JPH04211828A - Design supporting device - Google Patents

Abstract

PURPOSE:To obtain a design supporting device which realizes the manpower saving and the automation of the selection of parts to be used and can be easily interfaced by a speech recognition function capable of intellectual interaction by natural language. CONSTITUTION:A data inference device 2 connected to a data processor 1 in which an expert system consisting of a knowledge data base and an inference function part is constructed and a speech recognition device as the interface function of this expert system are connected to each other. Manpower required for design work can be drastically saved. Besides, a designer can design by the voice input/out of the natural language, and need not learn a command prepared by a system, and can easily interface it.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a design support system for designing highly reliable electronic equipment used, for example, in artificial satellites.

0002

[Prior Art] Fig. 6 shows, for example, ``Automatic design method of data path system of circuit based on specification description in high-level language'' (Information Processing Society of Japan paper Vol. 28 No. 1) and
176 is an overall system configuration diagram of a conventional design support device that searches for a circuit based on required specifications;
An inference device (22) is connected to the terminal device (21), and a CAD system (23) is connected to this inference device (22).

The above-mentioned CAD system (23) can be broadly classified into: a. A function to search, display, or create a circuit diagram of a logic circuit based on the conditions and specifications required for the circuit; b. A function to perform circuit simulation based on the obtained circuit diagram; c. It consisted of a function to optimize the circuit by inputting certain attributes of the obtained circuit diagram.

The function a above for searching, displaying, or creating a circuit diagram is achieved by the designer using a keyboard or mouse to input an algorithmic operation description written in a high-level language and constraints such as processing time. It is possible to determine the components of a logic circuit. At that time, you can save the circuit diagram, but you can only search and recall the exact same thing.

Further, the circuit simulation function (b) is for displaying in a graph what kind of output signal is obtained when a certain signal is input to the input section of the circuit diagram. The designer looks at the graph, compares it with past experience and knowledge, returns to function a for searching, displaying, or creating a circuit diagram, and makes corrections to the circuit.

[0006] The circuit optimization function (c) performs optimization according to a predetermined procedure by inputting a request such as, for example, the highest speed among equivalent circuits.

[0007] An example of using a CAD system and an inference device is shown in Japanese Patent Laid-Open No. 61-208573, but this system uses a normal terminal device as an interface.

[0008] Furthermore, as a technique for loading a CAD system with a voice input device, as shown in Japanese Patent Laid-Open No. 60-101627, the system prepares commands in advance and the user memorizes and understands them. There is something to input.

[0009]

[Problem to be Solved by the Invention] Conventional design support devices are configured as described above, and are equipped with functions for searching, displaying, or creating circuits that meet certain conditions, a circuit simulation function, and a circuit optimization function. However, selecting the parts to be used required manual work by the designer (manuals and database searches).

[0010] Further, the selection of the parts to be used has problems such as requiring the data processing device or inference device to understand and memorize commands prepared by the system in advance.

[0011] As a way to compensate for this problem, the fixed algorithm and search key shown in "Intelligent CAD/CAT System that Provides Integrated Support for Design and Testing" (Journal of the Japanese Society for Artificial Intelligence Vol. 5 No. 4) There is a means to search for similar circuits.

[0012] When the given circuit and the circuit specifications do not match, this circuit search means searches the circuit diagram stored in the circuit diagram storage device using a predetermined algorithm according to a specific search key. The circuit diagrams can be sorted in the order in which they appear, and the circuit diagrams can be taken out as needed. However, the algorithm and search key for sorting in order of similarity are fixed, so if this algorithm does not match the designer's intuition or heuristics, the designer can search using a unique identifier (file name). In that case, there is a problem that the search efficiency will not be improved, and since the search key is specific to the connection method of parts, for example, how the parts are used and how they are connected. It has been impossible to search for a circuit that outputs a similar signal in response to a given input signal even if the input signals are different.

[0013] This invention has been made to solve the above-mentioned problems, and the invention as claimed in claim 1 saves labor and automates the selection of parts to be used, and enables intellectual dialogue using natural language. The purpose of this invention is to provide a design support device that can easily create an interface using a voice recognition function.

[0014] Furthermore, the invention according to claim 2 does not use a fixed algorithm when determining similarity, but independently stores changeable data for similarity search, and performs circuit search. We have a circuit search method with a learning function that changes the similar search data to match the designer's intuition through dialogue during the process, increasing the probability of matching the similarity judgment made by the designer. The purpose is to obtain a design support device.

[0015]

[Means for Solving the Problems] A design support device according to the invention as claimed in claim 1 connects a data inference device in which an expert system consisting of a knowledge database and an inference function section is constructed to a data processing device, and A voice recognition device is connected as an interface function of the expert system.

In the design support device according to the invention set forth in claim 2, an expert system consisting of a knowledge database and an inference function section is constructed, a data inference device having a similarity search function is connected to a data processing device, and the expert system is A voice dialogue device is connected as an interface function.

[0017]

[Operation] The design support device according to the invention as claimed in claim 1 has the following features:
The expert system, which consists of a knowledge database section and an inference function section, can greatly reduce the labor involved in design work, and as a result, the time required for design can be shortened. Furthermore, by using a voice dialogue device as an interface function of the expert system, the designer can proceed with the design using natural language voice input and output, and there is no need to memorize the commands provided by the system.

The design support device according to the invention described in claim 2 changes the similarity search knowledge through dialogue in the process of circuit search to match the designer's intuition, and matches the similarity judgment made by the designer. Since it has a learning function that increases the probability of The parts to be used can be selected.

[0019]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the overall configuration of an embodiment of the present invention, focusing on hardware. In FIG. 1, a data inference device (2) is connected to a data processing device (1), and this data inference device ( 2) is connected to a voice dialogue device (3). Software is incorporated into the data processing device (1) and data inference device (2) as means, and each process is executed according to an algorithm.

FIG. 2 shows the configuration of the software built into the data processing device (1) and the data inference device (2). ) is the same as the conventional means.

[0021] Furthermore, the dialogue control means (5), intelligent search means (6), and constraint check means (7) incorporated in the data inference device (2) are software, and the selection of parts to be used is made possible by these. , which proceeds in the form of "a natural language dialogue between the designer and the system." The data used are circuit models (8), guidance information (9), parts database (10), and Japanese dictionary grammar for Japanese analysis (1).
1), circuit constraint equation (12).

The dialogue control means (5) uses a circuit model (8
) and the selection status of the parts, determines the next part to be selected and generates a guide sentence to prompt the selection. This prevents the repetition of trial and error that beginners tend to fall into, and the user inputs the required specifications in Japanese according to the guide text.

[0023] The intelligent search means (6) analyzes the input Japanese sentence and generates a database search command from the analysis result. This command is part database (10)
A search is performed, and the search results (part candidates) are displayed in an easy-to-read tabular format. The user narrows down the component candidates by repeatedly inputting requests, selects the component to be used from among them, and the selected component name is sent to the constraint check means (7). The advantage of inputting requirements specifications in Japanese is that it can be used easily even by users who are not familiar with databases, and that conditions can be specified using ambiguous expressions.

The constraint chuck means (7) determines whether the selected component satisfies the relationship (constraint) with other components. If the constraints are not met, the cause is presented and the user is prompted to select the next candidate. Furthermore, when the value of a component (for example, resistance value) is calculated from the use of an already selected component, or when the component name is uniquely determined, the component search is omitted and the component is automatically selected. The list of parts thus obtained is finally output in tabular form.

The actual operation of the design support apparatus configured as described above will be explained using the flowchart shown in FIG.

First, the dialog control means (5) that started reads circuit information from the circuit model (8) and guidance information (9) (step ST3-1) and creates guidance information (step ST3-2). Based on this, the designer (18) is prompted to input the part name, search conditions, etc. using the voice dialogue function (step ST3-3). The part name input or search condition input by the designer using natural language speech is input to the data inference device (2) via the voice recognition device (3) (step ST3-4).

The dialogue control means (5) also analyzes the designer's input sentence based on the Japanese dictionary grammar (11) (step ST3-5). It is determined whether the part names have been narrowed down to one (step ST3-6), and if the names have not been narrowed down, the parts database is searched (step ST3-7) using the intelligent search means (6), and the search results are is displayed (step ST3-8).

When the components have been narrowed down to one, the constraint checking means (7) reads the circuit constraint equation (12),
Determine whether the selected part satisfies the constraint conditions (
Step ST3-9). If not, step S
Return to T3-3 and input the part name or search condition again. If the constraint conditions are satisfied, one component has been determined (step ST3-10).

At that point, the dialog control means (5) sends data to the design data processing means (1) to display the circuit diagram. Furthermore, the dialogue control means (5) determines whether all the parts have been determined from the dialogue so far and the circuit model (8) (step ST3-11), and if all the parts have been determined, sets the parts. If there are any parts that have not been determined yet, the process returns to step ST3-2 and accepts input of parts again.

FIG. 4 is a block diagram showing the overall configuration of another embodiment of the present invention, focusing on hardware, and the same parts as in FIG. The data inference device (2) has a built-in similarity search function (13) as a circuit search means with a learning function. This similarity search function (13) consists of a similarity search means (14), similar search data (15), and a circuit diagram database (16), and is connected to the interaction control means (5).

The circuit diagram created by the circuit diagram creating means (17) incorporated in the data processing device (1) is checked by the constraint checking means (7) to see if required specifications are satisfied. If the required specifications are not met, or if only the required specifications are input, the similar search means (14) uses the circuit model (8), guidance information (9) data, and similar search data (15). Search for a circuit diagram from the diagram database (16). if,
If the designer's judgment differs from the result of the similar search, the similar search data (15) can be rewritten. After the circuit diagram is determined by the similar search means (14) and the circuit diagram creation means (17) of the data processing device (1), the intelligent search means (6), the parts database (10), the constraint check means (7), The components of the circuit can be set using the circuit constraint equation (12).

Next, the actual operation of the design support apparatus configured as in the above embodiment will be explained using the flowchart shown in FIG. First, the designer selects what kind of circuit to search for or to create a new circuit (step ST5-1). When performing a search based on specifications, input the specifications (step ST5-2), and extract the model circuit most suitable for the specifications (step ST5-2).
5-3). The constraint checking means (7) determines whether the circuit satisfies the specifications using the circuit constraint equation (12) (step ST5-4). If the specifications are met, parts selection (step ST5-9) is performed. If the specifications are not met, the similar search means (14) performs a similar search using the similar search data (15) (step ST5-5).
. If the result is desired to be changed, the change is made (step ST5-6). Among the circuits searched by the similar search means (14), the designer selects and calls an appropriate one (step ST5-7). If necessary, the circuit diagram is modified (step ST5-8), and settings are made after the circuit diagram is determined (step ST5-9).

As a result of the selection in step ST5-1, if the circuit diagram is to be searched using a unique identifier (file name), the process moves to step ST5-7 to call the circuit diagram, create/modify the circuit diagram, Selection of parts is performed sequentially. Furthermore, if a new circuit diagram is to be created as a result of the selection in step ST5-1, the process moves to step 5-8 to select parts after creating and correcting the circuit diagram.

[0034]

As described above, according to the invention set forth in claim 1, a voice dialogue device is added to CAD technology as an interface function that enables intellectual dialogue using natural language as an AI function, so that design work can be simplified. Significant labor savings can be achieved, and as a result, the time required for setting can be shortened. In addition, the designer can proceed with the settings through natural language voice input and output, and there is no need to memorize the commands provided by the system, making the interface easier to use.

Further, according to the invention as claimed in claim 2, the similarity search function changes the similarity search knowledge to match the designer's intuition, and the probability of matching the positive similarity judgment made by the designer is increased. There is a learning function that increases, so
Compared to conventional circuit search devices and search methods, this system performs a series of design tasks such as searching for circuits based on required specifications, making corrections if necessary, and selecting specification parts after determining the circuit diagram. This can lead to significant labor savings.

[Brief explanation of the drawing]

FIG. 1 is an overall system configuration diagram showing a design support apparatus according to an embodiment of the present invention.

FIG. 2 is a software configuration diagram according to an embodiment of the invention.

FIG. 3 is a flowchart diagram illustrating actual operations.

FIG. 4 is an overall system configuration diagram showing a design support apparatus according to another embodiment of the present invention.

FIG. 5 is a flowchart diagram illustrating the operation of the embodiment of FIG. 4;

FIG. 6 is an overall system configuration diagram showing a conventional design support device.

[Explanation of symbols]

(1) Data processing device (2) Data inference device (3) Voice dialogue device (13) Similar search function

Claims (2)

[Claims] Claim 1: An expert system consisting of a knowledge database and an inference function unit is constructed, a data inference device connected to a data processing device, and a voice dialogue device connected to the data inference device as an interface function of the expert system. A design support device equipped with 2. An expert system consisting of a knowledge database and an inference function section is constructed, has a similarity search function, a data inference device connected to a data processing device, and the data inference device as an interface function of the expert system. A design support device equipped with a voice dialogue device connected to.