vehicle to provide vehicle users with vehicle-related information. However, this image data is not evaluated against an actually visible object. In other words, the camera's perspective may not be suitable for perceiving a vehicle based on ongoing activity.

[0004] Therefore, it is necessary to eliminate or at least reduce the shortcomings of known vehicle monitoring methods and assemblies. In particular, it is necessary to be able to identify specific activities based on the output image data. Summary of the Invention

[0005] This objective is achieved through the subject matter of the independent claims. Advantageous embodiments are given in the dependent claims and the following description, each of which may represent aspects of this disclosure individually or in (sub)combinations. Some features are described for methods, while others are described for assemblies. However, the corresponding aspects are transferred to each other in a corresponding manner.

[0006] According to one aspect, some embodiments of this disclosure relate to a vehicle monitoring method. The method includes at least the following steps: At least vehicle-related image data is collected using at least multiple cameras in a workshop inaccessible to vehicle users.

[0007] The control device evaluates at least the acquired image data based on at least one algorithm, including generative artificial intelligence. This algorithm at least temporarily selects image data corresponding to a specific camera viewpoint of one of a plurality of cameras, based on which maintenance or repair measures to be performed on the vehicle can be optimally identified.

[0008] At least the evaluated image data is output through at least one output device.

[0009] This method is based on the understanding that specific algorithms with corresponding generative artificial intelligence can be used to evaluate which camera perspective best detects a particular maintenance or repair procedure. Consequently, the output image data can be adjusted to allow vehicle users to truly perceive the corresponding maintenance or repair procedure. For example, this avoids outputting video streams that don't allow the user to see a particular maintenance or repair procedure, such as one performed under the vehicle while the camera view is from above. According to this method, the image data selected for vehicle users is always corresponding to the camera perspective that best allows for the perception of the corresponding maintenance or repair procedure. Therefore, even if the vehicle is located in an area inaccessible to the user, they can monitor their vehicle. Furthermore, they can understand the repair process and better understand the potential costs. Thus, the repair process becomes more transparent. They can also learn more about their vehicle and use the output image data to understand how to use their vehicle in a more convenient way, such as preventing expensive parts replacements. This also creates a way to entertain vehicle users.

[0010] According to another aspect, some embodiments of this disclosure relate to a vehicle monitoring assembly. The assembly includes at least a plurality of cameras, a server data storage connected to the cameras, a control device connected to the server data storage, and at least one output device connected to the control device. The cameras are positioned in a workshop inaccessible to vehicle users and are configured to acquire at least vehicle-related image data in the workshop and transmit the acquired image data to the server data storage. The control device is at least configured to evaluate the acquired image data based on at least one algorithm including generative artificial intelligence and output the evaluated image data using at least one output device. The algorithm at least temporarily selects image data corresponding to a specific camera viewpoint of one of the plurality of cameras, based on which maintenance or repair measures performed on the vehicle can be optimally identified.

[0011] The advantages achieved by the method described in this article are also realized in a corresponding manner through the assembly.

[0012] In this article, "the workshop that vehicle users cannot enter" can be understood as an area where, during normal workshop operation, vehicle users are not allowed to enter at will, at least without the consent or permission of the workshop operator or employees. Typically, this restriction on unauthorized entry by vehicle users is to prevent unnecessary disruption during workshop operation.

[0013] In this article, "at least temporarily" can be understood as ensuring the best camera angle within a non-negligible timeframe.

[0014] Preferably, the control device can also evaluate the acquired image data in order to permanently select image data corresponding to the camera viewpoint of one of the multiple cameras (from which the appropriate maintenance or repair measures can be best identified).

[0015] Preferably, the camera can be located inside or outside the vehicle. For example, an existing environmental camera on the vehicle can be used to collect the relevant image data.

[0016] In another approach, cameras can be deployed in the workshop area and positioned so that they can capture specific maintenance or repair procedures.

[0017] In another approach, the camera can be carried by the assembler or workshop employee performing the appropriate maintenance or repair procedures (body camera). For example, workshop employees can wear the camera on their chest or as a head-mounted camera.

[0018] When evaluating the acquired image data, the control device assesses the image data in relation to maintenance or repair measures to be performed or currently being performed. Therefore, it determines the image data that corresponds to a specific viewing angle and provides the optimal perspective for maintenance or repair measures.

[0019] During the output process, the output image data provides the best perspective (angle) for providing appropriate maintenance or repair measures.

[0020] Preferably, the control device also evaluates the acquired image data in accordance with the maintenance protocol to be performed on the vehicle. When outputting the image data, progress information is also output, showing the progress of the maintenance protocol execution. Therefore, the vehicle user can understand the progress of maintenance and repair measures. Consequently, the vehicle user can better estimate when the maintenance protocol will end and when the vehicle can be put back into use.

[0021] Alternatively, the progress information may include additional information about the individual steps involved in the maintenance or repair procedures.

[0022] This can also be used for progress information: for many maintenance or repair procedures, the corresponding processes are clearly defined. In this way, the control unit can use algorithms to reliably assess which step of the work is currently in.

[0023] For example, progress information can be displayed as a progress bar or a progress percentage when outputting.

[0024] Optionally, the control device processes the acquired image data to prevent facial and / or personal information from being identified. This prevents the leakage of personal information during the implementation of the method. It also increases employee acceptance of the method.

[0025] In some embodiments, the control device outputs at least one piece of additional information when outputting image data. This can further increase the amount of information available to the vehicle user.

[0026] Preferably, the additional information relates to the maintenance or repair measures performed. For example, the additional information can be used to convey which components are installed in the vehicle or which components are used for maintenance purposes. Therefore, the additional information is specifically tailored to the maintenance or repair measures performed, thereby further enhancing its relevance to the vehicle user.

[0027] Preferably, the acquired image data is processed by the control device in the server's data storage. This means that the camera transmits the acquired image data to the server's data storage. This eliminates the need for complex computing infrastructure in the workshop area. Instead, the control device can process the data directly on the server's data storage. Therefore, only a single control device needs to be considered for processing different vehicles.

[0028] Optionally, the control device's algorithm includes at least one object tracking algorithm. Specifically, the control device's algorithm may include the YOLO V4 algorithm. This allows for the tracking of specific vehicle components, such as wheels. Using object tracking, image data that best allows for vehicle perception can be automatically selected. Furthermore, object tracking can also be used for components requiring maintenance or repair, such as wheels, fuel tanks, etc.

[0029] Diagnostic data is preferably read from the vehicle via at least one diagnostic device connected to the vehicle's data bus (preferably a CAN bus). The control unit takes the read diagnostic data into account when evaluating the acquired image data. For example, the diagnostic data can be used to determine which part of the vehicle is being repaired or maintained. This information can be used during image data evaluation.

[0030] In some embodiments, the algorithm of the control device includes at least a large language model. Data processed in the large language model is converted into a token sequence and then processed. This enables efficient computational processing of the data.

[0031] Preferably, the large language model includes different transducer architectures (also called transducers) for fine-tuning different tasks, such as identifying different objects or work steps in the context of maintenance and repair measures.

[0032] Large language models are preferably pre-trained and include SoftMax layers in a known manner.

[0033] Optionally, vehicle-related audio data is also collected via at least one microphone in the workshop, in areas inaccessible to the vehicle user, and this audio data is considered by the control unit during the evaluation. This allows the evaluation to be built on a broader foundation of information, thereby improving the accuracy and reliability of the evaluation. For example, assemblers or workshop employees can also provide voice input to instruct on the maintenance or repair procedures they are performing. This information can then be taken into account as part of the evaluation, ensuring the best possible outcome for the vehicle user.

[0034] Preferably, the evaluated audio data is also output by the output device. This creates an additional information channel for vehicle users.

[0035] Preferably, the output device is wirelessly connected to the control device. This means that data can also be output to devices located far from the control device, especially mobile devices. For example, wireless pairing can be performed using mobile communication standards such as Bluetooth, Wi-Fi, or 4G, 5G, or 6G.

[0036] In some embodiments, the output device includes at least one of the following: a screen, particularly a television or computer screen, and a mobile device, such as a smartphone, tablet, laptop, etc. For example, a television may be placed in a customer room (waiting room) associated with the workshop. This means that vehicle users can immediately learn the status of maintenance or repair measures without having to contact employees directly.

[0037] The output image data can also be provided through a web interface accessed by the vehicle user on the corresponding internet-connected device.

[0038] In some embodiments, the method is performed in real time, except for the time required to acquire and evaluate image data and output the evaluated image data. This avoids time loss, allowing vehicle users to receive information immediately.

[0039] Optionally, the method is designed to be implemented by a computer. This means that the steps of the method can be executed by means of one or more data processing devices. In particular, the data processing device of the control unit can trigger or execute the corresponding steps.

[0040] According to another aspect, this disclosure also relates to a computer program product comprising instructions that, when executed by a computer, cause the computer to perform the methods described herein. The advantages achieved by the methods described herein are also achieved in a corresponding manner by the computer program product.

[0041] According to another aspect, this disclosure also relates to a computer-readable storage medium containing instructions that, when executed by a computer, cause the computer to perform the methods described herein. The advantages achieved by the methods described herein are also achieved in a corresponding manner through the computer-readable storage medium.

[0042] For the purposes of this disclosure, "vehicle" may specifically include land vehicles, i.e., off-road and on-road vehicles, such as passenger cars, buses, trucks, and other commercial vehicles. Vehicles may be manned or unmanned. Vehicles may be at least partially electric, having an internal combustion engine and / or an electric motor used as a propulsion system.

[0043] All the features described in each aspect can be combined individually or in combination with other aspects (sub-aspects). Attached Figure Description

[0044] The present disclosure and its other advantageous embodiments and details are described and explained in more detail below with the examples shown in the accompanying drawings. In the drawings: Figure 1 A simplified schematic diagram of a vehicle monitoring assembly according to an embodiment is shown; Figure 2 A simplified schematic diagram of some components of a vehicle monitoring assembly according to an embodiment is shown; Figure 3 A simplified schematic diagram of a vehicle monitoring method according to an embodiment is shown; Figure 4 A simplified schematic diagram illustrating the operation of the assembly according to an embodiment is shown; Figure 5 A simplified schematic diagram of the architecture of the large language model used in the method is shown; Figure 6 A simplified schematic diagram of the architecture of the converter for the large language model used in the method is shown; and Figure 7 A simplified schematic diagram is shown illustrating an exemplary output of image data using an output device within the context of the method. Detailed Implementation

[0045] The following detailed description, taken in conjunction with the accompanying drawings, is intended to describe various embodiments of the disclosed object and is not intended to represent a single embodiment, wherein the same reference numerals in the drawings refer to the same elements. Each embodiment described in this disclosure is by way of example or illustration only and should not be construed as superior to other embodiments. The illustrative examples contained herein are not exhaustive and do not limit the claimed subject matter to the exact forms disclosed. Various variations of the described embodiments will be readily recognized by those skilled in the art, and the general principles defined herein can be applied to other embodiments and applications without departing from the spirit and scope of the described embodiments. Therefore, the described embodiments are not limited to the embodiments shown but have the broadest possible scope of application compatible with the principles and features disclosed herein.

[0046] All features disclosed below with respect to exemplary embodiments and / or drawings may be combined with features of various aspects of this disclosure, either individually or in any sub-combination, including features of preferred embodiments, provided that the resulting combination of features is reasonable to those skilled in the art.

[0047] For the purposes of this disclosure, the phrase "at least one of A, B, and C" means, for example, (A), (B), (C), (A and B), (A and C), (B and C), or (A, B, and C), including all other possible combinations when more than three elements are listed. In other words, the term "at least one of A and B" generally means "A and / or B," i.e., "A" alone, "B" alone, or "A and B."

[0048] Figure 1 A simplified schematic diagram of the vehicle monitoring assembly 10 according to an embodiment is shown.

[0049] According to this embodiment, the assembly 10 includes a plurality of cameras 12 and at least one (typically optional) microphone 14. The cameras 12 and microphone 14 are connected to a server data storage 16 (cloud storage) and are configured to transmit acquired image data or acquired audio data to the server data storage 16.

[0050] The server data storage 16 includes or is connected to at least one control device 18, which includes an algorithm 20 for evaluating the received data.

[0051] According to this embodiment, assembly 10 also includes a vehicle data bus 22, specifically a CAN data bus, which is also connected to server data storage 16. Using data bus 22, diagnostic data can be read from the vehicle and transmitted to server data storage 16. In this case, control device 18 can be used as a diagnostic device to determine the diagnostic data, or control device 18 can access either an internal or external diagnostic device of the vehicle, which determines the diagnostic data and transmits it to control device 18.

[0052] The control device 18 is configured to evaluate received data, including image data, audio data, and diagnostic data, in the server data storage 16 using algorithm 20. Furthermore, the control device 18 is configured to transmit the corresponding evaluated data to the output device 24.

[0053] Optionally, Algorithm 20 has a pre-trained large language model for evaluating the acquired image data. The data processed in the large language model is converted into token sequences and processed in this manner.

[0054] Assembly 10 includes at least one output device 24 configured to output data evaluated by control device 18 to a vehicle user. According to this embodiment, output device 24 can be designed in different ways. For example, output device 24 can be in the form of a web interface 26 accessible via the internet. Alternatively, output device 24 can be a laptop computer 28 or a computer with a screen and / or speakers for outputting the evaluated data. In another alternative, output device 24 can also be a mobile device 30, such as a smartphone 32 or tablet computer 34.

[0055] Figure 2 A simplified schematic diagram of some components of the vehicle monitoring assembly 10 according to an embodiment is shown.

[0056] The control device 18 includes an algorithm 20 and is connected to the server data storage 16. Furthermore, the control device 18 includes a main memory for processing data, at least one processor (data processing device), and internal data storage. Additionally, the control device 18 includes an input / output interface 36.

[0057] Using the input / output interface 36, the control device 18 can communicate with the vehicle control device 38 via the vehicle's data bus 22, for example, by reading the vehicle's diagnostic data via the data bus 22.

[0058] Furthermore, the control device 18 is connected to the output device 24 in the manner already explained, the output device 24 being, for example, a mobile device 30. In particular, the connection between the control device 18 and the output device 24 can be wireless, for example via Bluetooth, Wi-Fi, or mobile communication standards such as 4G, 5G, or 6G.

[0059] Figure 3 A simplified schematic diagram of a vehicle monitoring method 50 according to an embodiment is shown. Optional steps are shown in dashed lines. Typically, method 50 is configured to monitor and track maintenance or repair procedures performed on a vehicle in a workshop inaccessible to the vehicle user, based on output playback data.

[0060] Method 50 includes step S1, in which at least vehicle-related image data is acquired in a workshop inaccessible to vehicle users by at least a plurality of cameras 12.

[0061] Optionally, method 50 may also include step S2, in which vehicle-related audio data is also collected in a workshop inaccessible to the vehicle user via at least one microphone 14.

[0062] Furthermore, method 50 can be extended to read diagnostic data from the vehicle via at least one diagnostic device connected to the vehicle's data bus 22, according to optional step S3. For example, the diagnostic device may be located inside the vehicle or may be formed by the control device 18.

[0063] Method 50 then includes a follow-up step S4, in which the control device 18 evaluates at least the image data acquired in step S1 based on at least one algorithm 20. The algorithm includes generative artificial intelligence. During the evaluation, in step S4, the algorithm 20 at least temporarily selects those image data corresponding to a specific camera viewpoint of one of the plurality of cameras 12, based on which maintenance or repair measures performed on the vehicle can be optimally identified.

[0064] Step S4 can be extended in several ways. For example, according to optional step S5, during the evaluation of the acquired image data, the control device 18 may also consider acquired audio data and / or read diagnostic data. For example, this can determine what maintenance or repair work is being performed on the vehicle, or provide explanations that can be considered from the perspective of assemblers or workshop staff.

[0065] Optionally, method 50 may further include step S6, in which the control device 18 performs additional evaluation on the acquired image data in accordance with the maintenance protocol to be performed on the vehicle. For example, the control device 18 may determine which maintenance or repair measures need to be performed on the vehicle based on the maintenance protocol. Then, the control device 18 may compare the acquired image data with the maintenance or repair actions in the maintenance protocol to determine which maintenance or repair action is currently being performed. Thus, the control device 18 can determine the progress of the maintenance protocol execution.

[0066] Furthermore, step S4 can also be extended by optional step S7, in which the control device 18 processes the acquired image data (and any other acquired audio and / or read diagnostic data) to make the person's face and / or personal information unrecognizable. This prevents the leakage of personal information.

[0067] Generally, step S4 can also be extended by optional step S8, in which the control device 18 evaluates the acquired image data (and any other acquired audio data and / or read diagnostic data) in the server data storage 16.

[0068] Furthermore, according to optional step S9, when the acquired image data is evaluated by algorithm 20 of control device 18, an object tracking function for vehicle components can be generated. For this purpose, algorithm 20 may include an object tracking algorithm, such as the YOLO V4 algorithm.

[0069] Method 50 is followed by step S10, in which the evaluated image data is output through at least one output device 24 (e.g., a moving device 30).

[0070] If method 50 includes optional step S5, the evaluated audio data and / or read diagnostic data may also be optionally considered during output. This allows for the output of combined data to the vehicle user.

[0071] Optionally, method 50 may further include step S11, in which the control device 18 considers additional information when outputting the evaluated image data. This means that, for example, in addition to the pure image data, additional information, such as information about the consumables or replacement parts used, may be output. In this way, the vehicle user can understand which resources were used during the activity.

[0072] If method 50 includes optional step S6, the control device 18 also considers progress information indicating the progress of the maintenance protocol execution during the output of evaluated image data. For example, the progress information may include a progress bar considered during output, indicating the portions of the maintenance and repair measures that have been completed in the maintenance protocol.

[0073] Aside from the time required for image data acquisition and evaluation, and the output of the evaluated image data, method 50 preferably runs in real time. This allows the vehicle user to immediately understand the progress of method 50.

[0074] Figure 4 A simplified schematic diagram of an exemplary operation of assembly 10 according to an embodiment is shown. Optional steps are also shown in dashed lines here.

[0075] Figure 4 The operating mode of assembly 10 shown is merely exemplary, and may take different forms in other embodiments of method 50.

[0076] According to this embodiment, in step S1, multiple cameras 12 are used to collect image data of the vehicle in a workshop that the vehicle user cannot access.

[0077] In the next step S4, the control device 18 evaluates the acquired image data. According to this embodiment, an object tracking algorithm is used for this purpose, according to optional step S9.

[0078] Therefore, in optional step S12, the position and arrangement of the predefined components of the vehicle, which are the result of optional step S9, are determined and tracked in the acquired image data.

[0079] Meanwhile, diagnostic data is read from the vehicle’s data bus 22 according to optional step S3, and according to this embodiment, the diagnostic data is considered during the evaluation in step S4.

[0080] Subsequently, optional step S13 is used by the control device 18 to evaluate the acquired image data in the server data storage 16 using a large language model.

[0081] By tracking and using this evaluation method, control device 18 can determine which image data is best suited to provide the vehicle user with the optimal view to perceive appropriate maintenance or repair measures. This means that control device 18 can select the camera angle that best observes maintenance or repair measures based on object tracking. According to this embodiment, control device 18 will consider the diagnostic data it reads.

[0082] In optional step S14, the evaluation results are provided by the control device 18.

[0083] Subsequently, according to step S10, the output device 24 outputs the evaluated image data. Artificial intelligence is used to select the optimal camera angle for vehicle users so that they can perceive the performed maintenance or repair measures in the best way.

[0084] Figure 5A simplified schematic diagram of the architecture of the pre-trained large language model 52 used in method 50 is shown. The large language model 52 is well known.

[0085] 54 represents the high-level structure of the large language model 52. Based on text and location embeddings, corresponding input data is provided to twelve parallel working blocks 56. Working blocks 56 have a masked multi-head attention unit on the input side, followed by layer normalization and a feedforward controller. Starting from another unit used for layer normalization, the results of working blocks 56 are used as part of a unit for text prediction and / or task classifiers.

[0086] The classifier is represented by 58, where an excerpt is selected from the corresponding text starting from the starting unit, and the excerpt is fed to the converter 66 and then linearized.

[0087] 60 represents the resulting unit that establishes the premise starting from the initial unit and then restricts it. As a result, a hypothesis is obtained, from which an extract is extracted, which is then fed into converter 66 and linearized.

[0088] 62 represents a similarity unit that performs two parallel data processing operations. Starting from the starting unit, the first text portion "Text1" is fed to the constraint unit in the first chain 63A. This produces the second text portion "Text2", from which an excerpt is extracted and then fed to the converter 66. Simultaneously, starting from the starting unit in the second chain 63B, the second text portion "Text2" is fed to the constraint unit. This (optimally) results in the first part of the text "Text1", from which an excerpt is extracted and then fed to the converter 66. Starting from the two converters 66, the corresponding results are added and fed to the linearization unit.

[0089] In addition, a selection problem (multiple choice) unit 64 is used. The selection problem unit 64 typically comprises n (n is a natural number) parallel lines 65A to 65N. For each chain 65, starting from the initial unit, a scenario unit is used and constrained to produce the nth response corresponding to the nth chain 65N. The nth response is fed to an extraction unit, which extracts a part, which is then fed to a converter 66. The results of the various converters 66 are then combined.

[0090] like Figure 5 As shown, the processed data is converted into token sequences in a pre-trained large language model 52 using classifier unit 58, result unit 60, similarity unit 62 and selection question unit 64, and processed in this way.

[0091] Figure 6 A simplified schematic diagram of the architecture of the converter 66 of the large language model 52 used in the context of method 50 is shown.

[0092] The converter 66 includes a first section 68A. Within the first section 68A, data is fed from the input of a unit to an input embedding unit. Then, positional encoding of the data is considered. Next, the positionally encoded data is fed to a multi-head attention mechanism, followed by summation and normalization. This processed data is then fed to a feedforward controller, where it undergoes summation and normalization again, determining the result of the converter's first section 68A.

[0093] The converter 66 also includes a second part 68B, which considers the output. The output and input are staggered. Starting with the output, data is fed to an output embedding unit. Next, positional encoding of the data is considered. Afterward, the positionally encoded data is fed to a masked multi-head attention mechanism, followed by summation and normalization. This processed data is now fed to another multi-head attention unit, which also receives the output data (result) of the first part 68A of the converter 66. After a subsequent summation and normalization unit, the data is fed to another feedforward controller, followed by another summation and normalization unit. The processed data is then linearized and fed to the Softmax layer of the converter 66. As a result of the second part 68B, the converter 66 provides output probabilities for different data at different outputs relative to specific input data. Therefore, the output probabilities indicate the magnitude of the probability that a specific input data of the converter 66 will ultimately correspond to a specific output data of the converter 66.

[0094] The converter 66 is structure-based and ultimately output-probability-based, ensuring fine-tuning of the transformation in relation to different output data with respect to specific input data.

[0095] Figure 7 A simplified schematic diagram is shown of an exemplary output 70 of image data using output device 24 in the context of method 50.

[0096] Here, smartphone 32 is used as an example for output 70, acting as output device 24. Smartphone 32 has a screen 72. The evaluated image data is displayed on screen 72. Furthermore, according to this example output 70, the read diagnostic data 74 is displayed in an overlaid display window. Although the diagnostic data 74 is displayed directly here, this is not mandatory. The diagnostic data 74 may also be used only when evaluating the acquired image data without being displayed. Additionally, according to this exemplary output 70, optional additional information 76 is displayed, such as which components were used in the context of maintenance or repair measures. Furthermore, according to an exemplary embodiment, output 70 includes a progress bar to display progress information 78. The progress bar indicates the progress of activities relative to the maintenance protocol. The additional information 76 and progress information 78 are also optional for the output of the evaluated image data.

[0097] The specific embodiments disclosed herein use circuits (e.g., one or more circuits) to implement the standards, protocols, methods, or techniques disclosed herein to functionally connect two or more components, generate information, process information, analyze information, generate signals, encode / decode signals, convert signals, send and / or receive signals, control other devices, etc. Any type of circuit can be used.

[0098] In one embodiment, the circuitry, such as the control device, includes one or more data processing devices, such as a processor (e.g., a microprocessor), a central processing unit (CPU), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a system-on-a-chip (SoC), or similar devices, or any combination thereof, and may include discrete digital or analog circuit elements or electronic devices, or combinations thereof. In one embodiment, the circuitry includes hardware circuitry implementations (e.g., implementations in analog circuitry, implementations in digital circuitry, etc., and combinations thereof).

[0099] In one embodiment, the circuit comprises a combination of circuitry and a computer program product, wherein software or firmware instructions are stored in one or more computer-readable storage media and cooperate to cause the device to perform one or more of the protocols, methods, or techniques described herein. In one embodiment, the circuitry technology includes circuitry that requires software, firmware, etc., to operate, such as a microprocessor or components thereof. In one embodiment, the circuitry includes one or more processors or components thereof, along with associated software, firmware, hardware, etc.

[0100] This disclosure refers to quantities and numbers. Unless explicitly stated otherwise, these quantities and numbers should not be considered limitations, but rather as examples of possible quantities or numbers relating to this disclosure. In this context, the term "plural" may also be used to refer to quantities or numbers. In this context, the term "multiple" refers to any number greater than one, such as two, three, four, five, etc. The terms "approximately," "approximately," "close to," etc., indicate plus or minus 5% of the stated value.

[0101] Although this disclosure has been presented and described in conjunction with one or more embodiments, equivalent changes and modifications will be able to be made by those skilled in the art after reading and understanding this specification and the accompanying drawings.

Claims

1. A vehicle monitoring method, comprising the following steps: At least image data related to the vehicle is collected in a workshop inaccessible to vehicle users using at least multiple cameras; The control device evaluates at least the acquired image data based on at least one algorithm including generative artificial intelligence, wherein the algorithm at least temporarily selects those image data corresponding to a specific camera viewpoint of one of the plurality of cameras, and based on these image data, it is possible to optimally identify maintenance or repair measures performed on the vehicle. as well as At least the evaluated image data is output through at least one output device.

2. The method according to claim 1, wherein, The control device also evaluates the acquired image data in accordance with the maintenance protocol to be performed by the vehicle, and when outputting the image data, it also outputs progress information indicating the progress of the maintenance protocol execution.

3. The method according to claim 1, wherein, The control device processes the acquired image data to make it impossible to identify a person's face and / or personal information.

4. The method according to claim 1, wherein, The control device also outputs at least one piece of additional information when outputting the image data.

5. The method according to claim 1, wherein, The acquired image data is processed in the server's data storage via the control device.

6. The method according to claim 1, wherein, The algorithm of the control device includes at least one object tracking algorithm.

7. The method according to claim 1, wherein, Diagnostic data is read from the vehicle via at least one diagnostic device connected to the vehicle's data bus, and the control device takes the read diagnostic data into account when evaluating the acquired image data.

8. The method according to claim 1, wherein, The algorithm of the control device includes at least one large language model, wherein the data processed in the large language model is converted into a token sequence and processed.

9. The method according to claim 1, wherein, The system also collects vehicle-related audio data in the workshop, which is inaccessible to the vehicle user, via at least one microphone, and the control device takes the audio data into account during the evaluation.

10. The method according to claim 1, wherein, The output device is wirelessly connected to the control device.

11. The method according to claim 1, wherein, The method is performed in real time, except for the time required for the acquisition and evaluation of the image data and the output of the evaluated image data.

12. A vehicle monitoring system, comprising: Two or more cameras; Server data storage connected to the camera; Control device connected to the server data storage device; as well as At least one output device connected to the control device, The camera is located in a workshop inaccessible to vehicle users and is configured to collect vehicle-related image data in the workshop and transmit the collected image data to the server data storage. The control device is configured to evaluate the collected image data based on at least one algorithm including generative artificial intelligence and output the evaluated image data through the at least one output device. The algorithm at least temporarily selects those image data corresponding to a specific camera viewpoint of one of the plurality of cameras, based on which maintenance or repair measures performed on the vehicle can be optimally identified.

13. The system according to claim 12, wherein, The control device also evaluates the acquired image data in accordance with the maintenance protocol to be performed by the vehicle, and when outputting the image data, it also outputs progress information indicating the progress of the maintenance protocol execution.

14. The system according to claim 12, wherein, The control device processes the acquired image data to make it impossible to identify a person's face and / or personal information.

15. The system according to claim 12, wherein, The control device also outputs at least one piece of additional information when outputting the image data.

16. The system according to claim 12, wherein, The acquired image data is processed by the control device in the server's data storage.

17. The system according to claim 12, wherein, The algorithm of the control device includes at least one object tracking algorithm.

18. The system according to claim 12, wherein, Diagnostic data is read from the vehicle via at least one diagnostic device connected to the vehicle's data bus, and the control device takes the read diagnostic data into account when evaluating the acquired image data.

19. The system according to claim 12, wherein, The algorithm of the control device includes at least one large language model, wherein the data processed in the large language model is converted into a token sequence and processed.

20. The system according to claim 12, wherein, The system also collects vehicle-related audio data in the workshop, which is inaccessible to the vehicle user, via at least one microphone, and the control device takes the audio data into account during the evaluation.

Images