Description:Complete Specification:
The following specification describes and ascertains the nature of this invention and the manner in which it is to be performed

Field of the invention
[0001] The present disclosure relates to the field of requirement engineering. More specifically, it discloses a method and system to process engineering requirements using artificial intelligence techniques.

Background of the invention
[0002] Modern industrial systems such as a vehicle use multiple software’s that support and efficiently execute the functions of the multiple sensors and other hardware’s inside the vehicle. These software’s manifested in the hardware that need requirement engineering. Requirements engineering (RE) is the process of defining, documenting, and maintaining requirements in the engineering design process. When a new project starts, the customer delivers multiple engineering requirements documents at different intervals. These documents come in multiple formats such as DDT sheet, excel, csv and images and are written in different languages. Currently, whenever a new project starts, first the requirement engineer analyzes customer requirements to come up with system requirements, then product owners are involved to write software requirements. Once this is done, then review process starts with the testing team. All in all, this creates confusion for the requirement engineer with respect to different formats, multiple languages, and reference documents thereby causing delay in the final delivery of the project.

[0003] Chinese Patent CN106339366 B titled “Artificial intelligence-based demand recognition method and device” proposes a method based on artificial intelligence demands recognition method and apparatus, wherein, the artificial intelligence-based demand recognition method, comprising the steps of: obtaining demand information; said information word to form a plurality of phrases; obtaining the tightness between the plurality of phrase syntax information; according to the closeness of the syntax information and demand information preset dependency parsing tree needs identification. The present invention is based on artificial intelligence needs identification methods, can effectively improve the demand for recognition accuracy and recognition efficiency.

Brief description of the accompanying drawings
[0004] An embodiment of the invention is described with reference to the following accompanying drawings:
[0005] Figure 1 depicts a system (10) to process engineering requirements and a Requirements engineering (RE) environment;
[0006] Figure 2 illustrates method steps (200) to process engineering requirements;
[0007] Figure 3 is the graphic user interface depiction for one requirement specification extracted using the aforementioned method steps (200) and system (10).

Detailed description of the drawings

[0008] Figure 1 depicts a system (10) to process engineering requirements and a Requirements engineering (RE) environment. A conventional Requirements engineering (RE) environment involves multiple stakeholders like the requirement engineer, product owner and the testing team having to do a lot of manual work and collaboration before the engineering requirement specifications can be formulated and passed on for software development and testing. The proposed system (10) is intended to automate the manual work of the multiple stakeholders and directly come up with the engineering requirement specifications that can be used for software development and testing.

[0009] The system (10) to process engineering requirements comprises an input interface (11), an output interface (14) and at least a processor (12). The input interface (11) adapted to receive a plurality of documents in a plurality of formats, the input interface (11) is in communication with the processor (12). The input interface (11) receives different formats of documents such as word, html, excel, csv, DDT
(Diagnostic Definition Table), CDD (CANdelaStudio diagnostic description) sheet, images and PDF’s.

[0010] The processor (12) can either be a logic circuitry or a software residing in a cloud that responds to and processes logical instructions to get a meaningful result. A hardware processor (12) may be implemented as any or a combination of: one or more microchips or integrated circuits interconnected using a parent board, hardwired logic, software stored by a memory device and executed by a microprocessor (12), firmware, an application specific integrated circuit (ASIC), and/or a field programmable gate array (FPGA). In an exemplary embodiment of the present disclosure the processor (12) executes an AI module.

[0011] An AI module with reference to this disclosure can be defined as reference or an inference set of data, which is use different forms of correlation matrices. Using these models and the data from these models, correlations can be established between different types of data to arrive at some logical understanding of the data. A person skilled in the art would be aware of the different types of AI models such as linear regression, naïve bayes classifier, support vector machine, neural networks and the like. The AI module may be implemented as a set of software instructions, combination of software and hardware or any combination of the same. For example, neural network chips are specialized silicon chips, which incorporate AI technology and are used for machine learning. In the present disclosure the AI model is capable of running Natural language processing (NLP) based algorithms.

[0012] The processor (12) is configured to: read and parse information from the plurality of documents; execute an AI module on the parsed information based on a knowledge graph stored in a memory to extract engineering requirements in a user-defined standard template; update the knowledge graph based on modifications received from an output interface (14). The knowledge graph comprises relationships between keywords of various communication protocols and standards and the corresponding states, functions, values and signals. The AI module is configured to: annotate word vectors in the parsed information to identify states, functions, values and signals; infer engineering requirements based on the annotated word vectors and the knowledge graph.

[0013] The output interface (14) communicates and receives information from the processor (12). The output interface (14) is configured to: display the extracted engineering requirements in the user-defined standard template; receive modification from a user on the extracted engineering requirements. The components of the system (10) to process engineering requirements are distinguished by their functionality. The functionality of aforementioned components is elucidated further while explaining the method steps 200.

[0014] It should be understood at the outset that, although exemplary embodiments are illustrated in the figures and described below, the present disclosure should in no way be limited to the exemplary implementations and techniques illustrated in the drawings and described below.

[0015] Figure 2 illustrates method steps to process engineering requirements. The method steps (200) are performed by the components system (10) to process engineering requirements as elucidated in accordance with figure 1. The method to process engineering requirements uses the AI based system (10) described in accordance with figure 1.

[0016] Method step 201 comprises receiving a plurality of documents in a plurality of formats via the input interface (11). In the context of Requirements engineering (RE) environment the customer(s) input their requirements in different formats of documents such as word, html, excel, csv, DDT/CDD sheet, images and PDF’s.

[0017] Method step 202 comprises reading and parsing information from the plurality of documents by means of the processor (12). In an exemplary embodiment of the proposed method, the system would accurately parse all the above-mentioned documents using advanced python libraries to generate a dictionary as output that could be used as reference while handling the customer requirement. The processor (12) is further configured to read images and tables in the documents, then apply contour detection to identify the state machine or flow chart. In each of the identified contour, Image processing algorithms are applied such as dilation and erosion, thresholding to reduce the noise in the images. Once noise reduction algorithms are applied, then the contour passes through a model that identifies text in the image. This method step works with multiple languages and outputs the string. The output strings for each contour are saved in a linked list and finally a word vectored list is formed from the state machine or the flow chart which can be used for prediction by the AI model.

[0018] Method step 203 comprises executing the AI module on the parsed information based on a knowledge graph stored in a memory of the processor (12) to extract engineering requirements in a user-defined standard template. The knowledge graph comprises relationships between keywords of various communication protocols and standards and the corresponding states, functions, values and signals. The communication protocols and standards with reference to this disclosure include but are not limited to ISO, Automotive Open System (10) Architecture (AUTOSAR), Unified Diagnostic Services (UDS) etc. Here executing the AI module further comprises first annotating word vectors in the parsed information to identify states, functions, values and signals. Then infer engineering requirements based on the annotated word vectors and the knowledge graph.

[0019] In this method step 203 the AI model takes the outputs from the parsed documents including the image information extractors as input and uses the Natural language processing (NLP) algorithms to infer useful details from the extracted information. The parsed information is converted into word vectors, followed by first inferring with the data annotations to identify states, signals, values and keywords. Post this the learnt data (i.e. the signals, states and values of the communication protocols and standards) are matched with already available Knowledge graph to come to an internal requirement. Then the final engineering requirement is generated using either syntax-based output or NLP based algorithm for sentence formation.

[0020] Method step 204 comprises displaying the extracted engineering requirements in the user-defined standard template via the output interface (14). In an exemplary embodiment of the present invention, the output interface (14) can be hosted as a webpage through the server where the processor (12) resides. The output interface (14) includes an interactive UI, where the user can select and modify the internal requirements they choose to. Displaying the extracted engineering requirements further comprises receiving modifications (if any) from a user on the extracted engineering requirements via the output interface (14). If the requirements are modified, we can retrain the AI model for future predictions which completes the cycle. The knowledge graph stored in the memory of the processor (12) is also updated based on modifications received from the user.

[0021] Figure 3 is the graphic user interface depiction for one requirement specification extracted using the aforementioned method steps (200) and the system (10). An example for the same is described below;

Lets us assume a Customer Requirement :-
ID: DIAG_75
RQMT 5.3.1-1 Following table defines generic rule for the session control service, CDD (CANdelaStudio diagnostic description) will define requirements for applications.

Object_Type: Requirement
Inference By AI tool
Once the requirement file is uploaded, the first requirement is loaded, since there is an image in the requirement, the image/Table is first parsed, using image processing algorithms and data is extracted in the form of a dictionary. Then system identifies the keyword - “CDD”, now the tool gets reference from the uploaded CDD sheet.

After the CDD sheet is parsed, then the tool proceeds with NLP matching.
The dictionary that is formed with all the information from the text in the requirement, information from the image/table and the CDD sheet is then matched with the knowledge graph which forms the basis for matching. The best possible matches are then displayed for the user to select the internal requirements in accordance with figure 3. There is an option to select multiple internal requirements for one customer requirement. Also, some part of the internal requirement can be marked as description. If the user decides to modify some requirement, and if the modification is huge, then we retrain the model and update the knowledge graph with the corresponding information.

[0022] A person skilled in the art will appreciate that while these method steps describes only a series of steps to accomplish the objectives, these methodologies may be implemented with custom made modifications and adaptation to the system (10) described herein.

[0023] This idea to develop a method and system (10) to process engineering requirements aims to shorten the duration for SDLC (Software development life cycle) with agile approach. The usage of the proposed method and system (10) provides a common template of requirements for all product areas and multiple projects that is easy to for both developers and testers. The proposed method and system (10) can be integrated with the available frameworks for test case generation and execution to achieve end to end automation. This will reduce the time delay for the customer delivery expectation versus the actual delivery time.

[0024] It must be understood that the embodiments explained in the above detailed description are only illustrative and do not limit the scope of this invention. Any modification to the method and system (10) to process engineering requirements are envisaged and form a part of this invention. The scope of this invention is limited only by the claims.

, Claims:We Claim:
1. A system (10) to process engineering requirements, the system (10) comprising:
an input interface (11) adapted to receive a plurality of documents in a plurality of formats, the input interface (11) in communication with a processor (12);
the processor (12) configured to:
read and parse information from the plurality of documents;
execute an AI module on the parsed information based on a knowledge graph stored in a memory to extract engineering requirements in a user-defined standard template;
update the knowledge graph based on modifications received from an output interface (14).
the output interface (14) configured to:
display the extracted engineering requirements in the user-defined standard template;
receive modification from a user on the extracted engineering requirements.

2. The system (10) to process engineering requirements as claimed in claimed in claim 1, wherein the knowledge graph comprises relationships between keywords of various communication protocols and standards and the corresponding states, functions, values and signals.

3. The system (10) to process engineering requirements as claimed in claimed in claim 1, wherein the AI module is configured to:
annotate word vectors in the parsed information to identify states, functions, values and signals;
infer engineering requirements based on the annotated word vectors and the knowledge graph.

4. A method (200) to process engineering requirements using an AI based system (10), said system (10) comprising an input interface (11), a processor (12) executing an AI module, and at least an output interface (14), the method steps comprising: receiving (201) a plurality of documents in a plurality of formats via the input interface (11); reading and parsing (202) information from the plurality of documents by means of the processor (12); the method steps characterized by:
executing (203) the AI module on the parsed information based on a knowledge graph stored in a memory of the processor (12) to extract engineering requirements in a user-defined standard template;
displaying (204) the extracted engineering requirements in the user-defined standard template via the output interface (14).

5. The method (200) to process engineering requirements as claimed in claim 4, wherein the knowledge graph comprises relationships between keywords of various communication protocols and standards and the corresponding states, functions, values and signals.

6. The method (200) to process engineering requirements as claimed in claim 4, wherein executing the AI module further comprises:
annotate word vectors in the parsed information to identify states, functions, values and signals;
infer engineering requirements based on the annotated word vectors and the knowledge graph.

7. The method (200) to process engineering requirements as claimed in claim 4, wherein after displaying the extracted engineering requirements further comprises receiving modifications from a user on the extracted engineering requirements via the output interface (14).

8. The method (200) to process engineering requirements as claimed in claim 6, wherein receiving modifications further comprises updating the knowledge graph in the memory of the processor (12) based on modifications received the user.