JP2007121401A - Simulator for car driving - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a simulator for car driving, enabling evaluation of a road surface by reflecting localized road surface condition and expressing future traveling state wherein the road conditions are changed. <P>SOLUTION: In the simulator for car driving, the data of the position of the road surface to be the object of simulation traveling and the data of road surface height, corresponding to the position are classified into segments wherein waveform characteristics (1) of a reference road surface, are divided into waves (2a, 2b, 2c, 2d, 2e, 2f) of multiple kinds of frequencies (a, b, c, d, e, f) and defined, according to the characteristics of a localized oscillating part (4), wherein oscillation amplitude locally varies by a larger amount than at other continuous portions and appear in a waveform (3) obtained by division, and are inputted. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an automobile simulation driving apparatus that can express a driving state under a predetermined condition without actually driving an automobile.

  Traditionally, road investment has been focused on new construction, but in recent years it can be said that road construction in Japan has almost been completed. How to maintain and manage roads constructed so far has become an important issue. ing. And how much damage the road surface causes to allow people and cars to run comfortably, or what road factors are involved in potential hazards such as near-miss experiences etc. The influence of road surface conditions on driving safety and comfort is a particularly important factor in road maintenance.

  On the other hand, regarding road maintenance management, use of a simulated vehicle driving device has been proposed in road traffic evaluation (for example, Japanese Patent Application Laid-Open No. 2002-157673). The vehicle simulation driving device is a means for improving the skill of the driver while avoiding the risk of causing an accident in the training of an immature driver. It is used not only for training but also for a wide range of purposes such as collecting data necessary for vehicle design and data for road traffic evaluation. Various attempts have been made to improve the vehicle driving simulation device that can be used for various purposes as described above. For example, in the technology disclosed in Japanese Patent Application Laid-Open No. 2003-114607, ITS (Intelligent Transport System), etc. Attempts have been made to provide a low-cost virtual driving system that is effective for human interface evaluation utilized in the world.

  However, although the conventional automobile simulation driving device has been improved in accordance with various data collection purposes, the main point has been to reproduce the visual virtual space and pursue the force. Therefore, there is no conventional automobile driving simulation device that can be applied to data collection necessary for road surface evaluation.

Therefore, one of the present inventors has attempted to reflect the road surface state in the simulated traveling as disclosed in Japanese Patent Application No. 2004-131866 in order to more accurately reproduce the actual traveling state. According to this attempt, in addition to the video captured by the image capturing device and the data collected by the motion sensor (accelerometer and angular velocity meter) during actual vehicle running on the actual road surface, the measured road surface data is further converted into the power spectral density. (PSD) classifies and captures, and the running state is reproduced by the visual field display device and the shaking device based on these data. In addition, the method of expressing road surface property by PSD is determined by ISO8608.
JP 2002-157673 A JP 2003-114607 A Japanese Patent Application No. 2004-131866

  However, in the device of Japanese Patent Application No. 2004-131866, the road surface condition is considered, but only a rough change of the road surface is reflected, and the localized road surface conditions such as steps and fine cracks are reflected. It was not enough to be applied to the road surface evaluation.

  In addition, all the conventional simulation traveling devices including the device according to Japanese Patent Application No. 2004-131866 can reproduce the traveling state based on the current state of the road, but well express the future traveling state in which the road state has changed. I could not. On the other hand, in the road surface evaluation, since the data on the behavior of the vehicle in the road surface condition deteriorated with time is valid and necessary, the conventional vehicle driving simulation device which cannot express the future driving state reflecting the time deterioration of the road surface well is this point. However, it was not applicable to road surface evaluation.

  SUMMARY OF THE INVENTION An object of the present invention is to provide an automobile driving simulation device that can reflect a localized road surface condition and perform a road surface evaluation by expressing a future driving state in which the road state has changed.

In the automobile driving simulation apparatus according to the present invention, the road surface position to be simulated and the road height data corresponding to the position decompose the wavy characteristics of the reference road surface into waves of a plurality of types of frequencies. Are input after being classified into categories defined according to the characteristics of the local vibration part appearing in the waveform obtained by the above and whose amplitude fluctuates more locally than other continuous parts.
If the local vibration part does not appear in the waveform obtained by decomposition, it can be defined as a section where the local vibration part does not appear, for example, a normal road surface section, depending on the characteristics of the local vibration part. Included in the definition.

  The decomposition may be performed by spatial frequency analysis.

  In the spatial frequency analysis, discrete wavelets may be used.

  The discrete wavelet may be configured based on a scaling function, and the decomposition may be performed by the multi-resolution analysis or wavelet packet analysis.

  According to the automobile driving simulation apparatus according to the present invention, the local road surface condition such as a step or a fine crack is obtained by decomposing the wavy characteristic of the road surface into waves of a plurality of types of frequencies. By utilizing the feature that appears as a local vibration part that fluctuates more locally than the part, the local road surface condition can be reflected in the simulated traveling. For example, as shown in FIG. 2 (a), the wavy characteristics of the road surface from the reference point O to the point B through the region A having a step (the distance from the reference point O is the horizontal axis and the road surface height is the vertical axis The road 2a having the frequency a), the wave 2b having the frequency b, the wave 2c having the frequency c, the wave 2d having the frequency d, the wave 2e having the frequency e, and the wave 2f having the frequency f. When the waveform in which only the wave 2c and the wave 2d are strong is obtained in FIG. 5, by applying vibration based on the waveform of the wave 2c and the wave 2d in the region A, the step difference localized in the region A can be reflected in the simulated traveling. it can. Then, grasp the characteristics of the localized vibration parts corresponding to various localized road surface conditions such as steps and fine cracks, define the classification according to the characteristics of these localized vibration parts, and classify the road surface data into the classification If it is classified and input, it is possible to reflect various localized road surface conditions by applying vibration according to the classification in the simulated traveling.

  Further, since the local road surface state is expressed as a local vibration part, it is possible to express a change with time by processing the local vibration part. For example, in the above example, increase in the strength of the local vibration parts of the waves 2c and 2d, or increase the length in the horizontal axis direction, etc., to express the enlargement of the step due to the deterioration of the road surface over time. Can do. And, by adapting the vibration given according to the classification of the road surface data in the simulated traveling to the change considering the temporal deterioration of the local vibration part, it is possible to express the future traveling state in which the road state has changed Become.

  If the decomposition of the wavy characteristics of the reference road surface is executed by spatial frequency analysis, a known wavelet transform technique can be applied. In general wavelet theory, time signal data is often handled, and it corresponds to what is called time frequency analysis for time signals, and it corresponds to the wavy characteristics of the road surface that is space axis (distance axis) data. All wavelet transform analysis is called spatial frequency analysis.

  If a discrete wavelet is used in the spatial frequency analysis, it becomes suitable for handling by an electronic computer, so that it is possible to realize an automobile driving simulation device that requires processing by the electronic computer.

If a discrete wavelet is configured based on a scaling function, an approximation function (linear combination of scaling functions) of the road surface wavy characteristic is expressed by a hierarchical structure of resolution, so that the properties can be analyzed for each resolution. For this reason, it is possible to appropriately select the resolution according to the quality of the actually obtained data, for example, the size of the road surface height data collection interval, and perform the optimal analysis of the wavy characteristics of the road surface.
If the decomposition of the wavy characteristics of the reference road surface is executed by multi-resolution analysis, it is possible to mainly reflect the macro characteristics of the total length of the road surface to be traveled while reflecting the steps and fine cracks localized on the road surface. . This is because multi-resolution analysis is based on extraction of low-pass components (Approximation).
On the other hand, if the decomposition of the wavy characteristics of the reference road surface is executed by wavelet packet analysis, the high-pass component (Detail) can be extracted in the same degree as the low-pass component in addition to the low-pass component. Therefore, the micro features of arbitrary components such as steps and fine cracks in the entire length of the road surface to be traveled can be reflected using the high-pass component.

  FIG. 1 is a block diagram showing the concept of a specific example of an automobile driving simulation apparatus according to the present invention. 2 and 3 show examples of reference road surface data used to define the data classification of the vehicle driving simulation device, (a) shows the wavy characteristics of the reference road surface, (b) shows a plurality of types of the same wave characteristics. It is a graph which shows the waveform obtained by decomposing into a frequency wave.

  In this automobile driving simulation device, a subject performs a simulated running in a cockpit 11 equipped with a steering wheel, an accelerator pedal, a brake pedal, and the like as in a normal automobile. A visual display device 12 is provided in front of the cockpit 11 so that visual information similar to that in actual traveling is given to the subject. In addition, the cockpit 11 is provided with vibration information similar to that in actual traveling to the subject via the shaking device 13.

The operations of the visual display device 12 and the shaking device 13 are controlled by the control device 14. The control device 14 operates the visual display device 12 and the shaking device 13 based on the input data 15 and the operation information from the cockpit 11 to give appropriate information to the subject.
As the cockpit 11, the visual display device 12, the shaking device 13, and the control device 14, a known device, for example, a device disclosed in Japanese Patent Application No. 2004-131866 can be used.

  As the input data 15, the position of the road surface to be simulated and the road surface height corresponding to the position are used. FIGS. 2 (a) and 3 (a) are graphs showing the wavy characteristics of a reference road surface, which will be described later, and the input data 15 also has the same format as FIGS. 2 (a) and 3 (a). Use.

  The input data 15 is a waveform obtained by decomposing the wavy characteristic 1 of the reference road surface into waves 2a, 2b, 2c, 2d, 2e and 2f of a plurality of types of frequencies a, b, c, d, e and f. 3 is input after being classified into categories defined in accordance with the characteristics of the local vibration unit 4 whose amplitude varies locally more largely than other continuous portions. In addition, what is necessary is just to define a division suitably based on the state of the actual road surface selected as a reference | standard road surface, However, The case shown in FIG.2 and FIG.3 is demonstrated as an example of a division.

  The reference road surface shown in FIG. 2 has a step in the region A. When the position of the road surface is taken on the horizontal axis and the road surface height is taken on the vertical axis, the reference road surface vibrates greatly in the region A as shown in FIG. It has a wave characteristic 1. Further, in the waveform 3 obtained by decomposing the wave-like characteristic 1 into the waves 2a, 2b, 2c, 2d, 2e, and 2f, the amplitude of the waves 2c and 2d fluctuates more locally than other continuous portions. Local vibration part 4 appears. Therefore, a section where the localized vibration unit 4 appears in the waves 2c and 2d is defined as a “step road surface”. If there is a step on the road surface to be simulated, the data of that portion is input as belonging to the category “step road surface”.

  The reference road surface shown in FIG. 3 is not damaged at all. When the road surface position is taken on the horizontal axis and the road surface height is taken on the vertical axis, the road surface slightly vibrates due to small unevenness of the road surface as shown in FIG. It has a wavy characteristic 1. Further, the waveform 3 obtained by decomposing the wave-like characteristic 1 into the waves 2a, 2b, 2c, 2d, 2e, and 2f has a constant frequency, and there is no localized vibration portion. Therefore, the section where the localized vibration part does not appear is defined as “general paved road”. If the road surface to be simulated is paved, most data is input as belonging to the category “general paved road” unless there is a particular damage or step to be noted.

  Note that the number of waves for decomposing the wave characteristic 1 is not limited, and can be determined as appropriate according to the accuracy of the obtained data. However, as the number increases, the localized road surface condition can be reflected more accurately, and therefore it is preferable to increase the number of waves as much as possible.

  According to this automobile driving simulation device, the local road surface conditions such as steps and fine cracks are obtained by decomposing the wave surface characteristics of the road surface into waves of a plurality of types of frequencies, and the amplitude is greater than that of other continuous parts. By utilizing the feature that appears as a local vibration part that varies greatly locally, the local road surface condition can be reflected in the simulated traveling. For example, in the case of the stepped road surface shown in FIG. 2, the step localized in the region A can be reflected in the simulated traveling by applying vibration based on the waveforms of the waves c and d in the region A. Then, grasp the characteristics of the localized vibration parts corresponding to various localized road surface conditions such as steps and fine cracks, define the classification according to the characteristics of these localized vibration parts, and classify the road surface data into the classification If it is classified and input, it is possible to reflect various localized road surface conditions by applying vibration according to the classification in the simulated traveling.

  In addition, since the localized road surface condition (the step in the example of FIG. 2) is expressed as the localized vibration part 4, the temporal change can be expressed by processing the localized vibration part 4. For example, the amplitude of the localized vibration part 4 of the wave 2c and the wave 2d can be increased, or the length in the horizontal axis direction can be increased to express the increase in the level difference due to the deterioration of the road surface over time. Then, by adapting the vibration given according to the classification of the road surface data in the simulated traveling to the change in consideration of the temporal deterioration of the local vibration unit 4, it is possible to express the future traveling state in which the road state has changed. It becomes.

The decomposition of the wavy characteristic 1 of the reference road surface is performed by spatial frequency analysis. Specifically, it is executed by analysis using a known wavelet transform represented by Formula 1. Note that f (x) in Equation 1 is a function of the wave characteristic 1 to be decomposed.

なお、ｊはレベルと呼ばれ、その値の小さいものは高周波数になる。一方、ｋはシフト距離（横軸方向の移動）を示す。ここで、数式２の離散ウェーブレットを適当に選択して直交系（内積がｊ＝ｋ以外の場合すべて０になる関数群）とすれば、波状特性１の関数ｆ（ｘ）は数式３の級数に展開される。

なお、ｃ_ｊｋは展開係数またはウェーブレット係数とよばれる。 However, in the case of Equation 1, infinite integration using a and b is necessary, which is not realistic. In order to be suitable for handling by the controller 14 is a computer, the a in Equation 1 to 2 ^j, transformation by the discrete wavelet (Equation 2) obtained by the binary split representation by replacing b to 2 ^{j k} It is executed by spatial frequency analysis using

Note that j is called a level, and a small value has a high frequency. On the other hand, k represents a shift distance (movement in the horizontal axis direction). Here, if the discrete wavelet of Formula 2 is appropriately selected to form an orthogonal system (a group of functions that are all 0 when the inner product is other than j = k), the function f (x) of the wave characteristic 1 is a series of Formula 3. Expanded to

Note that c _jk is called an expansion coefficient or a wavelet coefficient.

ここに、ｄ_ｊｋはスケーリング係数とよばれる。 Furthermore, by constructing the discrete wavelet of Formula 2 based on the scaling function, the wave characteristic 1 can be approximated by a linear combination of the scaling function and expressed by a hierarchical structure of resolution. A general expression of an approximate function in which the scaling function is represented by φ is as shown in Expression 4. For example, a Haar function may be adopted as the scaling function.

Here, d _jk is called a scaling factor.

ここに、ｈとｇは、それぞれローパス及びハイパスフィルタであり、これはφ（スケーリング関数）及びψ（離散ウェーブレット）から導かれる。また、ｊ＋１次のスケーリング係数ｄ_ｊ＋１は、ｊ次より一つ下の低解像度表現であり、解析周波数及び距離解像度がｊ次の２分の１となる。このように、近似関数ｆを解像度の階層構造で表現すれば、解像度ごとにその性質を分析することができる。 A scaling coefficient and a wavelet coefficient having a resolution of jth order or less can be obtained by iterative calculation of Expression 5 and Expression 6 based on d _jk .

Here, h and g are low-pass and high-pass filters, respectively, which are derived from φ (scaling function) and ψ (discrete wavelet). The j + 1-order scaling coefficient d _{j + 1} is a low-resolution expression that is one order lower than the j-order, and the analysis frequency and the distance resolution are ½ of the j-order. As described above, if the approximate function f is expressed by a hierarchical structure of resolution, the property can be analyzed for each resolution.

  When the approximate function f is obtained by multi-resolution analysis that repeatedly extracts the wavelet coefficient of Equation 6 that is a low-pass component (Apploximation), the step of the road surface to be traveled is reflected while reflecting the step difference localized on the road surface and fine cracks. It can mainly reflect the macro characteristics of the full length.

On the other hand, when the approximation function f is obtained by wavelet packet analysis in which the scaling coefficient of Equation 5 which is a high-pass component (Detail) in addition to the low-pass component is extracted to the same extent as the low-pass component, steps and fine cracks in the entire length of the road surface to be traveled It is possible to reflect the micro-features of arbitrary components such as those using a high-pass component.
4 and 5 show specific examples of waveforms obtained when the wave characteristics are decomposed into waves of a plurality of types of frequencies by wavelet packet analysis.

4 shows the wave characteristics of the reference road surface (road surface having a step in a specific area) shown in FIG. 2, and FIG. 5 shows the wave characteristics of the reference road surface (road surface having no damage etc.) shown in FIG. It is a graph which shows the waveform obtained when it decomposes | disassembles into the wave of multiple types of frequency by analysis. In FIGS. 4 and 5, the waveform 1 and the waveform of each frequency represent the amplitude at the corresponding point as the height of the bar.
In FIG. 4 and FIG. 5, the raw data S before analysis wavy characteristics 1, first, are decomposed into a low-pass component A ₁ and a high-pass component D _1. Then, the low-pass component A ₁ is decomposed into a low-pass component A ₂ and a high-pass component D ₂ , the low-pass component A ₂ is further converted into a low-pass component A ₃ and a high-pass component D ₃ , and the low-pass component A ₃ is converted into a low-pass component A ₄ . decomposed into high-pass component D _4, technique using the low-pass component is multiresolution analysis. On the other hand, decomposes the high pass component _{D 1} to the low-pass component AD ₂ and a high-pass component DD _2, further decompose a high-pass component DD ₂ to the lowpass ADD ₃ and highpass DDD _3, degradation of the primary high-pass component _{D 1} Wavelet packet analysis is a technique that uses the components obtained in this way. However, the degree of decomposition may be determined as appropriate as long as the microscopic characteristics of the optional component can be reflected. For example, not only the high-pass component DD ₂ of the high-pass component D ₁ but also the low-pass component AAD ₃ and the high-pass component DAD ₃ obtained by decomposing the low-pass component AD ₂ are used, or the high-pass component D ₂ (DA ₂ ) is used. It is also possible to use a low-pass component ADA ₃ and a high-pass component DDA ₃ obtained by decomposition.

  4 and 5, the total length of the waveform obtained after the decomposition is half of the total length of the waveform before the decomposition, which is due to downsampling. A position corresponding to each data is a position when a zero value is inserted between each data of the waveform obtained after the decomposition, and the entire length is aligned with the waveform before the decomposition. That is, the low-pass component and the high-pass component do not correspond to the left half and the right half of the waveform before the decomposition, and both components correspond to the entire length of the waveform before the decomposition.

It is a block diagram which shows the concept of the specific example of the motor vehicle driving simulation apparatus concerning this invention. The example of the reference road surface data used for defining the data classification of the car driving simulation device is shown, (a) is the wave characteristic of the reference road surface, (b) is the wave characteristic decomposed into waves of plural kinds of frequencies. It is a graph which shows the waveform obtained by doing. The other example of the reference | standard road surface data used in order to define the data classification | category of the same vehicle driving simulation apparatus is shown, (a) is a wavy characteristic of a reference | standard road surface, (b) is the wave characteristic of several types of frequency. It is a graph which shows the waveform obtained by disassembling. It is a graph which shows the waveform obtained when the wavelike characteristic of the reference | standard road surface (road surface which has a level | step difference in a specific area | region) shown in FIG. 2 is decomposed | disassembled into the wave of several types of frequency by wavelet packet analysis. It is a graph which shows the waveform obtained when the wave-like characteristic of the reference | standard road surface (road surface which does not have any damage etc.) shown in FIG. 3 is decomposed | disassembled into the wave of several types of frequency by wavelet packet analysis.

Explanation of symbols

1 Wave characteristic a, b, c, d, e, f Frequency 2a, 2b, 2c, 2d, 2e, 2f Wave 3 Waveform 4 Local vibration part

Claims (5)

  The data on the road surface position and the road surface height corresponding to the position of the simulated driving are used to determine the wavy characteristics (1) of the reference road surface as the wave of multiple frequencies (a, b, c, d, e, f). (2a, 2b, 2c, 2d, 2e 2f), and the local vibration part (a) whose amplitude changes locally more greatly than other continuous parts appears in the waveform (3) obtained by the decomposition ( 4. A vehicle driving simulation device, wherein the vehicle driving simulation device is classified into the categories defined according to the characteristics of 4) and input.   The automobile driving simulation device according to claim 1, wherein the decomposition is performed by spatial frequency analysis.   The vehicle driving simulation apparatus according to claim 2, wherein discrete wavelets are used in the spatial frequency analysis.   The automobile driving simulation device according to claim 3, wherein the discrete wavelet is configured based on a scaling function, and the decomposition is performed by multi-resolution analysis.   The automobile driving simulation device according to claim 3, wherein the discrete wavelet is configured based on a scaling function, and the decomposition is performed by wavelet packet analysis.

Images