JP2014041390A - Design and development support system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve a problem that a designer/developer cannot completely recognize which parts of a target built-in system users using the built-in system feel unsatisfied with.SOLUTION: Various sensors are attached to a target built-in system, and various actions which users take when feeling unsatisfied are detected on the basis of information of the sensors. With a detection result as a trigger, behavior information of the system at timings of the actions is transmitted to a development environment accessible by a designer/developer. The designer/developer reads in the behavior information from an arbitrary design tool at an arbitrary timing. The design tool finds out which part on design causes disappointment indicated by the behavior information. Finally, the result is displayed on a screen to clearly specify which part in the system is problematic, to the designer/developer.

Description

  The present invention relates to an embedded system design / development support system.

  Conventionally, it has been difficult to improve the operational feeling (usability) of embedded systems. The reason is that there is a part that depends on the subjectivity of the user, and that designers / developers and product planners (hereinafter referred to as “designers / developers”) cannot collect information about customer dissatisfaction completely. . Several inventions have been made to improve such usability. Before collecting dissatisfaction in the first place, when a software failure occurs, a method of collecting failure information and transmitting the information to another device or referring to it from the outside has been studied. According to Patent Document 1, in a system for collecting logs when a failure occurs, when there are a plurality of log recording destinations, recording is performed without deviation of recording, and an interface that can refer to information from the outside is provided at the plurality of recording destinations. In this case, a technique is proposed in which information can be collected even if a failure occurs in any of the log recording functions. Next, for usability improvement, as shown in Patent Document 2, a usability improvement system that collects user's operation contents and compares it with the correct operation method to narrow down design elements that have deteriorated operability is proposed. Has been.

Japanese Patent Laid-Open No. 9-259013 JP 2004-287716 A

In the above-described prior art, it is possible to collect only information under conditions that can be assumed by the designer / developer, such as when a failure occurs or there is a mistake in the operation method. However, usability has a part that depends on the subjectivity of the user, and the embedded system user may be dissatisfied with the point that the designer / developer does not intend.
Therefore, it is generally difficult for the designer / developer to understand which part of the target embedded system the user using the embedded system is dissatisfied with. There is a problem that it is difficult to provide an easy-to-use system that does not depend on the user's subjectivity.

  In the present invention, various sensors are attached to the corresponding embedded system, and various actions when the user feels dissatisfied are detected based on the information. Using the detection result on the left as a trigger, the system behavior information at the timing when the action is taken is sent to the development environment accessible to the design / developer. The designer / developer reads the behavior information from an arbitrary design tool at an arbitrary timing. The design tool allocates where in the design the dissatisfaction indicated by the behavior information corresponds. Finally, the results are displayed on the screen, and the parts of the system that have problems are clearly shown to the designer / developer to clarify where the user is dissatisfied with the embedded system.

  The features of the present invention are (1) a design / development support system for embedded systems. When the embedded system is used after shipment, various behaviors of the user are shown when the user shows dissatisfaction with the processing. Detects from sensor information, collects system behavior information at the detected timing, notifies the host environment, records it in a recording device on the host environment, and the designer / developer shows dissatisfaction at any time It is to display information on functions that were operating in the host environment design / development tool.

  (2) In (1), based on information input from several sensor groups, whether the user is dissatisfied with the function being implemented is automatically identified and fed back to the designer / developer It is desirable.

  (3) In (1) and (2), by performing additional learning automatically based on information when the information input from the sensor group is not determined to be abnormal, high identification accuracy is matched to user characteristics You can also detect user dissatisfaction.

  (4) In (1), (2) or (3), even in a network-independent environment, user dissatisfaction can be fed back to the designer / developer via the recording medium.

  (5) In (1), (2) or (3), it can be realized by a single casing.

  (6) In (1), it is possible to provide a function capable of directly transmitting user dissatisfaction to the design / developer, and when dissatisfied, the information may be directly transmitted to the design / developer.

  According to the present invention, it is possible to associate a location where many users are potentially dissatisfied with design position information. As a result, it is possible to provide an easy-to-use system that does not depend on the subjectivity of the designer / developer or some loud users.

The figure which shows an example of the whole structure of embodiment of this invention. The figure which shows an example of the data structure at the time of preserve | saving the behavior information of a system. The figure which shows an example of the abnormality detection apparatus 005. The figure which shows an example of the process sequence of an abnormality detection process. The figure which shows an example of the process sequence which learns a user's abnormal behavior. The figure which shows an example of a function required for extending the design / development tool of a normal embedded system in implementing this invention. The figure which shows an example of the data structure of analysis information. The figure which shows an example of the process for reflecting a behavior information collection result in a design / development tool.

  First, a first embodiment of the present invention will be described.

  FIG. 1 is an explanatory diagram of an example of a function improvement system according to the present invention. The function improvement system according to the present invention includes a target system 1 that is a function improvement target and a host environment 2 that performs design and development of the target system to be improved.

  The target system 1 provides a buffer memory 001 for recording the behavior information of the target system, the control of the system, and various functions of the target system, and an investigation target program 002 to be improved by this function improvement system. Collecting behavior information of the program to be investigated 002 and recording it in the buffer memory 001, sensing the user's dissatisfaction, detecting an abnormality that is different from the usual behavior, and determining the abnormality Thus, the tracer 003 and the profiler 004 collect the behavior information of the investigation target program 002 recorded in the buffer memory 001 via the tracer 003 and the profiler 004, and deliver them to the communication module 006 for transmission to the host environment 2. Anomaly detection 005 and the communication module 006 that transmits the behavior information acquired from the abnormality detection device 005 to the communication module 011 of the host environment 2, receives the tracer 012 created in the host environment, and updates the tracer 003. And a recording device 007 for storing a learning result used when an abnormality is detected.

  The host environment 2 records in the recording device 010 the design / development tool 008 used when designing / developing the investigation target program of the target system and the behavior information of the investigation target program 002 recorded in the recording device 010. Using the behavior information and the design / development information, the data analysis function 009 for displaying the design / development tool superimposed on the corresponding function information, the behavior information of the investigation target program 002, the design / development information, The development tool 008 is used to communicate with the recording device 010 for recording design / development information generated when the investigation target program 002 is designed and developed, and the communication module 006 of the target system 1, and the investigation target To acquire behavior information of the program 002 and to send the newly created tracer 012 of the investigation target program A communication module 011, in advancing analysis of survey program 002 consists of tracer 012 updates created, or the recording spot new to change the traced.

Buffer memory data structure:
FIG. 2 is a diagram illustrating an example of a data structure of the buffer memory 001 for temporarily or permanently storing system behavior information collected by the tracer 003 and the profiler 004. The buffer memory follows the contents of the acquisition information of the behavior information collection function such as the tracer 003 and the profiler 004 included in the target system 1. FIG. 2 shows an example of the data structure of the buffer memory for the tracer 003. The sequence number (001-001), event name (001-002), time (001-003), content 1 (001-004), content 2 (001-005). The sequence number (001-001) is a sequence number of event information acquired by the tracer. When the resource is finite, wrap around with a predetermined threshold. The event name (001-002) is the name of the event to be acquired. For example, contents of process switching, interruption, memory allocation / release, timer activation, and the like are recorded. Although the application-specific event name is not written here, such contents may be recorded. When the object to be developed is a digital TV, examples include channel switching and volume adjustment. Time (001-003) is the time when the event name (001-002) occurs. Content 1 (001-004), content 2 (001-005). . As the contents, etc., auxiliary information necessary for analyzing the event name (001-002) is recorded. For example, in the event of process switching of sequence number 1, it is not only to know that process switching has occurred, but to know which process from which process has switched to which process. However, it is possible to accurately grasp the behavior of the system. It is supplementary information that supports behavioral analysis that is insufficient only with event names. The profiler 004 also records profiled information in the same manner, but is omitted here.

Example of an anomaly detector:
Next, details of the abnormality detection apparatus 005 shown in the target system 1 of FIG. 1 will be described with reference to FIG. The abnormality detection device 005 is a device for detecting dissatisfaction when the user uses the embedded system based on sensing data obtained from various sensors. The anomaly detection device 005 performs various learning / identification processes for determining the normal state and abnormal state of the user from various sensor groups 5-001 for sensing the user's action and the sensed information. In addition, a format conversion device 5-002 for performing various pre-processing such as A / D conversion and sampling of data, the data converted in format, and the normal state of the user are stored in the learning device 5-006. Using the learning result 5-004, which is the result of learning using, it is determined whether or not the user is dissatisfied with the embedded system sequentially, and it is determined that there is dissatisfaction (not normal… abnormal) The information collected by the behavior information collection tool is transferred to the communication module 006 via the tracer 003 and the profiler 004, and the communication mode of the host environment 2 is transferred. When the communication module 006 is requested to transmit to the module 011 and it is determined that there is no dissatisfaction, the normal state of the user of the target system 1 is detected using the sensing data stored in the buffer memory 5-005 at that time. To determine whether or not the data converted by the format conversion device 5-002 and the identification device 5-003 that makes a processing request to the learning device 5-006 can be used as learning data. The information stored in the buffer memory 5-005 is triggered by the buffer memory 5-005 for temporarily storing information until the determination of -003 and the processing request issued from the identification device 5-003. Using a machine learning algorithm or the like, the normal state is learned, and the learning result 5-004 used by the identification device and the recording device 007 are stored. Consisting learning device 5-006 for updating the learning result that is.

Anomaly detection device processing sequence:
FIG. 4 is a flowchart illustrating an example of a processing sequence of the abnormality detection processing. The abnormality determination process starts when the system is started. After starting the processing, the process proceeds to step S-01, where the past learning result stored in the recording device 007 is read out, read out as a learning result 5-004 on the working memory, and the process proceeds to step S-02. In step S-02, it is determined whether an abnormality determination request has been made, the memory is scanned, and processing is stopped. If the process is to be stopped, the process is terminated as it is. If not, the process proceeds to step S-03. In step S-03, the degree of abnormality is calculated based on the learning result 5-004 and the sensor information acquired from 5-002. In general, the degree of abnormality calculation process uses a score of a machine learning identification algorithm. As an example, there is a technique called clustering. Based on normal behavior information, the behavior information is classified into classes called classes in the vector space, and the distance between the center of gravity and representative vectors of those classes and the input vector data is obtained. Treated as a degree of abnormality. When such calculation of the degree of abnormality ends, the process proceeds to step S-04. In step S-04 abnormality level comparison, the threshold value recorded in the learning result 5-004 is compared with the abnormality level obtained in step S-03. If it compares, it will progress to step S-05. In step S-05, the branch of the process after the following step is implemented based on the comparison result calculated | required by step S-04. If the degree of abnormality does not exceed the threshold, it is determined that the user is in a normal state, and the process proceeds to step S-08 to perform additional learning of the normal state. If the threshold value is exceeded, the process proceeds to step S-06 to collect information on the function of the investigation target program 002 being processed at that time. In step S-06, the behavior information recorded in the buffer memory 001 by the tracer 003 and the profiler 004 is referred to the API provided by the tracer 003 and the profiler 004, the descriptor for accessing the recording area, the address information, and the like. And proceed to Step S-07. In step S-07, based on the collected information, the communication module 006 is requested to send the behavior information collected in step S-06 to the communication module 011 in the host environment 2. Return to step S-02. In step S-08, since it is determined to be normal from the value of the degree of abnormality calculated in step S-03 at that time, learning is performed on the learning device 5-006 to perform additional learning of the normal state. The process is requested so as to proceed to step S-02.

  FIG. 5 is a flowchart illustrating an example of a processing sequence of learning processing performed in the learning device 5-006, which is a part of the abnormality detection device 005. This process is performed when a learning request is received from the identification device 5-003. When the processing request is received, the learning process is started, and the process proceeds to step S-08. In step S-08, the learning result accumulated in the buffer memory 5-005 is read, and the process proceeds to step S-09. In step S-09, learning is performed based on the data read in step S-08. In general, learning is performed using a learning algorithm used in machine learning or the like. As an example, there is a technique called clustering. Based on the acquired behavior information at normal time, the behavior information is classified into classes called classes in the vector space, and the center of gravity and representative vectors of those classes are obtained. When the learning is finished, the process proceeds to step S-10. In step S-10, based on the learning result obtained in step S-09, the learning result 5-004 which is a memory for storing the learning result and the learning result stored in the recording device 007 are updated. After completion of the update, the process proceeds to step S-11. In step S-11, if the remaining number of learning target data is not checked, the process ends. If there is, the process returns to step S-08.

  To add to the design / development tool of the normal embedded system in the design / development tool 008 and the recording device 010 in FIG. 6 and reflect in the design / development information which function the embedded system user was dissatisfied with An example of a function added to a normal embedded system design / development tool in order to create the definition information is shown. There are a wide variety of design and development tools for ordinary embedded systems. For example, even when designing a system, a modeling language called UML may be used to describe using a modeling tool that supports UML model description, or a flowchart may be created using a document file creation program. May be described. In this way, there are a wide variety of design methods and design tools. Similarly, the development tools are diverse. There are quite a few programming languages alone, and the number of compilers and associated analysis tools that generate executable binaries from that language is even greater than the number of languages. Therefore, here, an example of a part to be added to a normal embedded system design / development tool that is necessary instead of the design environment and the development environment itself is shown. The design / development tool 008 includes design information 010-001 designed by the design tool and stored in the recording device 010, and a creation program 010-002 created by the development tool and stored in the recording device 010. A list of position information corresponding to the processing event is extracted and stored as recording device 010 analysis information 010-003, analysis information 010-003 analyzed and stored by the object analysis tool, and trace Trace location registration tool for extracting hooks and trace statements inserted for tracing from definition information 010-004 created by the user, in which label names such as processing function names are registered, and updating analysis information 010-003 008-002 and the design information 010-00 described above , Creation program 010-002, analysis information 010-003, and definition information 010-004, and them from recording to keep recording apparatus 010.

  FIG. 7 is a diagram illustrating an example of the data structure of the analysis information 010-003 stored in the recording device 010. The analysis information 010-003 varies depending on the tool used for design, the design method, the compiler used for creating the program, the tool chain, and the like. The data structure of the analysis information 010-003 includes a management ID (010-003-001), a label name (010-003-002), an event name (010-003-003), an address (010-003-004), and a problem. It consists of locations (010-003-005). The management ID (010-003-001) is an ID for managing the destination from which the trace is acquired. The label name (010-003-002) is a function name indicating a position where a trace is acquired, a function name and its offset, and the like. The event name (010-003-003) is a name indicating the contents of the trace or profile. The address (010-003-004) is an address corresponding to the label name from which the trace is acquired, that is, an address for taking the trace. The problem location (010-003-005) is an area in which the system writes a result indicating which function the embedded system user is dissatisfied with in the trace result reflection process described later.

  FIG. 8 shows related processing for updating the analysis information 010-003 by allocating which function of the embedded system was acquired from the acquired behavior information when the behavior information was acquired. The normal embedded system design / development tool 008 uses the design information 010-001, the creation program 010-002, and the analysis information 010-003, and the label name and event name recorded in the design information. An analysis result overlay process 008-003 for allocating the label name event name described as a problem in the problem location (010-003-005) of the analysis information 010-003 and overlaying the analysis result on the design / development information To add. Superimposition of analysis results can also be realized by changing the display method of the corresponding function. Moreover, you may implement | achieve by the method of superimposing a pointer etc. on the applicable function. Next, the data analysis function 009 will be described. The data analysis function 009 combines the following three processes to obtain a problem location (010-003-005) in the analysis information 010-003 shown in FIG. 7 from the acquired behavior information 010-005 and analysis information 010-003. Realize the update. Read behavior information 010-005 recorded in the recording device 010, update the work memory 009-002, and request behavior processing to the name resolution processing 009-003, behavior information reading processing 009-001, and behavior information reading processing 009- From 001, the data is received and the data is temporarily recorded in the data structure as shown in FIG. 2, and at the same time, the analysis information 010-003 is temporarily stored in the data structure shown in FIG. 7 during the name resolution process 009-003. Received by the name resolution process 009-003 and used as a temporary storage when updating the determination result of whether or not there is a problem location, the work memory 009-002 and the behavior information reading process 009-001 are received. The analysis information 010-003 is read from the recording device 010 onto the work memory, and is expanded on the work memory. In order from the top of the behavior information shown in FIG. 2, matching is performed to determine whether there is an event name that matches the analysis information 010-003 shown in FIG. 7. If there is matching, the problem location (010-003 in the analysis information 010-003 is found. -005) is updated, and the process is completed for all the behavior information 010-005, and the design information 010-001 recorded in the recording device 010 is updated for the design information update process 090-004. Based on the analysis information updated on the work memory in response to processing requests from the requested name resolution processing 009-003 and the name resolution processing 009-003, the labels and events of the corresponding functions in the design information 010-001 It consists of design information update processing 009-004 for adding / updating additional information to the name.

  In the first embodiment, the host environment and the target system are connected via the network. However, the host environment and the target system may not be connected via the network. For example, in this case, the communication modules 006 and 011 are realized by a recording device that can be inserted and removed. Usually, a behavioral information when a user using the target system 1 is dissatisfied by connecting a detachable recording device represented by an SD card or the like to the communication module 006 on the target system 1 side. Etc. are recorded. Then, when the developer or the user of the target system 1 desires, it is removed from the target system 1 and inserted into the host environment, and analysis is performed with the same contents as in the first embodiment. An example is shown in the tracer 012. When updating a tool for collecting behavior information such as a tracer, it is recorded in a recording device that can be inserted / removed from the host information at the timing to be updated, and the recorded recording device is inserted into the target system 1. Then, a tool update process for collecting behavior information is performed.

  In the first and second embodiments, the target system 1 and the host environment 2 are separate environments, but they may be contained in the same environment (enclosure). In that case, the exchange via the communication module is not carried out, and the exchange between the devices and functions is carried out directly.

  In the first embodiment, instead of automatically identifying sensor group information, the user may be provided with a function that can directly present dissatisfaction, and when dissatisfied, the intention may be displayed directly to the designer. .

  Locations that many users are potentially dissatisfied can be associated with location information in the design, and which part of the target embedded system the user using the embedded system is dissatisfied with It becomes easy for the designer / developer to grasp, and it becomes possible to provide an easy-to-use system that does not depend on the subjectivity of the designer / developer and some loud users.

DESCRIPTION OF SYMBOLS 1 ... Target system, 2 ... Host environment, 001 ... Buffer memory, 002 ... Investigation target program, 003 ... Tracer, 004 ... Profiler, 005 ... Abnormality detection device, 006 ... Communication module, 007 ... Recording device, 008 ... Design and development Tool, 009, data analysis function, 010, recording device, 011, communication module, 012, tracer.

Claims (6)

  An embedded system design / development support system that detects user behavior from information from various sensors when a user shows behavior that is unsatisfactory during processing of the embedded system after shipment. System behavior information is collected at the timing, notified to the host environment, recorded in a recording device on the host environment, and information on the function that was operating when the user indicated dissatisfaction at any time by the designer / developer. A design / development support system characterized by displaying it on the design / development tool of the host environment.   In the design / development support system according to claim 1, based on information input from several sensor groups, whether or not the user is dissatisfied with the function being executed is automatically identified, A design / development support system that features feedback to developers.   In the design / development support system according to claim 1, by performing additional learning automatically based on information when the information input from the sensor group is not determined to be abnormal, according to the characteristics of the user, A design / development support system that detects user dissatisfaction with high identification accuracy.   The design / development support system according to any one of claims 1, 2, and 3, wherein user dissatisfaction can be fed back to the design / developer through a recording medium even in a network-independent environment. Design and development support system.   4. The design / development support system according to claim 1, wherein the design / development support system is realized by a single housing.   The design / development support system according to claim 1 has a function capable of directly transmitting a user's dissatisfaction to the design / developer, and transmitting information to the design / developer directly when dissatisfied. Characteristic design and development support system.

Images