JP2009003686A - Apparatus, method and program for analyzing software operation - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide software operation analysis apparatus, method and program which can improve detection accuracy in the occurrence of software events in analyzing the operation of software having at least one and more tasks. <P>SOLUTION: Each of software operation analysis apparatuses 1, 1A is provided with: a task operation state characteristic value detection means 110 for detecting a task operation state characteristic value expressing the characteristic of an operation state of each task; a task operation state characteristic value storage means 140 for storing the detected task operation state characteristic value; a data processing means 120 for executing data processing for extracting the characteristic of the task operation state characteristic value for the stored task operation state characteristic values; and an event occurrence detection means 130 for setting a prescribed range in accordance with the characteristic obtained from a data processing result and detecting the occurrence of a software event under a condition that the task operation state characteristic value detected by the task operation state characteristic value detection means 110 is not included in the set prescribed range. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a software operation analysis apparatus, a software operation analysis method, and a software operation analysis program for analyzing the operation of software having at least one task.

  Conventionally, in order to verify software installed in a computer, an apparatus that analyzes data representing an operation state has been used. At this time, if the scale of the software increases, the amount of data to be analyzed becomes enormous and the operation state becomes complicated, so that it becomes difficult to verify the software. For this reason, a technique has been proposed for efficiently and accurately performing software verification by detecting and using an event (a specific operation different from a normal operation) that has a high possibility of a software malfunction (for example, Patent Documents 1 to 3).

  In the vehicle data recorder of Patent Document 1, when recording data for analyzing the operation of the vehicle control software, a signal indicating an abnormality from an acceleration sensor, an airbag ECU, a bumper switch, or the like mounted on the vehicle is used. Based on this, an emergency state of the vehicle is detected, and vehicle travel and operation information before and after the occurrence of the detected emergency state is recorded as data.

  In the observation apparatus of Patent Document 2, when a signal output from a logic circuit that is an observation target apparatus is input and its operation is observed, the state transition of the output signal is stored, and the number of occurrences of the state transition is determined in advance. When it is less than the specified threshold value or when the state transition matches a state transition designated in advance by the user, it is notified that the state transition has occurred as an operation with a high possibility of occurrence of a failure.

Moreover, in the monitoring apparatus of patent document 3, when monitoring the abnormal operation of the periodic software of the computer system, the processing time required for one cycle travel is calculated, and this processing time deviates from a predetermined specified range. An abnormal operation is detected when the software becomes abnormal.
JP-A-7-37133 JP 2006-113696 A JP-A-9-54612

  However, in the apparatus of Document 1, when an event (emergency state of a vehicle) is detected, a signal indicating an abnormality from an acceleration sensor or the like mounted on the vehicle is used. The signals from these acceleration sensors and the like are data that are directly used for controlling the vehicle that is controlled by software, and are not data that directly represents the operating state of the software. Therefore, since the apparatus of the above-mentioned literature 1 detects an event based on data used for such control, there is a possibility that a software event cannot be detected with high accuracy.

  Further, in the apparatus of the above-mentioned document 2, when detecting an event (an operation that is highly likely to cause a failure), the state transition of the output signal of the logic circuit that is the observation target apparatus is used. The state transition of the output signal is data that directly represents the operating state of the software (logic circuit), but is not necessarily a value that represents the characteristics of the operating state of the software suitable for software verification. Therefore, since the apparatus of the above-mentioned document 2 detects an event using a value that directly represents the operation state of the software, there is a possibility that the software event cannot be detected with high accuracy.

  On the other hand, in the apparatus of the above-mentioned document 3, when an event (abnormal operation) is detected, the processing time required for one cycle of the periodic software to be analyzed is calculated and used. It is considered that the value represents the characteristic of the operating state of the periodic software. However, the apparatus of the above-mentioned document 3 detects an event by comparing the processing time with a predetermined determination value. On the other hand, the processing time of software greatly depends on its operating state. For example, it is assumed that the processing time of a specific task of the vehicle control software varies depending on the vehicle control state by software, such that the processing time of the vehicle is long when the vehicle is traveling at high speed and short when the vehicle is in an idle stop state.

  For this reason, it is difficult to appropriately detect an event when compared with a determination value that is a mere fixed value as in the apparatus of Document 3 above. In other words, if the judgment value is set outside the range of values that can be taken in all states to prevent false detection of the event, it cannot be detected unless the characteristic value fluctuates outside that range. A situation that cannot be detected can occur. In addition, when setting the judgment value outside the normal value range so that no event is not detected, if the characteristic value fluctuates outside the normal range according to the control status, However, there was a problem that a situation where an event was detected could occur.

  In view of the above circumstances, the present invention provides a software operation analysis apparatus, a software operation analysis method, and a software operation analysis method capable of improving the detection accuracy of event occurrence of software when analyzing the operation of software having at least one task. It is another object of the present invention to provide a software operation analysis program.

  In order to achieve such an object, a software operation analysis apparatus according to the present invention is a software operation analysis apparatus that analyzes the operation of software having at least one task. Task operation state characteristic value detection means for detecting a value, task operation state characteristic value storage means for storing the task operation state characteristic value detected by the task operation state characteristic value detection means, and task operation state characteristic value storage means The data processing means for performing data processing for extracting the characteristics of the task operation state characteristic value for the task operation state characteristic value stored in step (b), and the data processing result of the data processing means from the data processing result A predetermined range is set according to the obtained characteristics, and the task detected by the task operation state characteristic value detecting means is detected. And an event occurrence detecting means for detecting that an event of the software is generated on condition that the operating state characteristic value is not included in the set predetermined range (first operation) invention).

  According to the software operation analysis apparatus of the first invention, the task operation state characteristic value representing the operation state characteristic of each task of the software is detected in real time by the task operation state characteristic value detecting means. Here, the task operation state characteristic value is a characteristic value representative of the software operation state, which is different from, for example, data directly used for control by the software or data directly representing the operation state of the software. That is, as the task operation state characteristic value, data representing the operation state characteristic of the software, which is different from the data acquired directly from the software, is obtained. Then, the detected task operation state characteristic value is stored in the task state value storage means, and data processing for extracting the characteristics of the task operation state characteristic value by the data processing means is performed on the stored task operation state characteristic value. Is called. As a result, data indicating the characteristics of the operating state characteristic value of the software is obtained as the data processing result.

  Therefore, the event occurrence detection means sets a predetermined range for the data processing result of the data processing means according to the characteristics obtained from the data processing result. The event occurrence detection means detects that a software event has occurred on condition that the task operation state characteristic value is not included in the predetermined range. Here, the software event is a unique operation different from a normal operation in software, and there is a high possibility of a software malfunction. At this time, by using the data processing result indicating the characteristic of the operation state characteristic value of the software and setting the predetermined range according to the characteristic obtained from the data processing result, the predetermined range corresponding to the normal operation of the software is appropriately set. Can be set. Therefore, it is possible to detect the event with high accuracy on condition that the task operation state characteristic value is not included in the predetermined range. By using the event detected in this way, it is possible to perform software verification efficiently and accurately.

  Therefore, according to the present invention, when analyzing the operation of software having at least one task, it is possible to improve the detection accuracy of the occurrence of the software event.

  In the software operation analysis apparatus according to the first aspect of the present invention, the software operation analysis device further includes a task type determination unit that determines a task type of the task operation state characteristic value detected by the operation state value detection unit, and the data processing unit includes the task operation The task processing state characteristic value stored by the state characteristic value storage means is subjected to the data processing for each type of task, and the event occurrence detection means has a type of task determined by the task type determination means. It is preferable to perform a process of detecting the occurrence of an event on the data processing result (second invention).

  In this case, by performing data processing for each task type on the task operation state characteristic value, the data processing result shows the characteristic of the operation state characteristic value more prominently for each task type of software. It becomes. By using this data processing result, a predetermined range corresponding to a normal operation can be set more appropriately for each type of task of software. Therefore, it is possible to detect an event for each task type with higher accuracy on the condition that it is not included in the predetermined range. Therefore, by detecting the task type of the task operation state characteristic value and performing the process of detecting the event occurrence on the data processing result of the task type, the detection accuracy of the event occurrence can be further improved. it can.

  In the software operation analysis device according to the first or second aspect of the invention, the data processing means may be configured such that the number of task operation state characteristic values stored for each task type by the task operation state characteristic value storage means is the task operation state. When there is a task exceeding the predetermined number from which the characteristic value characteristic is extracted, or a predetermined time for extracting the task operational state characteristic value characteristic from the start of storing the task operational state characteristic value When there is a task that has passed, it is preferable to perform the data processing on the operation state characteristic value of the task (third invention).

  That is, in data processing such as statistical processing, for example, by taking a sufficiently large amount of data to be processed, the influence of noise included in the data is eliminated, and data trends are appropriately extracted. Therefore, when the number of task operation state characteristic values exceeds a predetermined number from which the characteristics of the task operation state characteristic values are extracted, or the task operation state characteristic values are By performing data processing when a predetermined time elapses when characteristics are extracted, it is possible to obtain a large amount of data for task operation state characteristic values to be processed. The data processing result indicating the characteristics of the operating state characteristic value of the software can be obtained more stably. And the detection accuracy of the occurrence of an event can be further improved by performing the process for detecting the occurrence of the event on the data processing result.

  In the software operation analysis apparatus according to the first to third inventions, parameter value acquisition means for acquiring a parameter value used for the operation of the software, parameter value acquired by the parameter value acquisition means, and the task operation Either or both of the task operation state characteristic values detected by the state characteristic value detection means are sequentially collected as the internal data of the software at a predetermined timing, and the internal data for the predetermined time period or the predetermined number of times of collection. Internal data storage means for storing data, and internal data storage means for storing internal data stored in the internal data storage means on the condition that an event has been detected by the event occurrence detection means (4th invention).

  In this case, the internal data storage means sequentially collects one or both of the parameter value used for the operation of the software and the task operation state characteristic value at a predetermined timing as internal data of the software. Here, the parameter value used for the operation of the software is, for example, data directly used by the software for control, data directly representing the operation state of the software, or the like. The internal data storage means stores internal data for a predetermined time immediately before or a predetermined number of collection times. Specifically, the internal data storage means erases data before a predetermined time (or a predetermined number of collection times) stored in the memory, for example, and stores the newly collected data in the erased portion, thereby allowing the internal data Store data while rewriting data sequentially. As a result, the internal data of the software for the predetermined time (predetermined number of times of collection) is always stored with a constant memory amount.

  The internal data stored in the internal data storage means is stored in the internal data storage means on condition that the event occurrence detection means detects that an event has occurred. That is, for example, when the occurrence of an event is detected or a predetermined time after the occurrence of an event is detected, the internal data is stored. As a result, the internal data before and after the event occurrence is appropriately saved with the occurrence of the event as a trigger, so the amount of data to be saved can be reduced and the processing load for saving can be reduced. By using it when verifying software, it becomes possible to verify software efficiently and accurately.

  In the software operation analysis apparatus according to any one of the first to fourth inventions, the data processing means is configured to store the task operation state characteristic values stored for each task type by the task operation state characteristic value storage means at predetermined time intervals. The task operation state characteristic value for the predetermined time or the task operation state characteristic value for the predetermined number of times each time the stored number of the task operation state characteristic value becomes the predetermined number It is preferable to carry out the treatment (fifth invention).

  That is, it is assumed that the task operation state characteristic value characteristic varies depending on the operation state of the software. At this time, the task operation state characteristic value for the predetermined time for the predetermined time interval or the task operation state characteristic value for the predetermined number every time the number of stored task operation state characteristic values becomes the predetermined number By performing data processing on the target, a data processing result reflecting the operation state of the software at that time can be obtained. Therefore, by setting the predetermined range according to the characteristics obtained from the data processing result using the data processing result, the predetermined range can be set to follow the operating state of the software at that time. In other words, the range of task operation state characteristic values (the range of task operation state characteristic values corresponding to the normal operation of software) in which the occurrence of an event is not detected is set according to the software operation state. Even when a plurality of operation states are taken, it is possible to detect an event with high accuracy on condition that the task operation state characteristic value is not included in the predetermined range.

  Next, a software operation analysis method of the present invention is a software operation analysis method for analyzing the operation of software having at least one task, and detects a task operation state characteristic value representing the operation state characteristic of each task. A task operation state characteristic value detecting step, a task operation state characteristic value storage step for storing the task operation state characteristic value detected in the task operation state characteristic value detection step in a task operation state characteristic value storage unit, and the task operation With respect to the task operation state characteristic value stored in the state characteristic value storage step, a data processing step for performing data processing for extracting the characteristic of the task operation state characteristic value, and a data processing result of the data processing step, A predetermined range is set according to the characteristic obtained from the data processing result, and the task operation state characteristic value is detected. An event occurrence detecting step for detecting occurrence of the software event on condition that the task operation state characteristic value detected in the step is not included in the set predetermined range. 6 invention).

  According to the software operation analysis method of the sixth invention, as described with respect to the software operation analysis device of the first invention, the task operation state characteristic representing the operation state characteristic of each task of the software in the task operation state characteristic value detection step. The value is detected in real time. Further, in the data processing step, a data processing result that significantly shows the characteristics of the software operation state characteristic value is obtained using the task operation state characteristic value. Then, in the event occurrence detection step, a predetermined range corresponding to the normal operation of the software can be appropriately set using the data processing result, so that the event is accurately detected on condition that it is not included in this predetermined range. It becomes possible to do. Therefore, according to the present invention, when analyzing the operation of software having at least one task, it is possible to improve the detection accuracy of the occurrence of the software event.

  Next, the task software operation analysis program of the present invention is a software operation analysis program for causing a computer to execute a software operation analysis process for analyzing the operation of software having at least one task. Task operation state characteristic value detection processing for detecting task operation state characteristic values representing state characteristics, and task operation state characteristic values detected in the task operation state characteristic value detection processing are stored in the task operation state characteristic value storage unit A task operation state characteristic value storage process; a data process for extracting a characteristic of the task operation state characteristic value for the task operation state characteristic value stored in the task operation state characteristic value storage process; and For the data processing result, set a predetermined range according to the characteristics obtained from the data processing result, The event occurrence detection process for detecting the occurrence of the software event on condition that the task operation state characteristic value detected in the task action state characteristic value detection process is not included in the set predetermined range. A function to be executed by a computer is provided (seventh invention).

  According to the software operation analysis program of the seventh aspect of the invention, it is possible to cause a computer to execute a process that can achieve the effects described with respect to the software operation analysis apparatus of the first aspect of the invention.

[First Embodiment]
A first embodiment of the software operation analysis apparatus of the present invention will be described with reference to FIGS. 1 is a functional block diagram of the software operation analysis apparatus of the present embodiment, FIG. 2 is a flowchart showing software operation analysis processing in the software operation analysis apparatus of FIG. 1, and FIG. 3 is a software operation of FIG. It is a graph which shows the example of detection of the event generation | occurrence | production in an analysis process.

  As shown in FIG. 1, the software operation analysis apparatus 1 of the present embodiment is mounted on a vehicle 2 and incorporated in an electronic control unit (ECU) 3 that controls the operation of the vehicle 2. The software operation analysis apparatus 1 analyzes the operation of the software by using control software having tasks 1 to k installed in the ECU 3 as an analysis target. In order to control the operation of the vehicle 2, sensors 4-1 to 4 -m for detecting the state of the vehicle 2 and control devices 5-1 to 5 -n controlled by the ECU 3 are connected to the ECU 3. .

  The software operation analysis apparatus 1 is composed of a computer (comprising a CPU, ROM, RAM, I / O, etc.). The computer that constitutes the software operation analysis apparatus may be a computer that constitutes the ECU 3 or may be another computer that is communicably connected to the ECU 3.

  The software operation analysis apparatus 1 includes a measurement data input unit 10, a characteristic value calculation unit 110, a characteristic value data processing unit 120, an event determination unit 130, and a reader 200 as its functions. Each of these functions may be constituted by one computer or may be constituted by a plurality of independent computers.

  Each of these functions is realized by a computer executing a program installed in advance in a computer memory. This program includes the software operation analysis program of the present invention. The program may be stored in the memory via a recording medium such as a CD-ROM. Alternatively, the program may be distributed or broadcast from an external server via a network or an artificial satellite, received by a communication device mounted on the vehicle 2 or the like, and stored in a memory.

  The measurement data input unit 10 inputs measurement data from the control software of the ECU 3 to be analyzed. The measurement data corresponds to parameter values used for the operation of the software. For example, the software controls the detection data of the sensors 4-1 to 4-m and the control data of the control devices 5-1 to 5-n. Software data that directly represents the operating state of software, such as data to be used directly and task switching signals.

  The characteristic value calculation unit 110 detects the task operation state characteristic value representing the operation state of each task 1 to k in software in real time from the measurement data input to the measurement data input unit 10. Here, the task operation state characteristic value representing the operation state of each of the tasks 1 to k corresponds to a characteristic value representing the operation of the software different from the measurement data directly acquired from the software.

  The characteristic value data processing unit 120 performs data processing on the characteristic value calculated by the characteristic value calculation unit 110. Specifically, the characteristic value data processing unit 120 stores the characteristic value calculated by the characteristic value calculating unit 110 and performs data processing for extracting the characteristic of the characteristic value for the stored characteristic value. At this time, the characteristic value data processing unit 120 performs data processing on the characteristic values for the predetermined time at predetermined time intervals among the stored characteristic values.

  The event determination unit 130 performs event occurrence detection processing for determining whether an event has occurred during the operation of the software, using a part or all of the data processing result by the characteristic value data processing unit 120. At this time, the event determination unit 130 sets a predetermined range for the data processing result of the characteristic value data processing unit 120 according to the characteristic obtained from the data processing result. Then, the event determination unit 130 determines that a software event has occurred on the condition that the characteristic value calculated by the characteristic value calculation unit 110 is not included in the set predetermined range.

  Data relating to the event occurrence determined by the event determination unit 130 is read from the PC 6 via the reader 200 and used when the control software of the ECU 3 in the PC 6 is verified.

  The measurement data input unit 10 corresponds to the parameter value acquisition unit of the present invention, the characteristic value calculation unit 110 corresponds to the task operation state characteristic value detection unit of the present invention, and the characteristic value data processing unit 120 The task operation state characteristic value storage means, task type determination means, and data processing means correspond to the invention, and the event determination unit 130 corresponds to the event occurrence detection means of the present invention.

  Next, the operation (software operation analysis process) of the software operation analysis apparatus 1 of the present embodiment will be described with reference to FIGS. As shown in FIG. 2, the software behavior analysis apparatus 1 repeats the processing of STEP1 to STEP4 and executes the software behavior analysis processing at every predetermined calculation processing cycle. In this embodiment, STEP1 to STEP4 are set as one cycle, but the present invention is not limited to this. For example, the processing of STEP1, the processing of STEP2 and STEP4, and the processing of STEP3 are executed at different cycles. May be.

  First, in STEP 1, the measurement data input unit 10 inputs measurement data from the control software of the ECU 3 to be analyzed. Specifically, for example, the measurement data input unit 10 inputs a task switching signal in software as measurement data.

  Next, in STEP2, the characteristic value calculation unit 110 calculates the characteristic value from the input measurement data in real time. Specifically, for example, the characteristic value calculation unit 110 calculates a processing cycle of a specific task from the task switching signal as the characteristic value. Thereby, data representing the characteristics of the operating state of the software, which is different from the measurement data obtained directly from the software, is obtained as the characteristic value.

  Next, in STEP 3, the characteristic value data processing unit 120 stores the calculated characteristic value and performs data processing on the stored characteristic value (characteristic value data processing). Specifically, the characteristic value data processing unit 120 calculates a histogram for characteristic values at intervals of unit time N (for example, 1 sec) among the stored characteristic values as characteristic value data processing. The unit time N is a value that is determined in advance as a sufficiently large value that allows the characteristic to be extracted by eliminating the influence of noise included in the characteristic value with respect to the stored characteristic value. Then, the characteristic value data processing unit 120 performs processing for calculating the average x and the variance σ of the characteristic values within the unit time N from this histogram. The average x and the variance σ, which are data processing results, indicate the characteristics of the operating state characteristic value of the software. In addition, by performing data processing on the characteristic value for each unit time N interval as described above, a data processing result reflecting the operation state of the software at that time (control state of the vehicle 2 by software) can be obtained.

  Here, an example of a histogram of characteristic values stored in the unit time N is shown in the graph of FIG. In FIG. 3, the horizontal axis indicates the processing cycle of a specific task, and the vertical axis indicates the frequency of occurrence. The data representing the processing cycle of a specific task stored in the unit time N is indicated by ♦ in FIG. At this time, as shown in FIG. 3, the average x and the variance σ within the unit time N of the processing cycle of the specific task are obtained as the data processing result.

  Next, in STEP 4, the event determination unit 130 determines whether an event has occurred during the operation of the software. Specifically, for example, the event determination unit 130 determines the event determination from x ± 3σ with a preset threshold 3σ with respect to the average x and variance σ of the processing cycle calculated by the characteristic value data processing unit 120. A predetermined range that is a threshold is set. Then, the event determination unit 130 determines that an event has occurred when the processing cycle calculated by the characteristic value calculation unit 110 is outside the predetermined range.

  As described above, by using the average x and the variance σ that remarkably indicate the characteristics of the operating state characteristic value of the software, a predetermined range corresponding to the normal operation of the software is appropriately set. Therefore, it is possible to detect an event with high accuracy on the condition that it is not included in the predetermined range. In addition, as described above, the event determination threshold is calculated based on the average x and the variance σ, which are the data processing results for the characteristic values for the unit time N, so that a predetermined event determination range (determination threshold) is obtained. It can be set to follow the operating state of the software at that time. Therefore, since the determination threshold is calculated by eliminating the influence of the fluctuation of the characteristic value according to the software operation state (control state of the vehicle 2 by software), even when a plurality of operation states are taken during the software operation, By using the determination threshold, it becomes possible to detect the event with high accuracy.

  Specifically, as shown in FIG. 3, predetermined ranges Ra to Rb for event determination are calculated from x ± 3σ using the average x and the variance σ. Here, the requirement time Rth in FIG. 3 is a value set as a software specification, and the range where the processing cycle is equal to or less than the requirement time Rth is a range where the specification is OK, and the processing cycle is the requirement time Rth. Exceeding range is a range that violates the specification. The requirement time Rth is set outside the range of values that can be taken in all states so as not to cause erroneous detection of events. When this requirement time Rth is used as a determination threshold, an event A that is separated from the data distribution of the processing cycle by one and has a high possibility of abnormal operation is not detected. On the other hand, since the predetermined ranges Ra to Rb are set to match the current operating state of the software, the event A is appropriately detected by using the predetermined ranges Ra to Rb as determination thresholds. be able to.

  Furthermore, when it is determined that an event has occurred, data relating to the occurrence of the event is stored in the event determination unit 130.

  The above is the software operation analysis processing in this embodiment. Data relating to the event occurrence determined by the event determination unit 130 in the software operation analysis process is read from the PC 6 via the reader 200 and used when the control software of the ECU 3 in the PC 6 is verified. Specifically, data related to the event occurrence is presented to the software designer, etc., and the designer examines the execution history of the software stored separately, for example, based on this presentation, thereby efficiently verifying the software and It becomes possible to carry out with high accuracy.

As described above, according to the present embodiment, it is possible to improve the detection accuracy of the occurrence of an event of the software when analyzing the operation of the software having at least one task.
[Second Embodiment]
Next, a software operation analysis apparatus 1A according to the second embodiment of the present invention will be described with reference to FIGS. Note that the present embodiment is different from the first embodiment in that the data temporary storage unit 140 and the data storage unit 150 are provided, and only the configuration relating to the data temporary storage unit 140 and the data storage unit 150 is different. In the following description, the same components as those in the first embodiment are denoted by the same reference numerals, and the description thereof is omitted.

  In the software operation analysis apparatus 1A of the present embodiment, the temporary data storage unit 140 sequentially collects and stores software internal data for a predetermined time or a predetermined number of times of collection. The internal data of the software includes measurement data such as a task switching signal input from the measurement data input unit 10 and characteristic values such as a processing cycle of a specific task calculated by the characteristic value calculation unit 110. At this time, since the amount of memory that can be implemented is limited, the data temporary storage unit 140 uses a function such as a ring buffer to erase, for example, data stored in the memory before a predetermined time (predetermined number of collection times). Then, by storing newly collected data in the erased portion, the internal data for the immediately preceding predetermined time (for the predetermined number of times of collection) is stored while the data is sequentially rewritten. Note that the method used when storing data in the temporary data storage unit 140 is not limited to the ring buffer method.

  If the event determination unit 130 determines that an event has occurred, the event determination unit 130 outputs an event signal to the data storage unit 150. Then, the data storage unit 150 saves the internal data stored in the temporary data storage unit 140 on the condition that the event determination unit 130 determines that an event has occurred. That is, the data storage unit 150 stores the data of the temporary data storage unit 140 according to the event signal output from the event determination unit 130. Specifically, the data storage unit 140 stores the data temporarily after a predetermined time has elapsed since the event occurrence determination unit 130 determined the occurrence of an event (when an event signal due to the occurrence of the event was output). The internal data (internal data before and after the occurrence of the event) stored in the billion part 140 is stored. And the data memorize | stored in the data memory | storage part 150 are read from PC6 via the reader 200, and are utilized for verification of the control software of ECU3 in PC6. Other configurations are the same as those of the first embodiment.

  The temporary data storage unit 140 corresponds to the internal data storage means of the present invention.

  Next, the operation (software operation analysis process) of the software operation analysis apparatus 1A of the present embodiment will be described with reference to FIG. As shown in FIG. 5, the software behavior analysis apparatus 1A repeats the processing of STEP11 to STEP17 for each predetermined calculation processing cycle to execute the software behavior analysis processing. In the following description, the same processing as in the first embodiment will be described with reference to the first embodiment. In this embodiment, STEP 11 to STEP 17 are set as one cycle. However, the present invention is not limited to this. For example, STEP 11 and 13 processing, STEP 12 and 15 to 17 processing, and STEP 14 processing are separately performed. It is also possible to execute at a cycle of

  First, in STEP 11, as in STEP 1 of the first embodiment, the measurement data input unit 10 inputs a task switching signal in software as measurement data from the control software of the ECU 3 to be analyzed.

  Next, in STEP 12, as in STEP 2 of the first embodiment, the characteristic value calculation unit 110 calculates the processing cycle of a specific task in real time as a characteristic value from the input task switching signal.

  Next, in STEP 13, the measurement data input in STEP 11 and the characteristic value calculated in STEP 12 are stored in the data temporary storage unit 140 as internal data of the software. The temporary data storage unit 140 stores internal data of software for a predetermined time (a predetermined number of times of collection) immediately before the time (discrete system time) of a certain calculation processing cycle while rewriting data sequentially. As a result, the internal data of the software for the predetermined time (predetermined number of times of collection) is always stored with a constant memory amount.

  Next, in STEP14, as in STEP3 of the first embodiment, the characteristic value data processing unit 120 stores the calculated characteristic value and performs data processing on the stored characteristic value (characteristic value data processing). ).

  Next, in STEP 15, as in STEP 4 of the first embodiment, the event determination unit 130 determines whether or not an event to be noticed during later analysis occurs during the operation of the software. If the determination result in STEP 15 is NO (when it is not determined that an event has occurred), the software analysis process is terminated.

  When the determination result in STEP 15 is YES (when it is determined that an event has occurred), the process proceeds to STEP 16 and the event occurrence determination unit 130 outputs an event signal to the data storage unit 150.

  Next, in STEP 17, the data storage unit 150 executes a process of storing the internal data of the software stored in the temporary storage unit 140 in accordance with the event signal. At this time, the data storage unit 150 executes storage processing for the event after a predetermined time has elapsed since the time when the event occurred (time when the event signal was output). As a result, the internal data of the software before and after the time when the event occurs is stored.

  The above is the software operation analysis processing in this embodiment. Then, the data stored in the data storage unit 150 in the software operation analysis process is read from the PC 6 via the reader 200 and used in the control software verification of the ECU 3 in the PC 6. At this time, since the history of the measurement data and the characteristic value before and after the event occurrence can be confirmed from the stored data, the software can be verified efficiently and accurately. Other operations are the same as those in the first embodiment.

  As described above, according to the present embodiment, as in the first embodiment, when analyzing the operation of software having at least one task, it is possible to improve the detection accuracy of the occurrence of the software event. .

  In the first and second embodiments, the processing cycle of a specific task is used as the characteristic value. However, as the characteristic value, the frequency of task occurrence, the amount of change in predetermined control data, and the writing to a predetermined memory are used. It is possible to use various characteristic values representing the characteristics of the operating state of the software, such as frequency.

  In the first and second embodiments, the characteristic value data processing includes processing for calculating the average and variance of the characteristic values. However, the characteristic value data processing includes the adjustment rate of the characteristic values and the amount of change. For example, it is possible to perform various processes according to software characteristics and event determination methods.

  In the first and second embodiments, as an event determination process, an event occurs when the calculated characteristic value exceeds a predetermined range Ra to Rb set based on the average x and the variance σ based on x ± 3σ. Event determination processing sets various determination conditions depending on the situation, such as when the average value exceeds the threshold or when the amount of change from the previous value exceeds the threshold ratio Is possible.

  In the first and second embodiments, the characteristic value data processing unit 120 performs data processing on characteristic values at predetermined time intervals among the stored characteristic values (the data processing unit is set to unit time). However, as another embodiment, the characteristic value data processing unit 120 performs data processing on the predetermined number of characteristic values every time the predetermined number of characteristic values are stored. It may be performed (the data processing unit is set to the unit generation count).

  Further, in the first and second embodiments, there is provided task type discrimination means for discriminating the task type of the characteristic value calculated by the characteristic value calculation unit 10, and the characteristic value data processing unit 120 performs data processing for each task type. The event determination unit 130 may perform the event occurrence detection process on the data processing result of the task type determined by the task type determination unit. That is, the characteristic value data processing unit 120 sets the data processing unit for each task type, and the event determination unit 130 performs event occurrence detection processing on the data processing result for each task type. Specifically, for example, a data processing unit is set for each task activation factor, a task is classified into a fixed time period task, an engine rotation synchronization task, and the like, and data processing is performed for each type.

  In the second embodiment, the data storage unit 150 executes a process of storing data after a predetermined time has elapsed since the occurrence of the event. At this time, processing for storing data may be executed.

  In the second embodiment, the data storage unit 150 may use a storage device different from the data temporary storage unit 140. Alternatively, for example, the data storage unit 140 may generate an event when an event occurs. A function of prohibiting rewriting of data for a predetermined period of time before and after may be provided, thereby functioning as the data storage unit 150.

The functional block diagram of the software operation | movement analysis apparatus which is 1st Embodiment of this invention. The flowchart which shows the software operation | movement analysis process in the software operation | movement analysis apparatus of FIG. The graph which shows the example of a detection of the event generation | occurrence | production in the software operation | movement analysis process of FIG. The functional block diagram of the software operation | movement analysis apparatus which is 2nd Embodiment of this invention. 5 is a flowchart showing software operation analysis processing in the software operation analysis apparatus of FIG.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1,1A ... Software operation | movement analysis apparatus, 10 ... Measurement data input part (parameter value acquisition means), 110 ... Characteristic value calculation part (task operation state characteristic value detection means), 120 ... Characteristic value data processing part (Task operation state characteristic) Value storage means, task type determination means, data processing means), 130 ... event determination section (event occurrence detection means), 140 ... temporary data storage section (internal data storage means), 150 data storage section (internal data storage means) , 200 ... reader.

Claims (7)

In a software operation analysis apparatus for analyzing the operation of software having at least one task,
Task operation state characteristic value detecting means for detecting a task operation state characteristic value representing the operation state characteristic of each task;
Task operation state characteristic value storage means for storing the task operation state characteristic value detected by the task operation state characteristic value detection means;
Data processing means for performing data processing for extracting the characteristics of the task operation state characteristic value for the task operation state characteristic value stored by the task operation state characteristic value storage means;
A predetermined range is set for the data processing result of the data processing means according to the characteristic obtained from the data processing result, and the task operation state characteristic value detected by the task operation state characteristic value detection means is set. A software operation analysis apparatus comprising: event occurrence detection means for detecting that the software event has occurred on the condition that the event is not included in a predetermined range. The software operation analysis apparatus according to claim 1,
Task type determination means for determining the task type of the task operation state characteristic value detected by the operation state value detection means;
The data processing means performs the data processing for each task type on the task operation state characteristic value stored by the task operation state characteristic value storage means,
The event occurrence detection means performs a process for detecting occurrence of an event on the data processing result of the task type determined by the task type determination means. In the software operation analysis apparatus according to claim 1 or 2,
The data processing means is configured such that the number of task operation state characteristic values stored for each task type by the task operation state characteristic value storage means exceeds a predetermined number from which the characteristics of the task operation state characteristic values are extracted. If there is a task that has passed a predetermined time such that the task operation state characteristic value characteristic is extracted from the start of storage of the task operation state characteristic value, the task operation state characteristic A software operation analysis apparatus that performs the data processing on a value. In the software operation analysis apparatus according to any one of claims 1 to 3,
Parameter value acquisition means for acquiring a parameter value used for the operation of the software;
One or both of the parameter value acquired by the parameter value acquisition unit and the task operation state characteristic value detected by the task operation state characteristic value detection unit are sequentially used as internal data of the software at a predetermined timing. Internal data storage means for collecting and storing internal data for a predetermined time immediately before or a predetermined number of times of collection;
Software operation comprising: internal data storage means for storing internal data stored in the internal data storage means on the condition that the event occurrence detection means has detected that an event has occurred Analysis device. In the software operation analysis apparatus according to any one of claims 1 to 4,
The data processing means, for the task operation state characteristic value stored for each type of task by the task operation state characteristic value storage means, for the task operation state characteristic value for the predetermined time every predetermined time interval, Alternatively, each time the number of stored task operation state characteristic values reaches a predetermined number, the data processing is performed for the predetermined number of task operation state characteristic values,
The event occurrence detection means is for a result of data processing performed for a predetermined time or a predetermined number of task operation state characteristic values including the task operation state characteristic value detected by the task operation state characteristic value detection means. And a software operation analysis device that performs processing for detecting the occurrence of an event. A software behavior analysis method for analyzing the behavior of software having at least one task,
A task operation state characteristic value detection step for detecting a task operation state characteristic value representing the operation state characteristic of each task;
A task operation state characteristic value storage step for storing the task operation state characteristic value detected in the task operation state characteristic value detection step in a task operation state characteristic value storage unit;
A data processing step for performing data processing for extracting the characteristics of the task operation state characteristic value for the task operation state characteristic value stored in the task operation state characteristic value storage step;
A predetermined range is set for the data processing result of the data processing step according to the characteristic obtained from the data processing result, and the task operation state characteristic value detected in the task operation state characteristic value detection step is set. A software operation analysis method comprising: an event occurrence detection step of detecting occurrence of the software event on condition that the software event is not included in the predetermined range. A software operation analysis program for causing a computer to execute a software operation analysis process for analyzing the operation of software having at least one task,
A task operation state characteristic value detection process for detecting a task operation state characteristic value representing the operation state characteristic of each task;
A task operation state characteristic value storage process for storing the task operation state characteristic value detected in the task operation state characteristic value detection process in a task operation state characteristic value storage unit;
Data processing for extracting the task operation state characteristic value for the task operation state characteristic value stored in the task operation state characteristic value storage process;
A predetermined range is set for the data processing result of the data processing according to the characteristic obtained from the data processing result, and the task operation state characteristic value detected in the task operation state characteristic value detection process is set to the predetermined predetermined value. A software operation analysis program comprising a function for causing the computer to execute an event occurrence detection process for detecting an occurrence of an event of the software on condition that the event is not included in a range.

Images