JP7095278B2 - Vehicle weight estimation device and vehicle weight estimation method - Google Patents

Description

The present disclosure relates to a vehicle weight estimation device and a vehicle weight estimation method, and more particularly to a vehicle weight estimation device and a vehicle weight estimation method for estimating the weight of a vehicle with high accuracy.

When the smoothed value of the estimated vehicle weight deviates from the preset range, a device for correcting the value so as to fall within the range has been proposed (see, for example, Patent Document 1). This device estimates the weight of a vehicle by using an upper limit value based on the maximum value of the vehicle weight and a lower limit value based on the minimum value of the vehicle weight as a range.

Japanese Unexamined Patent Publication No. 2004-301576

By the way, the estimated weight of the vehicle is used for controlling the running of the vehicle. Therefore, an error occurs in the parameters used for estimation due to the influence of noise and the like, and each time the estimated weight of the vehicle changes, the control regarding the running of the vehicle also changes. As such a change in control, a change in timing for changing the shift stage of the transmission is exemplified. This change in control gives the driver a sense of discomfort and is a factor that reduces drivability.

That is, in the device of the above example, even if the actual weight of the vehicle does not change, it is updated to the estimated weight of the vehicle, so that the control regarding the running of the vehicle may change frequently. There is sex.

The present disclosure has been made in view of the above, and an object thereof is a vehicle weight estimation device and a vehicle weight estimation method for estimating the weight of a vehicle with high accuracy according to the timing when the weight of the vehicle changes. To provide.

The vehicle weight estimation device of one aspect of the present invention that achieves the above object is a parameter acquisition means for acquiring parameters that change while the vehicle is running, and the weight of the vehicle based on the parameters input. When the difference between the estimated value and the preset reference value deviates from a predetermined range by the selection means including an estimation means for outputting the estimated value of the above and a selection means for inputting the estimated value. In addition, when an output value corresponding to the estimated value is selected and the difference falls within the predetermined range, the previous time, which is the output value output from the selection means before estimating the estimated value. It is characterized in that an output value is selected and either the selected output value or the previous output value is output.

In the vehicle weight estimation method of one aspect of the present invention that achieves the above object, the parameters that change during the running of the vehicle are acquired, and the estimated value is estimated and estimated as the weight of the vehicle based on the parameters. The difference between the estimated value and the preset reference value is calculated, it is determined whether or not the calculated difference is out of the predetermined range, and when it is determined that the difference is out of the predetermined range, the vehicle. When an output value corresponding to the estimated value is selected as the weight and it is determined that the difference falls within the predetermined range, the output value output as the weight of the vehicle before estimating the estimated value is used. It is characterized in that a certain previous output value is selected and either the selected output value or the previous output value is output.

According to one aspect of the present invention, the weight of the vehicle can be estimated with high accuracy according to the timing when the weight of the vehicle changes.

It is explanatory drawing which illustrates the 1st Embodiment of the vehicle weight estimation apparatus. It is a block diagram which illustrates the control device of FIG. It is a part of the block diagram illustrating the vehicle weight calculation unit of FIG. It is a part of the block diagram illustrating the vehicle weight calculation unit of FIG. It is a flow figure which illustrates the 1st Embodiment of the vehicle weight estimation method. It is a relational figure which illustrates the relationship between an engine rotation speed and a fuel injection amount, and an engine torque.

Hereinafter, embodiments of the vehicle weight estimation device and the vehicle weight estimation method will be described.

The vehicle weight estimation device 30 of the embodiment illustrated in FIGS. 1 to 4 is a device mounted on the vehicle 10 and estimates the weight of the vehicle 10 based on the parameters changing during the traveling of the vehicle 10. Hereinafter, those with x and y attached to the code are variables, and indicate the detected values and estimated values of the sensor, etc., and those with k attached to the code indicate the time series, and this k indicates "1" for each sampling period ts. "Increases by.

As illustrated in FIG. 1, in the vehicle 10 on which the vehicle weight estimation device 30 is mounted, the driver's cab (cab) 12 is arranged on the front side of the chassis 11, and the body 13 is arranged on the rear side of the chassis 11.

The engine 14, the clutch 15, the transmission 16, the propeller shaft 17, and the differential gear 18 are installed in the chassis 11. The rotational power of the engine 14 is transmitted to the transmission 16 via the clutch 15. The rotational power shifted by the transmission 16 is transmitted to the differential gear 18 through the propeller shaft 17 and distributed as driving force to the pair of driving wheels 19 which are the rear wheels. A torque converter may be interposed between the engine 14 and the clutch 15.

The control device 20 is electrically connected to the engine 14, the clutch 15, the transmission 16, and various sensors via a signal line indicated by an alternate long and short dash line. As various sensors, an accelerator opening sensor 22 that detects the accelerator opening Ax from the amount of depression of the accelerator pedal 21 and a position sensor 24 that detects the lever position Px of the shift lever 23 are installed in the driver's cab 12. A rotation speed sensor 25, a vehicle speed sensor 26, and an acceleration sensor 27 for detecting the rotation speed Nx of the crankshaft (not shown) of the engine 14 are installed in the chassis 11.

The control device 20 is hardware composed of a CPU that performs various information processing, an internal storage device that can read and write programs and information processing results used for performing the various information processing, and various interfaces.

As illustrated in FIG. 2, the control device 20 has a control unit 28 for controlling an engine 14, a clutch 15, and a transmission 16 and a vehicle weight calculation unit 31 for calculating the weight of the vehicle 10 as functional elements. .. In this embodiment, each functional element is stored as a program in the internal storage device, but each functional element may be configured by individual hardware.

The vehicle weight estimation device 30 includes a position sensor 24, a rotation speed sensor 25, a vehicle speed sensor 26, an acceleration sensor 27, a control unit 28, and a vehicle weight calculation unit 31. The vehicle weight calculation unit 31 functions as a parameter acquisition means, an estimation means, and a selection means, and the detection values of those sensors and the calculation results of the calculation unit are input, and the calculation results are calculated based on each detection value and the calculation result. Is output as the output value mx.

The control unit 28 is a functional element that functions as a part of the parameter acquisition means, and in this embodiment, the fuel injection amount Qx in the engine 14 is acquired for each sampling cycle ts. Since the fuel injection amount Qx is proportional to the injection time (drive pulse) of the injector (not shown) of the engine 14, it can be obtained from the total value of the injection times.

The control unit 28 inputs the accelerator opening degree Ax detected by the accelerator opening degree sensor 22, and calculates the reference injection time based on the accelerator opening degree Ax. Next, the control unit 28 is added based on whether or not the in-vehicle device mounted on the vehicle 10 and driven by the engine 14 is driven, and whether or not the exhaust gas purifying device that purifies the exhaust gas discharged from the engine 14 is regenerated. Calculate the injection time. When the fuel injection amount based on this additional injection time is injected, the driving force of the in-vehicle device is supplemented or the exhaust gas purification device is regenerated. Next, the control unit 28 calculates the fuel injection amount Qx based on the total value of the reference injection time and the additional injection time for each sampling cycle ts.

The control unit 28 is not limited to this configuration as long as it can acquire the fuel injection amount Qx actually injected in the engine 14. The control unit 28 may calculate the reference injection time from the internal pressure, volumetric efficiency, and required air-fuel ratio of the intake manifold (not shown), or from the intake air amount and the engine rotation speed Nx. Examples of the in-vehicle device include an air compressor and a motor generator. As the exhaust gas purification device, a collection filter that collects particulate matter in the exhaust gas can be exemplified.

The position sensor 24 is a device that functions as a part of the parameter acquisition means, and detects the lever position Px requested by the driver by electrically detecting the position of the shift lever 23 operated by the driver of the vehicle 10. do. The position sensor 24 detects the gear ratio ix of the transmission 16 according to the lever position Px for each sampling cycle ts. Examples of the lever position Px include a parking position (P position), a reverse position (R position), a neutral position (N position), and a forward position (D range). As the forward position, for example, a plurality of 1st to 6th speeds are set. In each forward position, a gear ratio ix that decreases as the number of gears increases from the first gear is set. When the transmission 16 is composed of AMT, one having a function of reading the shift stage of the transmission 16 controlled by the control unit 28 may be used instead of the position sensor 24. Further, when the gear ratio ix of the transmission 16 is acquired from the control signal of the control unit 28, the gear ratio ix can be obtained based on the vehicle speed vx and the engine rotation speed Nx.

The vehicle speed sensor 26 is a device that functions as a part of the parameter acquisition means, reads a pulse signal proportional to the rotation speed of the propeller shaft 17, and calculates the vehicle speed vx for each sampling cycle ts by the vehicle speed calculation process (not shown) of the control device 20. It is a sensor to acquire. Since the vehicle speed sensor 26 acquires the vehicle speed vx based on the pulse signal proportional to the rotation speed, the acquired vehicle speed vx is not a negative value but a value of zero or more. As the vehicle speed sensor 26, a sensor that acquires the vehicle speed vx from the rotation speed of the output shaft, the drive wheel 19, the driven wheel, etc. (not shown) of the transmission 16 may be used. When a sensor that acquires the vehicle speed vx from the rotation speeds of the driving wheels 19, the driven wheels, and the like is used, it is preferable to acquire the rotation speeds of each of the pair of left and right wheels and use the average value as the vehicle speed vx.

The acceleration sensor 27 is a device that functions as a part of the parameter acquisition means, and operates by an acceleration component accompanying a speed change in the front-rear direction of the vehicle 10 and a gravitational acceleration component accompanying a posture change of the vehicle 10. The acceleration sensor 27 is a sensor that acquires an acceleration component parallel to the road surface, that is, an acceleration Gx in the front-rear direction of the vehicle 10, for each sampling period ts. Examples of the acceleration sensor 27 include a mechanical displacement measurement method, an optical method, and a semiconductor method.

As illustrated in FIGS. 3 and 4, in this embodiment, the vehicle weight calculation unit 31 has a parameter acquisition unit 32, an estimation unit 33, and a selection unit 34 as functional elements. Each functional element of the vehicle weight calculation unit 31 is stored in the internal storage device as a program, but each functional element may be configured by individual hardware.

The parameter acquisition unit 32 functions as a parameter acquisition means, and the detection values of the control unit 28 and each sensor are input. The parameter acquisition unit 32 is a functional element that outputs the first parameter Φx and the second parameter Φy as parameters that change while the vehicle 10 is traveling in each sampling cycle ts to the estimation unit 33. The parameter acquisition unit 32 has a first parameter calculation block 32a, a second parameter calculation block 32b, and an engine torque calculation block 32c.

The first parameter calculation block 32a is a functional element in which the acceleration Gx is input and the first parameter Φx is calculated. The second parameter calculation block 32b is a functional element for calculating the second parameter Φy by inputting the vehicle speed vx, the gear ratio ix of the transmission 16, and the engine torque Te. The engine torque calculation block 32c is a functional element that calculates the engine torque Te that is actually output from the engine 14 by inputting the engine rotation speed Nx and the fuel injection amount Qx.

The estimation unit 33 functions as an estimation means, and the first parameter Φx and the second parameter Φy are input to them and the previously estimated value mx (k-1) estimated immediately before in the sampling cycle ts. Based on this, it is a functional element that outputs an estimated value mx (k) estimated by using the smoothing process. The estimation unit 33 is composed of an RLS estimation block. The RLS estimation block updates the variables in the estimation operation every sampling period ts. That is, the RLS estimation block has the previously estimated value mx (k-1), the covariance matrix P (k-1), and the gain K (k-1) calculated by the RLS algorithm when a new parameter is input. It is in a memorized state.

The selection unit 34 functions as a selection means, and the estimated value mx (k) output from the estimation unit 33 is sequentially input. The selection unit 34 selects and outputs either the estimated value mx (k) or the previous output value m (x-1) as the output value mx based on the difference between the estimated value mx (k) and the reference value ma. It is a functional element to do.

The difference between the estimated value mx (k) and the reference value ma is the integrated value Σmx ( k). That is, the selection unit 34 sets the change amount Δmx (k), which is the difference between the input estimated value mx (k) and the previous estimated value mx (k-1), into the previous estimated value mx (k-1) and the reference value ma. It is a functional element that calculates the integrated value Σmx (k) integrated into the difference Δmx (k-1) of.

The reference value ma is a value updated at the timing when the estimated value mx (k) is output as the output value mx. As the initial value of the reference value ma, the reference vehicle weight W0 is exemplified. As the reference vehicle weight W0, the vehicle weight which is the weight of the vehicle 10 when the vehicle is empty, that is, the weight of the main body (chassis 11, driver's cab 12, body 13) of the vehicle 10 (including the engine 14 and the like) is exemplified. The vehicle weight may include the weight of fuel, oil, cooling water, spare tire, tool, and the like. Further, as the reference value ma, the total weight of the vehicle, that is, the weight obtained by adding the total weight of the maximum passenger capacity and the weight of the load having the maximum load capacity to the vehicle weight is also exemplified.

In this embodiment, the selection unit 34 includes a previous estimation value acquisition block (delay block) 34a, an addition block 34b, an integration block (also referred to as an integrator or an integrator) 34c, an absolute value block 34d, a comparison block 34e, and a switch block 34f. It also has a previous output value acquisition block (delay block) of 34 g.

The previous estimated value acquisition block 34a, the addition block 34b, and the integration block 34c are functional elements for calculating the difference between the input estimated value mx (k) and the reference value ma, and the integrated value Σmx (k) is used as the difference. It is a functional element to be calculated. Specifically, the previous estimated value acquisition block 34a, the addition block 34b, and the integration block 34c are the differences between the estimated value mx (k-1) and the previous estimated value mx (k-1) at regular intervals (sampling time). It is a functional element that calculates a certain change amount Δmx (k) and calculates an integrated value Σmx (k) by integrating the calculated change amount Δmx (k).

The amount of change Δmx (k) is the difference between the acquired current estimated value mx (k) and the previously acquired previously estimated value mx (k-1). As the amount of change Δmx (k), for example, the amount of change per fixed period (sampling time) may be used. The amount of change Δmx (k) is positive (+) when it increases from the previous estimated value mx (k-1) to the estimated value mx (k), and negative (-) when it decreases. The integrated value Σmx (k) is the sum of the change amount Δmx (k) calculated for each fixed cycle (sampling time), and the change amount Δmx (k) for each fixed cycle from the reset to the present time is sequentially sequentially calculated. Calculated by adding. For example, when reset in the previous cycle, the integrated value Σmx (k) becomes the change amount Δmx (k), and when it has never been reset, the integrated value Σmx (k) changes from the change amount Δmx (1) to the change amount. It is the sum of Δmx (k). When the binary signal "1" started from the comparison block 34e is input, the integration block 34c resets the integrated value Σmx (k), that is, sets it to zero ("0"). In addition, resetting the integrated value Σmx (k) substantially updates the reference value ma to the estimated value mx (k), that is, sets the estimated value mx (k) to the next reference value ma. Is synonymous with.

The absolute value block 34d and the comparison block 34e are functional elements for determining whether or not the integrated value Σmx (k) is out of the error range (−Wa to + Wa). The comparison block 34e transmits "1" (a signal indicating true) as a binary signal when the integrated value Σmx (k) is out of the error range (| ΣWx |> Wa). On the other hand, when the integrated value Σmx (k) falls within the error range (| ΣWx | ≦ Wa), "0" (a signal indicating false) is transmitted as a binary signal. In addition, in this specification, when the integrated value Σmx (k) is a value of + Wa or -Wa, it is assumed that it is within the range of error.

Next, the estimation calculation of the estimated value mx (k) by the estimation unit 33 of this embodiment will be described. Based on each parameter and the previous estimated value mx (k-1), the estimation unit 33 regards the equation of motion in the front-rear direction of the vehicle 10 as a transfer function, and uses an adaptive algorithm as a smoothing process to estimate the value mx. (K) is estimated. As an adaptive algorithm, an RLS (Recursive Last Squares) algorithm (sequential least squares algorithm) is used.

The equation of motion of the vehicle 10 in the front-rear direction is expressed by the following equation (1). In the formula (1), vx'is a derivative value obtained by time-differentiating the vehicle speed vx, Tw is the drive torque transmitted to the drive wheel 19, rw is the wheel diameter of the drive wheel 19, and Δmx is the mass corresponding to the rotating portion described later. , B is a constant, g is a gravitational acceleration, and μ is a rolling resistance coefficient. The constant B is a constant multiplied by "0.5", the air density ρ, the front projected area Af of the vehicle 10, and the air resistance coefficient Cd. The wheel diameter rw, the constant B, and the rolling resistance coefficient μ are obtained as values peculiar to the vehicle 10.

上記の数式（１）を変形すると、車両１０の重量ｍｘは、下記の数式（２）に表される。

By modifying the above formula (1), the weight mx of the vehicle 10 is expressed by the following formula (2).

上記の数式（２）は、第一パラメータΦｘを入力値、第二パラメータΦｙを出力値、推定値ｍｘを変数とした伝達関数として見做せる。そこで、その伝達関数（Φｙ（ｋ）＝Φｘ（ｋ）・ｍｘ（ｋ））を、ＲＬＳアルゴリズムに従って自己適応させて、推定値ｍｘ（ｋ）を推定する。推定値ｍｘ（ｋ）は以下の数式（３）～（５）で表される。以下の数式で、ｍｘ（ｋ－１）はサンプリング周期ｔｓにおける一つ前に推定した推定値である前回推定値を、Ｋ（ｋ）はＲＬＳアルゴリズムで計算されるゲインを、Ｐ（ｋ）は共分散行列を、Ｉは単位行列を、「^T」は転置行列をそれぞれ示している。

The above formula (2) can be regarded as a transfer function in which the first parameter Φx is an input value, the second parameter Φy is an output value, and the estimated value mx is a variable. Therefore, the transfer function (Φy (k) = Φx (k) · mx (k)) is self-adapted according to the RLS algorithm to estimate the estimated value mx (k). The estimated value mx (k) is expressed by the following mathematical formulas (3) to (5). In the following formula, mx (k-1) is the previous estimated value estimated immediately before in the sampling period ts, K (k) is the gain calculated by the RLS algorithm, and P (k) is. The covariance matrix, I represents the identity matrix, and " ^T " represents the transposed matrix.

共分散行列Ｐ（ｋ）の初期値Ｐ（０）を定めれば、パラメータΦｘにより、上記の数式（５）に基づいて共分散行列Ｐ（ｋ）を、及び数式（４）に基づいてゲインＫ（ｋ）をそれぞれ算出できる。つまり、新しく第一パラメータΦｘ及び第二パラメータΦｙが得られる度に、共分散行列Ｐ（ｋ）とゲインＫ（ｋ）を新たに更新する。そして、それらと数式（３）に基づいて、直前に推定した前回推定値ｍｘ（ｋ－１）を修正していく方式で推定
値ｍｘ（ｋ）を算出できる。

If the initial value P (0) of the covariance matrix P (k) is determined, the covariance matrix P (k) is obtained based on the above equation (5) and the gain is obtained based on the equation (4) according to the parameter Φx. K (k) can be calculated respectively. That is, every time the first parameter Φx and the second parameter Φy are newly obtained, the covariance matrix P (k) and the gain K (k) are newly updated. Then, based on these and the mathematical formula (3), the estimated value mx (k) can be calculated by a method of modifying the previously estimated value mx (k-1) estimated immediately before.

The initial value P (0) is represented by the product of the constant α and the identity matrix I. As the constant α, a value of about 1000 is usually used, but when the noise is large, the constant α may be set small. This constant α is determined by the magnitude of noise. Further, as the initial value m (0), for example, the vehicle weight when the vehicle is empty excluding the driver and the load, the gross vehicle weight when the maximum load is applied, or the average value of the estimated values mx (k) may be used.

When the covariance matrix P (k) becomes large, the estimated value mx (k) moves away from the true value, and when the covariance matrix P (k) converges small, the estimated value mx (k) approaches the true value.

As described above, by considering the above mathematical formula (2) as a transfer function and estimating the estimated value mx (k) by the adaptive algorithm, the estimated value mx (k) can be sequentially smoothed. This is advantageous for speeding up the convergence to the true value and improving the robustness against noise, disturbance, or changes in the statistical properties of the detected values of each sensor, and the estimation error can be reduced. Along with this, the weight of the vehicle 10 can be estimated with high accuracy.

In this embodiment, the estimated value mx (k) can be obtained from the above mathematical formulas (3) to (5) by using the RLS algorithm among the adaptive algorithms. This is advantageous for online estimation, and the estimated value mx (k) can be calculated in real time. Further, as compared with the method of removing noise by a low-pass filter for the detected value acquired by each sensor, it is advantageous for ensuring the responsiveness of the estimation of the weight of the vehicle 10.

In addition, it is only necessary to update the previously estimated value mx (k-1), the covariance matrix P (k-1), and the gain K (k-1) for each sampling period ts, and the internal storage device of the control device 20 can be used. The number to be memorized can be minimized. Therefore, the storage required for estimation is compared with the method by offline estimation (batch processing estimation) in which an infinite storage area must be secured in the internal storage device of the control device 20 for an uncertain traveling period of the vehicle 10. It is advantageous for reducing the capacity. The offline estimation referred to here can be exemplified by a batch processing least squares method, a method of calculating the average value of all estimated values mx (0) to mx (k), and the like.

Further, by using the RLS algorithm, the estimated value mx (k) can be calculated for each sampling period ts. This is advantageous for real-time estimation as compared with the method of estimating when the state of the vehicle 10 (for example, gear ratio ix or drive torque Tw) changes or when the vehicle travels a predetermined distance.

Next, the vehicle weight estimation method will be described as each function of the vehicle weight calculation unit 31 with reference to the flow chart of FIG. The following vehicle weight estimation method is started when the control device 20 of the vehicle 10 is energized, and is repeatedly performed every sampling cycle ts to estimate the weight of the vehicle 10 in real time. That is, processing from the start to the return is performed in one sampling period ts. Then, when the control device 20 loses power, it ends.

When started, the vehicle weight calculation unit 31 acquires parameters that change while the vehicle 10 is traveling by the function of the parameter acquisition unit 32 (S110). The parameters are the first parameter Φx and the second parameter Φy.

Specifically, the parameter acquisition unit 32 acquires those parameters from the detection values detected by the control unit 28 and each sensor. First, the control unit 28 determines the fuel injection amount Qx, the position sensor 24 determines the gear ratio ix of the transmission 16, the rotation speed sensor 25 determines the engine rotation speed Nx, the vehicle speed sensor 26 determines the vehicle speed vx, and the acceleration sensor 27 determines the acceleration Gx. Get each.

Next, the first parameter calculation block 32a calculates the first parameter Φx shown in the following mathematical formula (6) by each block.

加速度Ｇｘは、上述したとおり車両１０の前後方向での速度変化に伴う加速度成分と車両１０の姿勢変化に伴う重力加速度成分とを合成した路面に平行な加速度成分である。つまり、加速度Ｇｘは、微分値ｖｘ’と重力加速度成分ｇ・ｓｉｎβとを加算した値になる。したがって、数式（６）は、上記の数式（２）の分子と同義である。

As described above, the acceleration Gx is an acceleration component parallel to the road surface, which is a combination of an acceleration component accompanying a speed change in the front-rear direction of the vehicle 10 and a gravitational acceleration component accompanying a posture change of the vehicle 10. That is, the acceleration Gx is a value obtained by adding the differential value vx'and the gravitational acceleration component g · sin β. Therefore, the mathematical formula (6) is synonymous with the numerator of the above mathematical formula (2).

As the variables of the first parameter Φx, as shown in the above formula (2), instead of the acceleration Gx, the differential value vx'of the vehicle speed vx and the gravitational acceleration component based on the road surface gradient on which the vehicle 10 is traveling. You may use g · sinβ. In this case, instead of the acceleration sensor 27, it is preferable to use a vehicle speed sensor 26, a gradient sensor that acquires the road surface gradient on which the vehicle 10 is traveling, and a functional element that calculates the road surface gradient.

Next, the engine torque calculation block 32c calculates the actual engine torque Te output from the engine 14 based on the fuel injection amount Qx and the engine rotation speed Nx.

As illustrated in FIG. 6, the engine torque Te output from the engine 14 has a positive relationship with each of the engine rotation speed Nx and the fuel injection amount Qx, and the engine rotation speed Nx is high and the fuel injection amount Qx. The larger the value, the larger the value. This map data is obtained in advance by experiments and tests, and is stored in the engine torque calculation block 32c, which is a data block.

In this embodiment, the engine torque Te is calculated from the relationship between the engine rotation speed Nx and the fuel injection amount Qx, but the accelerator opening Ax acquired by the accelerator opening sensor 22 may be used instead of the fuel injection amount Qx. , Other acquisition methods may be used.

Next, the second parameter calculation block 32b calculates the drive torque Tw transmitted to the drive wheels 19 by using the following mathematical formula (7). In the formula (7), if indicates the gear ratio of the differential gear 18, and η indicates the transmission efficiency different depending on the gear ratio. The drive torque Tw may be obtained by using a torque sensor or by another method.

次いで、第二パラメータ算出ブロック３２ｂは、ルックアップテーブルブロック３２ｄにより回転部分相当質量Δｍｘを算出する。

Next, the second parameter calculation block 32b calculates the rotating portion equivalent mass Δmx by the look-up table block 32d.

The mass corresponding to the rotating portion Δmx is a value determined according to the gear ratio ix which is a variable. The lookup table block 32d is set with a plurality of masses Δmx corresponding to a plurality of rotating portions for each gear ratio ix, and the one corresponding to the gear ratio ix is selected. The mass corresponding to the rotating portion Δmx may be calculated from the relationship with the vehicle weight when the vehicle is empty, the gear ratio ix, and a predetermined coefficient.

Next, the second parameter calculation block 32b calculates the second parameter Φy shown in the following mathematical formula (8) by each block.

この実施形態では、第二パラメータΦｙを上記の数式（８）で示したが、エンジン１４、クラッチ１５、トランスミッション１６、ディファレンシャルギア１８などに働く摩擦トルクＴｆを考慮してもよい。この場合は、駆動トルクＴｗから摩擦トルクＴｆを減算した値を駆動輪１９の車輪径ｒｗで除算するとよい。摩擦トルクＴｆを考慮すると、推定精度の向上には有利になる。

In this embodiment, the second parameter Φy is shown by the above equation (8), but the friction torque Tf acting on the engine 14, the clutch 15, the transmission 16, the differential gear 18, and the like may be taken into consideration. In this case, the value obtained by subtracting the friction torque Tf from the drive torque Tw may be divided by the wheel diameter rw of the drive wheel 19. Considering the friction torque Tf, it is advantageous for improving the estimation accuracy.

When each parameter is acquired as described above, the vehicle weight calculation unit 31 estimates the estimated value mx (k) by the above-mentioned estimation method of the estimation unit 33 (S120).

Next, the vehicle weight calculation unit 31 has the estimated value mx (k) input by the selection unit 34 and the previous estimated value mx (k) which is a value input immediately before the estimated value mx (k) is input. The amount of change Δmx (k), which is the difference from -1), is calculated (S130). Specifically, in this step, the amount of change Δmx (k) is calculated by the previous estimated value acquisition block 34a and the addition block 34b. The amount of change Δmx (k) may be calculated as the amount of change in the estimated value mx (k) per unit time. Further, the previous estimated value mx (k-1) immediately after the control device 20 is energized may be an estimated value input immediately before the control device 20 is stopped, and is a reference which is an initial value of the above-mentioned reference value ma. The vehicle weight may be W0.

Next, the vehicle weight calculation unit 31 uses the selection unit 34 to integrate the change over time from the reference value ma up to the time when the estimated value mx (k) is input to the estimated value mx (k) Σmx ( k) is calculated (S140). Specifically, in this step, the amount of change calculated in step S130 from the previous integrated value Σmx (k-1), which is the difference between the previous estimated value mx (k-1) and the reference value ma, by the integration block 34c. The integrated value Σmx (k) is calculated by adding Δmx (k). That is, when the previous integrated value Σmx (k-1) is zero (“0”), the integrated value Σmx (k) calculated in this step is the amount of change Δmx (k).

Next, the vehicle weight calculation unit 31 determines whether or not the integrated value Σmx (k) is out of the error range (−Wa to + Wa) by the selection unit 34 (S150). In this step, when the absolute value of the integrated value Σmx (k) exceeds the numerical value Wa in the error range by the selection unit 34, the process proceeds to the step of outputting the tentative estimated value mx (k) as the output value mx. On the other hand, if the absolute value of the integrated value Σmx (k) is equal to or less than the numerical value Wa in the error range, the process proceeds to the step of maintaining the previous output value m (x-1).

When the absolute value of the integrated value Σmx (k) exceeds the numerical value Wa in the error range, the vehicle weight calculation unit 31 selects the estimated value mx (k) by the selection unit 34 (S160). Next, the vehicle weight calculation unit 31 resets the integrated value Σmx (k) by the selection unit 34 (S170). In this step, the integrated value Σmx (k) is reset to zero, so that the reference value ma is substantially updated to the estimated value mx (k). Next, the vehicle weight calculation unit 31 outputs the estimated value mx (k) selected by the selection unit 34 as the output value mx (S180), and returns to the start.

On the other hand, when the absolute value of the integrated value Σmx (k) is equal to or less than the numerical value Wa in the error range, the vehicle weight calculation unit 31 selects the previous output value m (x-1) by the selection unit 34 (S190). .. Next, the vehicle weight calculation unit 31 outputs the previous output value m (x-1) selected by the selection unit 34 as the output value mx (S180), and returns to the start.

Specifically, the selection unit 34 determines whether or not the absolute value of the integrated value Σmx (k) exceeds the numerical value Wa in the error range by the comparison block 34e (S150). When the absolute value of the integrated value Σmx (k) exceeds the numerical value Wa in the error range, the estimated value mx ( k) is selected (S160). Next, the integrated value Σmx (k) is reset by the integrated block 34c to which the binary signal "1" started from the comparison block 34e is input (S170).

On the other hand, when the absolute value of the integrated value Σmx (k) is equal to or less than the numerical value Wa in the error range, the previous output value m (x-1) is selected by the switch block 34f (S190).

As described above, the vehicle weight calculation unit 31 considers the time when the integrated value Σmx (k) is out of the error range (-Wa to + Wa) as the timing when the weight of the vehicle changes, and estimates at that time. The estimated value mx (k) is output as the output value mx. On the other hand, the vehicle weight calculation unit 31 regards the time when the integrated value Σmx (k) falls within the error range (-Wa to + Wa) as the time when the weight of the vehicle has not changed, and at that time, the previous output value. Maintain the output of m (x-1).

That is, the vehicle weight calculation unit 31 eliminates the error in sensing, determines the timing when the weight of the vehicle 10 changes, and outputs the output value mx to suppress the frequency of updating the weight of the vehicle 10. Will be advantageous. Further, even if the estimated value mx (k) gradually changes within the error range, the change is judged by the integrated value Σmx (k) to avoid the situation where the change in the weight of the vehicle 10 is overlooked. Will be advantageous. In this way, the vehicle weight calculation unit 31 can estimate the weight of the vehicle 10 with high accuracy according to the timing when the weight of the vehicle 10 changes, so that the driver feels uncomfortable due to the fluctuation of the control using the weight. It is possible to improve drivability without having to.

For example, when the control using the weight of the vehicle 10 is the shift control of the transmission 16, when the weight of the vehicle 10 is updated, the shift timing in the transmission 16 changes with the update. Therefore, the vehicle weight calculation unit 31 of the embodiment updates the output value mx when the weight of the vehicle 10 actually changes. Therefore, since the shift timing can be changed according to the change in the weight of the vehicle 10, the drivability can be improved without giving the driver a sense of discomfort due to the change in the shift timing.

Further, the vehicle weight calculation unit 31 outputs the previous output value m (x-1) when the integrated value Σmx (k) falls within the error range, that is, when the weight of the vehicle does not actually change. Therefore, it is advantageous to reduce the estimation error of the vehicle weight.

The vehicle weight calculation unit 31 uses an error range as a threshold value for the integrated value Σmx (k), but a range other than the error range may be used as long as it can be determined that the weight of the vehicle has changed. .. However, by using the error range, it is possible to eliminate the influence of the error due to the accuracy and sensitivity of each sensor and the error due to the vibration of the sensor caused by the vibration of the vehicle 10, and the estimation error of the road surface gradient is reduced. It will be advantageous to.

In the above-described embodiment, a configuration in which the integrated value Σmx (k) is calculated as the difference between the estimated value mx (k) and the reference value ma is illustrated, but the estimated value when the reference value ma is reset each time. It may be configured to update to mx (k), and the difference between the estimated value mx (k) and the reference value ma may be calculated at regular intervals (sampling time).

In this embodiment, an example in which the RLS algorithm is used as the estimation calculation of the estimated value mx (k) has been described, but the estimation unit 33 only needs to be able to estimate the weight of the vehicle 10, and the estimation calculation is not limited to this. For example, instead of the RLS algorithm, an LMS (Least Mean Square) algorithm, an NLMS (Nomalized Last Mean Square) algorithm, or the like may be used as an adaptive algorithm. Further, an averaging process may be performed to output the average value of all the estimated estimated values mx (0) to the estimated value mx (k). The averaging process is a specific pattern of the smoothing process.

The tentative estimated value mx (k) may be calculated by a simple method, and is not limited to the method using only the equation of motion in the front-rear direction of the vehicle 10 shown in the above mathematical formula (2). For example, when the vehicle 10 is equipped with an air suspension, a method based on a change in the vehicle 10 in the vertical direction may be used. Further, a method based on the torque input to the transmission before and after the shift and the amount of change in the number of revolutions output from the transmission may be used. Further, the tentative estimated value Mx may be a value obtained by adding the weight of the body 13 due to the change in the load capacity to the value acquired by a weight sensor such as a load cell and the weight of the vehicle when the vehicle is empty.

Further, in the above-described embodiment, an example in which the vehicle weight estimation device 30 is composed of the vehicle weight calculation unit 31 and each sensor has been described, but the present disclosure is not limited to this. For example, the vehicle weight estimation device 30 may be composed of one sensor that functions as a parameter acquisition means and a provisional estimation means, and hardware that functions as an estimation means and a maintenance means.

10 Vehicle 24 Position sensor 25 Rotation speed sensor 26 Vehicle speed sensor 27 Acceleration sensor 28 Control unit 30 Vehicle weight estimation device 31 Vehicle weight calculation unit 32 Parameter acquisition unit 33 Estimating unit 34 Selection unit Φx First parameter Φy Second parameter mx (k) Estimated value mx (k-1) Previous output value mx Output value ma Reference value Δmx (k) Change amount Σmx (k) Difference between estimated value and reference value (integrated value of change amount)

Claims (8)

Parameter acquisition means for acquiring parameters that change while the vehicle is running, and
An estimation means in which the parameter is input and an estimated value of the weight of the vehicle based on the parameter is output.
With a selection means in which the estimated value is input,
When the difference between the estimated value and the preset reference value deviates from a predetermined range by the selection means, an output value corresponding to the estimated value is selected, and the difference falls within the predetermined range. In this case, the previous output value, which is the output value output from the selection means, is selected before the estimated value is estimated, and either the selected output value or the previous output value is output. A vehicle weight estimation device characterized by having such a configuration. The first aspect of the present invention is the means for calculating the difference between the estimated value and the reference value as an integrated value of changes from the reference value to the estimated value until the estimated value is input. Vehicle weight estimation device. The selection means according to any one of claims 1 or 2, which is a means for setting the estimated value to the next reference value when the difference between the estimated value and the reference value deviates from the predetermined range. The vehicle weight estimation device described. The vehicle weight estimation device according to claim 2, wherein the selection means is a means for setting the integrated value to zero when it is determined that the integrated value is out of the predetermined range. The estimation means is a means for estimating the estimated value within a preset error range.
The vehicle weight estimation device according to any one of claims 1 to 4, wherein the predetermined range is set within the error range. The estimation means is a means for estimating the estimated value by using a smoothing process based on the input parameter and the previous estimated value which is an estimated value estimated before acquiring the parameter. The vehicle weight estimation device according to any one of 5 to 5. The estimation means is a means for estimating the estimated value by using an adaptive algorithm as the smoothing process by regarding the equation of motion in the front-rear direction of the vehicle as a transmission function based on the parameter and the previous estimated value. The vehicle weight estimation device according to claim 6. Get the parameters that change while the vehicle is running,
Based on the parameters, an estimated value as the weight of the vehicle is estimated.
Calculate the difference between the estimated estimated value and the preset reference value,
It is determined whether or not the calculated difference is out of the predetermined range, and the difference is determined.
When it is determined that the difference is out of the predetermined range, an output value corresponding to the estimated value is selected as the weight of the vehicle.
When it is determined that the difference is within the predetermined range, the previous output value, which is the output value output as the weight of the vehicle before estimating the estimated value, is selected.
A vehicle weight estimation method comprising outputting either the selected output value or the previous output value.

Images