JP2015040850A - Vehicle stop determination apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To propose a vehicle stop determination apparatus for determining a vehicle stop reliably even in the case of incorporating an appliance that cannot obtain vehicular information such as a signal of a vehicle speed sensor in a vehicle that has large variation of vibration even when the vehicle is stopped and makes it possible to secure a space for installing an acceleration sensor.SOLUTION: A vehicle stop determination apparatus 1 includes: a three-axis acceleration sensor 12, disposed in a vehicle 2, for measuring acceleration in three different directions; a high-frequency pass filter 22 for attenuating a lower frequency component lower than a predetermined cut-off frequency for a measurement value of the three-axis acceleration sensor 12; and a determination process part 23 for determining whether the vehicle 2 is stopped, on the basis of a comparison between a frequency component of the measurement value that has passed through the high-frequency pass filter 22 and a determination threshold Vth.

Description

  The present invention relates to a stop determination device.

  In an in-vehicle device that provides driving support using electronic information obtained from a navigation system or an inter-vehicle communication system, information indicating whether the vehicle is stopped or traveling (hereinafter simply referred to as “stop information”). May be required). Stop information is, for example, allowing operation of in-vehicle devices only when the vehicle is stopped, or correcting GPS (Global Positioning System, global positioning system) positioning errors that are likely to occur while the vehicle is stopped. Required in case.

  By the way, usually, the stoppage determination can be realized by inputting a signal of a vehicle speed sensor that is normally equipped in the vehicle to the in-vehicle device.

  However, an interface for receiving various vehicle side information signals typified by a vehicle speed sensor, such as a small car navigation device such as PND (Portable Navigation Device, Personal Navigation Device) and a simple inter-vehicle communication system. There are some simple in-vehicle devices that are not equipped. Further, for example, a vehicle that employs a mechanical vehicle speed sensor, such as an inexpensive vehicle, may not be able to input the detection result of the vehicle speed sensor as an electrical signal to an in-vehicle device.

  Therefore, there is known a stop determination device that does not require a signal from a vehicle speed sensor and determines whether or not the vehicle is stopped based on a measurement result of an acceleration sensor.

  This conventional stoppage determination device has a degree of acceleration variation (that is, dispersion) in a state where the vehicle is stopped (hereinafter simply referred to as “stopped state”) and a state in which the vehicle is traveling (hereinafter simply referred to as “ Since the two variations of the acceleration variation (that is, the variance) in the “running state”) are different, the acceleration sample values for a certain period are statistically processed to calculate the variance of the acceleration sample values. It is determined that the vehicle is stopped depending on whether or not the variance exceeds a predetermined threshold (see, for example, Patent Document 1).

JP 2011-33604 A

  For example, in the case of a vehicle with relatively large engine vibration such as a motorcycle, the dispersion of acceleration samples is large even when the vehicle is stopped due to vibration during idling of the engine. That is, when the conventional stop determination device is applied to a vehicle with a large engine vibration during idling, it may be erroneously determined as a running state even though it is actually in a stop state.

  Moreover, the conventional stop determination device must measure the acceleration in the longitudinal direction of the vehicle, and needs to point the measurement axis of the acceleration sensor in the longitudinal direction of the vehicle. This is because, for example, in vehicles where it is difficult to secure an installation space such as a motorcycle, it is difficult to install an acceleration sensor in the first place, or a new stay is installed on the vehicle body to install the acceleration sensor. Or the aesthetics will be damaged by the stay.

  Therefore, the present invention provides a case where a device that cannot acquire vehicle-side information such as a signal of a vehicle speed sensor is mounted on a vehicle in which variation in vibration is large even when the vehicle is stopped and it is difficult to secure an installation space for an acceleration sensor. Even so, a stop determination device that makes a stop determination more reliably is proposed.

  In order to solve the above problems, a stop determination device according to the present invention includes a triaxial acceleration sensor that is provided in a vehicle and measures accelerations in three different directions, and a cutoff frequency that is determined in advance for measurement values of the triaxial acceleration sensor. A high-pass filter that gradually reduces a low-frequency component, and a determination processing unit that determines whether the vehicle is stopped based on a comparison between a frequency component of the measurement value that has passed through the high-pass filter and a determination threshold value; .

  According to the present invention, when a device that cannot acquire vehicle-side information such as a signal from a vehicle speed sensor is mounted on a vehicle in which vibration variation is large even when the vehicle is stopped and it is difficult to secure an installation space for an acceleration sensor. Even so, it is possible to provide a stop determination device that makes a stop determination more reliably.

The side view which shows the vehicle with which the stop determination apparatus which concerns on embodiment of this invention is applied. The block diagram which shows the stop determination apparatus which concerns on embodiment of this invention. The flowchart which shows the stop determination process which the stop determination apparatus which concerns on this embodiment performs. The flowchart which shows the determination threshold value update process which the stop determination apparatus which concerns on this embodiment performs. The flowchart which shows the determination threshold value update process which the stop determination apparatus which concerns on this embodiment performs. The conceptual diagram which shows the other example of the determination threshold value update process which the stop determination apparatus which concerns on this embodiment performs. The flowchart which shows the other example of the determination threshold value update process which the stop determination apparatus which concerns on this embodiment performs. The flowchart which shows the other example of the determination threshold value update process which the stop determination apparatus which concerns on this embodiment performs. The flowchart which shows the stop determination process which the stop determination apparatus which concerns on this embodiment performs.

  Hereinafter, embodiments of a stop determination device according to the present invention will be described with reference to FIGS. 1 to 9.

  FIG. 1 is a side view showing a vehicle to which a stop determination device according to an embodiment of the present invention is applied.

  As shown in FIG. 1, the vehicle 2 to which the stop determination device 1 according to this embodiment is applied is a so-called motorcycle. The vehicle 2 is disposed behind the vehicle body 3, a vehicle body 3 extending in the front-rear direction, a front wheel 5 as a steering wheel disposed in front of the vehicle body 3, a steering mechanism 6 that supports the front wheel 5 on the vehicle body 3 so as to be steerable. And a power unit 9 having an engine 8 that is integrated with the rear wheel 7 so as to be swingable in the vertical direction on the vehicle body 3.

  The vehicle 2 is not limited to a scooter type motorcycle, and may be another type of motorcycle.

  The stop determination device 1 is attached to a part of the vehicle 2 where the vibration of the engine 8 is transmitted, for example, the front end portion of the vehicle body 3. The stop determination device 1 is driven by being supplied with power from the vehicle 2, for example, DC 12V.

  By the way, the stop determination device 1 is a simple vehicle-mounted device that does not include an interface for receiving various vehicle-side information signals represented by a vehicle speed sensor, such as a small car navigation device or a simple inter-vehicle communication system. 11 is provided with stop information and is a part of the function of the in-vehicle device 11.

  FIG. 2 is a block diagram showing the stop determination device according to the embodiment of the present invention.

  As shown in FIG. 2, the stop determination device 1 according to this embodiment includes a triaxial acceleration sensor 12 that is provided in the vehicle 2 and measures accelerations in three different directions, and an acceleration measurement value from the triaxial acceleration sensor 12. The control unit 13 that determines whether or not the receiving vehicle 2 is stopped, the power supply circuit 15 that supplies power to the triaxial acceleration sensor 12 and the control unit 13, and whether or not the vehicle 2 is stopped from the control unit 13 And an output port 16 for extracting the determination result.

  In addition, when incorporating the stop determination apparatus 1 as a part of function of the vehicle equipment 11, for example, a part of PND function, the control unit 13 and the power supply circuit 15 can be shared with other functions of the PND. Moreover, since the determination result as to whether or not the vehicle 2 is stopped can be provided to other functions of the PND inside the control unit 13, the output port 16 is not necessary.

  Furthermore, examples of the in-vehicle device 11 incorporating the stop determination device 1 include a vehicle-to-vehicle communication system, a road-to-vehicle communication system, a mobile phone, and a smartphone in addition to PND. When the stop determination device 1 is incorporated in these devices, the CPU, the acceleration sensor, the power supply circuit, and the like that are originally provided in the device can be used for the stop determination device 1.

  The triaxial acceleration sensor 12 measures acceleration in three different directions, for example, three orthogonal directions, and outputs each value to the control unit 13. The measured values of acceleration in each of the three directions are represented as Ax (n), Ay (n), and Az (n). n is the number of samples, for example, n = 1 to 100.

  The control unit 13 includes a microprocessor (not shown), a storage device (not shown) that stores various arithmetic programs executed by the microprocessor, parameters, and the like. The control unit 13 determines in advance the acceleration detection unit 21 that acquires the acceleration measurement values Ax (n), Ay (n), and Az (n) from the triaxial acceleration sensor 12 and the measurement values of the triaxial acceleration sensor 12. It is determined whether or not the vehicle 2 is stopped based on a comparison between the high-pass filter 22 that gradually decreases the frequency component lower than the cutoff frequency, and the frequency component of the measurement value that has passed through the high-pass filter 22 and the determination threshold value Vth. A determination processing unit 23 and a determination threshold learning unit 25 that learns and updates the determination threshold Vth according to the vehicle 2 mounted are provided.

  The cutoff frequency of the high-pass filter 22 is set to a value obtained by adding a predetermined value to the primary vibration frequency when the engine 8 is idle. The frequency components of the measured values that have passed through the high-pass filter 22 for the accelerations in the three directions are represented as Ahx (n), Ahy (n), and Ahz (n).

  The power supply circuit 15 converts power supplied from the vehicle 2, for example, DC 12 V, and adapts it to the triaxial acceleration sensor 12 and the control unit 13.

  Next, the stop determination process performed by the stop determination device 1 according to the present embodiment will be described in detail.

  FIG. 3 is a flowchart showing a stop determination process performed by the stop determination device according to the present embodiment.

  As shown in FIG. 3, the stop determination device 1 according to the present embodiment firstly measures acceleration values Ax (n), Ay (n), Az (in three directions) from the triaxial acceleration sensor 12 to the acceleration detection unit 21. n) is received (step S1), and the acceleration measurement values Ax (n), Ay (n), Az (n) are sequentially passed through the high-pass filter 22 and the filtered values Ahx (n), Ahy ( n), Ahz (n) is temporarily stored in the memory of the control unit 13 (from step S2 to step S4). Steps S1 to S4 are performed every execution cycle Ts, for example, every Ts = 3 milliseconds.

  Next, the stop determination device 1 reads the filtered values Ahx (n), Ahy (n), Ahz (n) from the memory of the control unit 13 to the determination processing unit 23, and performs three directions at predetermined time intervals. Root mean square (RMS) Vix, Viy, and Viz are calculated (step S11 to step S13), and the arithmetic average Vi of the three-way calculated values Vix, Viy, and Viz is obtained. (Step S14), the arithmetic average Vi is compared with the determination threshold value Vth to determine whether or not the vehicle 2 is stopped (Step S16). At this time, the stop determination device 1 learns and updates the determination threshold Vth by the determination threshold learning unit 25 before step S16 (step S15). Steps S11 to S16 are performed every execution cycle T1, which is longer than the execution cycle Ts, for example, T1 = 100 milliseconds.

  Details of each processing from step S1 to step S16 will be described.

First, in step S <b> 1, the acceleration detection unit 21 acquires three-direction acceleration measurement values Ax (n), Ay (n), and Az (n) measured by the three-axis acceleration sensor 12 for each sample time Ts. The sample time Ts is set to a value that can measure the frequency component of the secondary vibration of the engine 8. For example, the sampling time Ts with respect to the maximum value Rmax of the engine speed to be measured (rpm: revolution per minute, or rotation per minute) is expressed by the conditional expression of [Equation 1].
[Equation 1]
Sampling time Ts ≦ 1 ÷ (maximum value Rmax ÷ 60 × 2 × 2)

As for the engine speed to be measured, that is, the engine speed when the engine 8 is idle, the maximum value Rmax = 5000 times per minute is sufficient for the general engine 8. In this case, the sampling time Ts is obtained by the conditional expression [Equation 2].
[Equation 2]
Sampling time Ts ≦ 1 ÷ (5000 ÷ 60 × 2 × 2) = 0.003 = 3 milliseconds

  In steps S2 to S4, the high-pass filter 22 determines the acceleration from the acceleration measurement values Ax (n), Ay (n), and Az (n), the acceleration due to acceleration / deceleration of the vehicle and the attitude change, and the engine during idling. A frequency component equal to or lower than the cutoff frequency, such as primary vibration, is cut off, and a frequency component higher than the cutoff frequency mainly including engine vibration when the engine speed increases is extracted. The values Ahx (n), Ahy (n), and Ahz (n) after the filtering process are stored in buffers provided in the memory space. The buffer accumulates the filtered values Ahx (n), Ahy (n), and Ahz (n) for the execution period T1.

In step S11 to step S13, the determination processing unit 23 calculates the average of the square values of the values Ahx (n), Ahy (n), and Ahz (n) after the filter processing, and takes the square root to obtain the magnitude. The thickness (amplitude) is calculated. The magnitude of the acceleration vibration in each of the three directions is expressed as Vix, Viy, Viz. The relationship between acceleration vibration magnitudes Vix, Viy, Viz and Ahx (n), Ahy (n), Ahz (n) is shown in [Expression 3] to [Expression 5].

In step S <b> 14, the determination processing unit 23 calculates the arithmetic average Vi of the magnitudes Vix, Viy, and Viz of acceleration vibrations in the three directions, as shown in the equation [Equation 6]. The arithmetic average Vi is an index Vi that comprehensively indicates the magnitude of engine vibration at a location where the triaxial acceleration sensor 12 is installed.
[Equation 6]
Arithmetic mean Vi (index Vi) = (Vix + Viy + Viz) ÷ 3

  In general, engine vibration shows a positive correlation between frequency (that is, engine speed) and magnitude of vibration. However, when observing at any location of the vehicle 2, the engine vibration is affected by vibration characteristics such as a resonance point or non-resonance on the vehicle 2 side. The influence of the vibration characteristics on the vehicle 2 side also varies depending on the direction in which engine vibration is measured. That is, the measurement result of the triaxial acceleration sensor 12 installed in any part of the vehicle 2 indicates that the frequency and the magnitude of vibration are different in each of the three directions compared to the case where the engine vibration is directly measured by the engine 8. Correlation strength is different and frequency characteristics are also different.

  Therefore, the inventor of the present invention uses an index Vi that averages the three directions of the three-axis acceleration sensor 12 to cancel the influence of the vibration characteristics on the vehicle body 3 side, and to show the correlation between the frequency and the magnitude of vibration. It was experimentally confirmed that the correlation was stabilized and the correlation closely observed with the engine 8 can be obtained. That is, the stop determination device 1 of the present embodiment improves the determination accuracy of the stop determination by evaluating the measurement result of the triaxial acceleration sensor 12 affected by the vibration characteristics on the vehicle body 3 side by the arithmetic average Vi.

  In step S15, the determination threshold value learning unit 25 learns and updates the determination threshold value Vth from the time series information of the arithmetic average Vi acquired up to that point. Two methods of learning the determination threshold value Vth will be described.

  4 and 5 are flowcharts showing determination threshold update processing performed by the stop determination device according to the present embodiment.

  As shown in FIG. 4 and FIG. 5, the determination threshold value learning unit 25 of the stop determination device 1 according to the present embodiment is configured to detect acceleration vibration in each of the three directions when the determination processing unit 23 determines that the vehicle stops. If the arithmetic average Vi of the magnitudes Vix, Viy, Viz, that is, a value slightly larger than the index Vi is smaller than the determination threshold Vth, the determination threshold Vth is updated to a value slightly larger than the index Vi.

  Specifically, the determination threshold value learning unit 25 sets the maximum value that can take the initial value of the determination threshold value Vth. Then, the determination threshold value learning unit 25 determines whether or not the vehicle is stopped, that is, whether or not step S16 is established (step S21). When the stop determination is established (step S21 True), the determination threshold value learning unit 25 temporarily stores the maximum value Vimax of the indicator Vi during the update time Tr (FIG. 5, step S31 to step S40) ( Steps S22 and S23).

After the execution cycle T1 from step S11 to step S16 has elapsed (steps S24 and S25), the maximum value Vimax is corrected to a slightly larger value to stabilize the stoppage determination result (step S26). . The relationship between the maximum value Vimax and the correction value Vimax ′ is as [Equation 7]. [Expression 7] represents a calculation formula for calculating a value slightly larger than the index Vi.
[Equation 7]
Correction value Vimax ′ = a × maximum value Vimax + b
The constants a and b are set in a range of 1 ≦ a ≦ 1.2 and 0.1 × Viid ≦≦≦ 0.2 × Vidle, for example. Here, Vidle is an average value of the index Vi during idling.

  If the correction value Vimax 'is smaller than the current determination threshold value Vth, the determination threshold value Vth is updated with the correction value Vimax' (steps S27 and S28). Since there is a characteristic that the indicator Vi becomes the smallest in the stop state, the determination threshold Vth finally converges to a value Vi ′ obtained by correcting the indicator Vi in the stop state, and the stop determination is made by adopting this as the determination threshold Vth. It becomes possible.

  The determination whether or not the determination threshold update process is completed (step S30) is the time when the stop determination is established at the current determination threshold Vth each time the determination threshold Vth is updated (step S31 to step S33), specifically Specifically, the counter timer 1 is counted (step S34 True, step S35), the time when the stop determination is not established, specifically the counter timer 2 is counted (step S34 False, step S36), When the set update completion determination threshold count value Tth is exceeded (steps S37 and S38), the update is completed (step S39). For example, when the update completion determination threshold period is 30 seconds, the update completion determination threshold count value Tth is changed from an execution cycle T1, for example, T1 = 100 milliseconds to Tth = 300.

  FIG. 6 is a conceptual diagram illustrating another example of the determination threshold update process performed by the stop determination device according to the present embodiment.

  7 and 8 are flowcharts illustrating another example of the determination threshold value update process performed by the stop determination device according to the present embodiment.

  As shown in FIGS. 6 to 8, the determination threshold value learning unit 25 of the stop determination device 1 according to the present embodiment uses the arithmetic average Vi of the magnitudes Vix, Viy, and Viz of acceleration vibrations in the three directions, that is, the index Vi. The frequency of appearance is counted for each size, and between the size of the size where the frequency of appearance increases when the vehicle 2 is stopped and the size of the size where the frequency of appearance increases when the vehicle 2 is traveling. The determination threshold value Vth is statistically obtained and updated.

Specifically, the determination threshold value learning unit 25 counts the appearance frequency of the index Vi for each size. As shown in FIG. 6, this counting result draws a histogram in which the horizontal axis is the size of the index Vi and the vertical axis is the appearance frequency. As the vehicle 2 repeats stopping and running, two peaks appear in the histogram, one peak centering on the indicator Vi in the stopped state and the other peak centering on the indicator Vi in the running state. The positions of the valleys of the two peaks on these histograms are specified, and the determination threshold Vth is updated with the size of the index Vi at the valley positions.
As an algorithm for specifying the position of the valley, for example, there is a kittler method generally known as a threshold selection method for binarization processing in image processing.

  The kittler method obtains an optimum threshold value from a histogram by a statistical method. The kittler method calculates a threshold candidate k that minimizes E (k) calculated from [Equation 8] to [Equation 15] when the threshold candidate k is changed. This threshold candidate k is updated as the determination threshold Vth.

ただし、［数８］中の各項は以下の通り。

ｐｉ：ある指標Ｖｉの大きさｉにおけるヒストグラムの頻度値ｎｉを正規化した値。
σ０（ｋ）：ある閾値候補ｋの時の閾値候補ｋ以下の区間における指標Ｖｉの大きさｉの分散。
σ１（ｋ）：ある閾値候補ｋの時の閾値候補ｋ以上の区間における指標Ｖｉの大きさｉの分散。

However, each item in [Equation 8] is as follows.

pi: A value obtained by normalizing the histogram frequency value ni for the size i of a certain index Vi.
σ 0 (k): Variance of the size i of the index Vi in a section equal to or less than the threshold candidate k at the time of a certain threshold candidate k.
σ1 (k): variance of the size i of the index Vi in a section equal to or greater than the threshold candidate k at the time of a certain threshold candidate k.

  The determination of whether or not the determination threshold update process has been completed (step S59) is based on the difference between the maximum frequency and the frequency at the threshold candidate k in the two sections below the threshold candidate k and above the threshold candidate k on the histogram. When the calculation is completed (step S65) and both of them exceed a predetermined update completion determination threshold difference Hth, for example, 50 (steps S66 and S67), the update is completed (step S68).

  FIG. 9 is a flowchart showing stop determination processing performed by the stop determination device according to the present embodiment.

  As illustrated in FIG. 9, the determination processing unit 23 of the stop determination device 1 according to the present embodiment determines that the index Vi continuously falls below the latest determination threshold Vth during the stop determination period Ton in step S16. Then, it is determined that the vehicle 2 is in a stopped state (from step S71 to step S77). On the other hand, when the index Vi continuously exceeds the latest determination threshold value Vth during the travel determination period Toff, it is determined that the vehicle 2 is not in a stopped state and is in a travel state (steps S78 to S83). .

  The three-axis acceleration sensor 12, the high-pass filter 22, the determination processing unit 23, and the determination threshold learning unit 25 can be partly or entirely implemented with a digital circuit, and can be implemented with an analog circuit. Is also possible. For example, an analog circuit implements up to the square root calculation of the mean value of the square values of Ahx (n), Ahy (n), and Ahz (n), and inputs the processing result to the analog input port of the CPU. It is possible to perform digital processing in the CPU.

  The stop determination device 1 according to the present embodiment can remove gravity acceleration, acceleration / deceleration of the vehicle 2 and acceleration due to posture change by the high-pass filter 22, so any vehicle 2 can be used as long as it can detect engine vibration. Can be installed even in

  Moreover, since the stop determination apparatus 1 according to the present embodiment determines whether or not the vehicle 2 is stopped from the three-direction acceleration measured by the three-axis acceleration sensor 12, an erroneous determination based on a certain one-way vibration characteristic is performed. It is difficult to wake up, the accuracy of judgment is improved, and the directionality of installation is also highly flexible.

  Furthermore, the stop determination device 1 according to the present embodiment performs stop determination with the arithmetic average Vi, thereby determining the vehicle type of the vehicle 2, the installation location of the stop determination device 1, the installation direction (the measurement axis of the triaxial acceleration sensor 12). Direction), acceleration measurement values Ax (n), Ay (n), Az (n), and values Ahx (n) after the filtering process, which vary depending on the transmission characteristics between the engine 8 and the stop determination device 1. ), Ahy (n), Ahz (n), and fluctuations in the frequency characteristics of the acceleration vibration magnitudes Vix, Viy, Viz in the three directions, respectively, to prevent erroneous determination.

  Furthermore, since the stop determination device 1 according to the present embodiment learns and updates the determination threshold Vth, the appropriate determination threshold Vth is applied even if the vehicle type of the vehicle 2 or the installation location of the stop determination device 1 changes. Thus, stop determination can be made. In particular, since the stop determination device 1 self-learns the determination threshold value Vth, a preliminary experiment is performed for each vehicle type of the vehicle 2 or each installation location of the stop determination device 1, and an appropriate determination threshold value Vth is obtained. After the stop determination device 1 is installed without performing any operation, it is possible to effectively perform stop determination with high accuracy simply by repeating running and stopping for a while.

  In addition, the stop determination device 1 according to the present embodiment has a value Ahx (after filtering) that has passed through the high-pass filter 22 to which a cutoff frequency obtained by adding a predetermined value to the primary vibration frequency when the engine 8 is idle is applied. n), Ahy (n), Ahz (n) are compared with the determination threshold value Vth to determine whether or not the vehicle 2 is stopped. The difference from when the engine speed is increased (when the vehicle 2 is traveling) is clarified, and the accuracy of stop determination is improved. In a single-cylinder engine, the engine primary vibration during idling is relatively large, and the difference in acceleration measurement values Ax (n), Ay (n), Az (n) between idling and engine speed increase is small. It is particularly effective to suppress the influence of the engine primary vibration during idling.

  Therefore, according to the stop determination device 1 according to the present invention, the vehicle 2 has a large variation in vibration even when the vehicle is stopped, and it is difficult to secure an installation space for the acceleration sensor. Even when a vehicle-mounted device 11 that cannot acquire information is installed, it is possible to make a stop determination more reliably.

DESCRIPTION OF SYMBOLS 1 Stop determination apparatus 2 Vehicle 3 Car body 5 Front wheel 6 Steering mechanism 7 Rear wheel 8 Engine 9 Power unit 11 In-vehicle equipment 12 Triaxial acceleration sensor 13 Control part 15 Power supply circuit 16 Output port 21 Acceleration detection part 22 High-pass filter 23 Determination processing part 25 determination threshold learning unit

Claims (5)

A triaxial acceleration sensor that is provided in the vehicle and measures acceleration in three different directions;
A high-pass filter that gradually decreases a frequency component lower than a predetermined cutoff frequency for the measurement value of the three-axis acceleration sensor;
A stop determination device comprising: a determination processing unit that determines whether or not the vehicle is stopped based on a comparison between a frequency component of the measurement value that has passed through the high-pass filter and a determination threshold value. The determination processing unit calculates a square root of an average value of square values in three directions at predetermined time intervals for the frequency component of the measurement value that has passed through the high-pass filter, and further calculates a value in three directions. The stop determination device according to claim 1, wherein the arithmetic average of the vehicle is compared with the determination threshold value. A determination threshold learning unit that updates the determination threshold to the slightly larger value if a value slightly larger than the arithmetic average is smaller than the determination threshold when the determination processing unit determines that the vehicle is stopped. The stop determination device according to 2. About the arithmetic average, the frequency of appearance is counted by size, the size of the appearance frequency increases when the vehicle is stopped, and the size of the frequency of increase appearance when the vehicle is traveling The stop determination apparatus according to claim 2, further comprising a determination threshold learning unit that statistically obtains and updates the determination threshold. The stop determination device according to any one of claims 1 to 4, wherein the cutoff frequency is a value obtained by adding a predetermined value to a primary vibration frequency when the engine is idling.

Images