System and method for validating coding data of a vehicle

Abstract

The present disclosure relates to a system (100) for validating coding data (COD) of a vehicle, comprising: an input module (110) configured to receive a coding file containing coding data (COD) to be validated; a validation module (120) comprising at least one trained speech recognition algorithm (122), wherein the at least one trained speech recognition algorithm (122) is configured to check the coding data (COD) in a speech-based manner in order to detect errors in the coding data (COD); and an output module (130) configured to output a check result from the validation module (120).

Description

[0001] System and method for validating coding data of a vehicle

[0002] The present disclosure relates to a system for validating coding data of a vehicle, a method for validating coding data of a vehicle, a method for training a speech recognition algorithm for validating coding data of a vehicle, and a system for training a speech recognition algorithm for validating coding data of a vehicle. In particular, the present disclosure relates to intelligent algorithms for validating vehicle coding files in order to avoid errors that occur, for example, during the configuration of vehicle control units.

[0003] State of the art

[0004] In vehicle construction, there are often problems with so-called coding data errors that lead to functional impairments in the vehicle's software. One example of this is the requirement in the USA that the headlight range adjustment must not work with high-beam headlights. The coding data that represents this could be, for example, [country = US, HighBeam = YES, BeamAdaptation = AO]. Incorrect coding can lead to a function being incorrectly deactivated or violating legal regulations. The multitude of such rules for different countries, vehicle variants, and equipment variants makes the susceptibility to errors particularly high. Furthermore, often only the vehicle manufacturer has the necessary knowledge of which functions should or may work with which variants.

[0005] One coding data review method relies on the dual control principle, where human reviewers monitor the coding data. However, this approach is prone to human error, knowledge gaps, and the oversight of details. An alternative method for reviewing coding data relies on specialized software-based tools that vehicle manufacturers can purchase externally. The problem, however, is that the expertise of many of the vehicle manufacturer's experts must be integrated into the tool and continuously updated. This process is not only time- and resource-intensive but also carries the risk of errors.

[0006] Disclosure of the invention

[0007] It is an object of the present disclosure to provide a system for validating coding data of a vehicle, a method for validating coding data of a vehicle, a method for training a speech recognition algorithm for validating coding data of a vehicle, a system for training a speech recognition algorithm for validating coding data of a vehicle, and a storage medium for executing the methods, which enable correct control and / or configuration of technical vehicle functions. In particular, it is an object of the present disclosure to detect errors in coding data of a vehicle.

[0008] This object is achieved by the subject matter of the independent claims. Advantageous embodiments are specified in the subclaims.

[0009] According to an independent aspect of the present disclosure, a system for validating coding data of a vehicle, in particular a motor vehicle, is provided. The system comprises an input module configured to receive a coding file with coding data to be validated; a validation module comprising at least one trained speech recognition algorithm, wherein the at least one trained speech recognition algorithm is configured to perform a speech-based check of the coding data in order to detect errors in the coding data; and an output module configured to output a check result of the validation module.

[0010] According to the invention, the coding data is treated as a language that the speech recognition algorithm learns. The trained speech recognition algorithm can then review the coding data and find errors. This allows the coding data to be treated like human speech, checking for errors, e.g., grammar, etc. Detected errors can then be corrected in the coding data automatically or by a user. This results in correct configuration and thus control of technical vehicle functions.

[0011] The input module, the validation module and the output module may comprise software components / algorithms configured to be executed on at least one processor and thereby to perform the functionalities of the respective module.

[0012] The input module and / or the validation module and / or the output module can be implemented in a common software and / or hardware module. Alternatively, the input module and / or the validation module and / or the output module can each be implemented in separate software and / or hardware modules.

[0013] The term "vehicle" includes cars, trucks, vans, buses, mobile homes, motorcycles, etc., used to transport people, goods, etc. In particular, the term includes motor vehicles used to transport people.

[0014] The term "coding data," as used in this disclosure, refers to information contained in coding files (CAF files = Coding Application Files) used to customize the software installed on a vehicle. These coding files play a crucial role in fine-tuning the vehicle's electronic control units (ECUs) to ensure that the software meets the specific requirements of, for example, a customer and that various vehicle functions, such as engine performance, exhaust emissions, braking, steering, safety systems, and other aspects of vehicle operation, function properly.

[0015] The coding data is generated through Boolean expression evaluation, which determines which parameters apply to which functions within the software. This "fine-tuning" allows for setting function values ​​and ensuring that each control unit in the vehicle is configured according to its specific requirements. Coding is therefore performed at the software parameter and function level to ensure a tailored software configuration for each vehicle.

[0016] Preferably, the coding data is provided to individually configure software installed on the vehicle's control units. Control units are electronic modules that control and / or monitor various functions and systems in the vehicle. Each control unit can be specialized for a specific function or system. Examples of control units include, but are not limited to, engine control units, transmission control units, brake control units, climate control units, infotainment control units, body control units, and steering control units. These control units are connected via an internal network, such as system buses (CAN, LIN, MOST, FlexRay, Ethernet, etc.), and can exchange information with each other.

[0017] Preferably, the coding data is provided to individually configure software installed on the vehicle's control units by setting function values. The term "function values," as used in the present disclosure, refers to parameters for functions within the software.

[0018] The term "speech recognition algorithm," as used in the present disclosure, is a type of artificial intelligence (AI) capable of understanding and interpreting speech. The speech recognition algorithm may, for example, use various machine learning and deep neural network techniques to grasp the complexity of the encoding data (speech) and detect errors.

[0019] Preferably, the at least one trained speech recognition algorithm is configured to perform a speech-based check of the coding data by analyzing a grammar of the coding data to detect errors in the coding data. The term "grammar" refers to learned rules according to which, for example, function values ​​are combined to form valid configurations.

[0020] Preferably, the at least one speech recognition algorithm is trained using machine learning. Machine learning (ML) refers to the artificial generation of knowledge from experience. Here, algorithms learn from empirical data and, after the learning phase, are able to generalize the experiences. To do this, the algorithms build a statistical model based on the empirical data. The empirical data used to train the algorithms is generally referred to as "training data."

[0021] Preferably, the at least one speech recognition algorithm is trained using valid coding data. In other words, the training data can include or be valid coding data. The term "valid coding data" refers in particular to coding data with correctly set function values. However, the present disclosure is not limited thereto, and the at least one speech recognition algorithm can additionally or alternatively be trained using invalid coding data. The term "invalid coding data" refers in particular to coding data with incorrectly set function values.

[0022] Preferably, the at least one speech recognition algorithm comprises, or is, at least one neural network, in particular a trained neural network. Neural networks consist of a series of processing units that are connected to one another via communication channels, the so-called neurons. The neurons process input data and pass the result of the processing on to other neurons via the communication channels. Through the parallel processing of data, which is made possible by linking the neurons and their processing functions, complex, non-linear dependencies in the input data can be mapped. Neural networks are trained using a suitable learning technique, e.g. machine learning, so that they learn these dependencies. The training of the neural networks can be carried out using empirical data, the so-called training data.

[0023] The neural network may be selected from the group comprising or consisting of a deep neural network, a convolutional neural network and a recurrent neural network.

[0024] Preferably, the at least one speech recognition algorithm comprises, or is, a Large Language Model (LLM). LLMs are artificial neural networks characterized by their ability to generate and understand speech. LLMs acquire these capabilities by learning statistical relationships from training data, e.g., during a self-supervised or semi-supervised training process.

[0025] Preferably, the system for validating a vehicle's coding data is implemented in a central unit (e.g., server or backend) and / or is cloud-based. For example, the coding data can be validated by the system before and / or after it is uploaded to the vehicle. For this purpose, suitable communication connections, e.g., between the system and the vehicle, can be used to retrieve and validate the coding data. In this example, the system's input module can be, for example, a communication module.

[0026] The system's output module is configured to output a verification result from the validation module so that detected errors can be corrected.

[0027] In some embodiments, the system or another designated system automatically corrects the detected errors. For example, the system can correct the detected errors and directly import the corrected version of the coding file back into the vehicle or the corresponding control units.

[0028] In other embodiments, the correction can be performed by a user. For example, the output module can comprise or be a user interface module configured to display the verification result, in particular, detected errors in the coding file, to the user. The user can then manually correct the detected errors through user input or confirm a correction suggested by the system through user input.

[0029] The user interface module may comprise at least one display device and at least one input device. The at least one display device may comprise or be a display, in particular an LCD display, a plasma display, or an OLED display. The at least one input device may comprise a voice input device and / or a touch-sensitive input device, such as a touch panel or touch pad, and / or a tactile input device, such as a switch (e.g., push-button switch and / or rotary switch) or other mechanically actuatable key elements.

[0030] Preferably, the user interface module comprises, or is, a touchscreen providing the at least one display device and the at least one input device.

[0031] According to a further independent aspect of the present disclosure, a method for validating coding data of a vehicle, in particular a motor vehicle, is provided. The method comprises receiving a coding file with coding data to be validated; performing a speech-based check of the coding data using at least one trained speech recognition algorithm to detect errors in the coding data; and outputting a check result or validation result.

[0032] The method for validating coding data of a vehicle may implement the aspects of the system for validating coding data of a vehicle described in this document.

[0033] According to a further independent aspect of the present disclosure, a method for training a speech recognition algorithm for validating coding data of a vehicle, in particular a motor vehicle, is provided. The method comprises training at least one speech recognition algorithm using valid coding data to enable the at least one speech recognition algorithm to perform a speech-based check of the coding data to be validated in order to detect errors in the coding data.

[0034] According to a further independent aspect of the present disclosure, a software (SW) program is provided. The SW program can be configured to run on one or more processors and thereby execute the method described in this document for validating coding data of a vehicle and / or the method for training a speech recognition algorithm for validating coding data of a vehicle.

[0035] According to a further independent aspect of the present disclosure, a storage medium is provided. The storage medium can comprise a software program configured to be executed on one or more processors and thereby to execute the method described in this document for validating coding data of a vehicle and / or the method for training a speech recognition algorithm for validating coding data of a vehicle.

[0036] According to a further independent aspect of the present disclosure, software with program code is provided. The software is configured to perform the method for validating coding data of a vehicle and / or the method for training a speech recognition algorithm for validating coding data of a vehicle when the software runs on one or more software-controlled devices.

[0037] According to a further independent aspect of the present disclosure, a system is provided. The system comprises one or more processors; and at least one memory connected to the one or more processors and containing instructions executable by the one or more processors to perform the method described in this document for validating coding data of a vehicle and / or method for training a speech recognition algorithm for validating coding data of a vehicle.

[0038] A processor or processor module is a programmable computing unit, i.e. a machine or an electronic circuit that controls other elements according to given instructions and thereby drives an algorithm (process).

[0039] Short description of the drawings

[0040] Embodiments of the disclosure are illustrated in the figures and are described in more detail below. They show:

[0041] Figure 1 schematically shows a system for validating coding data of a vehicle according to embodiments of the present disclosure,

[0042] Figure 2 is a flowchart of a method for validating coding data of a vehicle according to embodiments of the present disclosure, and

[0043] Figure 3 is a flowchart of a method for training a speech recognition algorithm to validate coding data of a vehicle according to embodiments of the present disclosure.

[0044] Embodiments of the disclosure

[0045] Unless otherwise noted, the same reference numerals are used below for identical and equivalently functioning elements. Figure 1 schematically shows a system 100 for validating coding data COD of a vehicle according to embodiments of the present disclosure.

[0046] The coding data (COD) can be used to individually configure software installed on a vehicle's control units by setting function values. The coding data is generated, for example, using Boolean term evaluation, which determines which parameters apply to which functions within the software. This allows the function values ​​to be set and ensures that each control unit in the vehicle is configured according to its specific requirements. Coding is therefore performed at the level of software parameters and functions to ensure a tailored software configuration for each vehicle.

[0047] The system 100 comprises an input module 110 configured to receive a coding file COD with coding data COD to be validated; a validation module 120 comprising at least one trained speech recognition algorithm 122, wherein the at least one trained speech recognition algorithm 122 is configured to check the coding data COD in a speech-based manner in order to detect errors in the coding data COD; and an output module 130 configured to output a check result of the validation module 120.

[0048] In some embodiments, the system 100 can be implemented in a central unit (e.g., server or backend) and / or cloud-based. For example, the coding data COD can be validated by the system 100 before being uploaded to the vehicle and / or after being uploaded to the vehicle. In the latter case, suitable communication connections, e.g., between the system 100 and the vehicle, can be used to retrieve and validate the coding data COD. In this example, the input module 110 of the system 100 can, e.g., be a communication module configured for wireless or wired communication with the vehicle. The at least one speech recognition algorithm 122 is trained such that it can understand and interpret speech.The at least one speech recognition algorithm can, for example, use various machine learning and deep neural network techniques to capture the complexity of the coding data (speech) and detect errors. For example, the at least one trained speech recognition algorithm 122 can be configured to check the coding data COD based on language by analyzing a grammar of the coding data COD to detect errors in the coding data. The term "grammar" refers to learned rules according to which, for example, function values ​​are combined to form valid configurations.

[0049] The output module 130 of the system 100 is configured to output a verification result of the validation module 120 so that detected errors can be corrected. In some embodiments, the system 100 can automatically correct the detected errors. For example, the system 100 can correct the detected errors and directly import the corrected version of the coding file COD back into the vehicle or the corresponding control units.

[0050] In other embodiments, the correction can be performed by a user. For example, the output module 130 can include or be a user interface module configured to display the verification result, in particular, detected errors in the coding file, to the user. The user can then manually correct the detected errors through user input or confirm a correction suggested by the system through user input.

[0051] Figure 2 shows a flowchart of a method 200 for validating coding data of a vehicle according to embodiments of the present disclosure. Method 200 can be implemented by appropriate software executable by one or more processors (e.g., a CPU). Method 200 includes, in block 210, receiving a coding file with coding data to be validated; in block 220, a speech-based checking of the coding data by at least one trained speech recognition algorithm to detect errors in the coding data; and, in block 230, outputting a check result or validation result.

[0052] Figure 3 shows a flowchart of a method 300 for training a speech recognition algorithm to validate coding data of a vehicle according to embodiments of the present disclosure. Method 300 may be implemented by appropriate software executable by one or more processors (e.g., a CPU).

[0053] The method 300 comprises, in block 310, training at least one speech recognition algorithm using valid coding data to enable the at least one speech recognition algorithm to perform a speech-based check of coding data to be validated in order to detect errors in the coding data; and, in block 320, providing the trained speech recognition algorithm for validating coding data of a vehicle.

[0054] In some embodiments, the at least one speech recognition algorithm can be trained using machine learning. The data used to train the at least one speech recognition algorithm is generally referred to as "training data."

[0055] The at least one speech recognition algorithm can be trained, for example, using valid coding data. In other words, the training data can include or be valid coding data. The term "valid coding data" refers in particular to coding data with correctly set function values. However, the present disclosure is not limited thereto, and the at least one speech recognition algorithm can additionally or alternatively be trained using invalid coding data. The term "invalid coding data" refers in particular to coding data with incorrectly set function values.

[0056] In some embodiments, the at least one speech recognition algorithm comprises or is at least one neural network, in particular a trained neural network. The neural network may, for example, comprise or be a deep neural network, a convolutional neural network, or a recurrent neural network.

[0057] In some embodiments, the at least one speech recognition algorithm comprises or is a Large Language Model (LLM). LLMs are artificial neural networks characterized by their ability to generate and understand speech. LLMs acquire these capabilities by learning statistical relationships from training data, e.g., during a self-supervised or semi-supervised training process.

[0058] According to the invention, the coding data is treated as a language that the speech recognition algorithm learns. The trained speech recognition algorithm can then review the coding data and find errors. This allows the coding data to be treated like human speech, checking for errors, e.g., grammar, etc. Detected errors can then be corrected in the coding data automatically or by a user. This results in correct configuration and thus control of technical vehicle functions.

[0059] Although the invention has been illustrated and explained in detail by means of preferred embodiments, the invention is not limited by the disclosed examples, and other variations may be derived therefrom by those skilled in the art without departing from the scope of the invention. It is therefore clear that a multitude of possible variations exist. It is also clear that the embodiments mentioned by way of example are truly only examples and should not be construed as limiting the scope, possible applications, or configuration of the invention in any way.Rather, the preceding description and the description of the figures enable the person skilled in the art to implement the exemplary embodiments in concrete terms, whereby the person skilled in the art, with knowledge of the disclosed inventive concept, can make various changes, for example with regard to the function or the arrangement of individual elements mentioned in an exemplary embodiment, without departing from the scope of protection defined by the claims and their legal equivalents, such as further explanations in the description.

Claims

Patent claims 1. System (100) for validating coding data (COD) of a vehicle, comprising: an input module (110) configured to receive a coding file with data to be validated to receive coding data (COD); a validation module (120) comprising at least one trained speech recognition algorithm (122), wherein the at least one trained speech recognition algorithm (122) is configured to check the coding data (COD) in a speech-based manner in order to detect errors in the coding data (COD); and an output module (130) configured to output a check result of the validation module (120).

2. System (100) according to claim 1, wherein the coding data (COD) are provided to individually configure software applied to control units of the vehicle, in particular by setting function values.

3. System (100) according to claim 1 or 2, wherein the at least one trained speech recognition algorithm (122) is configured to check the coding data (COD) in a speech-based manner by analyzing a grammar of the coding data (COD) to detect errors in the coding data (COD).

4. System (100) according to one of claims 1 to 3, wherein the at least one speech recognition algorithm (122) is trained by means of machine learning.

5. System (100) according to one of claims 1 to 4, wherein the at least one speech recognition algorithm (122) is trained using valid and / or invalid coding data.

6. System (100) according to one of claims 1 to 4, wherein the at least one speech recognition algorithm (122) comprises or is a large language model.

7. A method (200) for validating coding data (COD) of a vehicle, comprising: Receiving (210) a coding file with coding data (COD) to be validated; speech-based checking (220) of the coding data (COD) by at least one trained speech recognition algorithm (122) to detect errors in the coding data (COD); and Output (230) a verification result.

8. A method (300) for training a speech recognition algorithm (122) for validating coding data (COD) of a vehicle, comprising: Training (310) at least one speech recognition algorithm (122) using valid coding data in order to enable (122) the at least one speech recognition algorithm to check coding data (COD) to be validated in a speech-based manner in order to detect errors in the coding data (COD).

9. A storage medium comprising a software program configured to be executed on one or more processors and thereby to carry out the method (200, 300) according to claim 7 or 8.

10. A system for training a speech recognition algorithm (122) for validating coding data (COD) of a vehicle, comprising: one or more processors; and at least one memory connected to the one or more processors and containing instructions executable by the one or more processors to perform the method (300) for training a speech recognition algorithm (122) for validating coding data (COD) of a vehicle according to claim 8.

Images