!294376 九、發明說明： 【發明所屬之技術領域】 本發明係涉及一種車輛動態預測系統與方法，尤其是關 於利用三個量測之物理量，搭配利用數值演算法則之一估測 器以推算出相關之物理量，並利用利用數值演算法則之一預 測器接受這些物理量而可預測車輛之動態行為。 【先前技術】 對於車輛的行駛狀況，較先進的車輛電子系統皆會配備 有車輛翻覆警告裝置，用以預測車輛於何種環境、何種駕駛 情況之下，可能造成車輛翻覆，而可適時的提出警告或是直 接予以心正。相關技術如美國專利公報公告號第6002974號 專利係利用側傾角感測器、俯仰角速度感測器、縱向加速規、 側向加速規與垂直加速規，並利用擴增卡曼濾波器 (Extended Kalman Filter)來估測目前的側傾角與俯 仰角，且再利用側傾角速度與側傾角套用至經驗加權函數來 預測未來的側傾角，同樣地利用俯仰角速度與俯仰角套用至 經驗加權函數來預測未來的俯仰角。如第61923〇5號專利採 用側傾角感、俯仰祕度感·、縱向加速規、側向加 速規與垂直加速規，並_擴增卡曼濾波器來估測目前的側 傾角、俯仰角，再利用估測的側向加速度與縱向速度套用至 經驗公式來麟由麵獅所產生之偏差值，並此偏差 5 1294376 值與側傾角套用至經驗加權函數來預測未來的侧傾角，而此 偏差值與俯仰角套用至經驗加權函數來預測未來的俯仰角。 如第6556908號專利感測器的選用採用緃向加速規、側向加 速規、橫擺角感測器、側傾角速度感測器與輪胎速度感測器， 經由側向加速度與側傾角速度套用至經驗加權函數來算出過 去與目前之側傾角速度，並藉由另—經驗加權函數來運用過 去與目前之·肖速絲計算其鎖校正後之麵梯度，並 藉由此側傾梯賴出較為轉性的側則以判斷翻覆事件。 如第6631317號專利感測器包含銳向加速規、側向加速規、 橫擺角感測H、側则速度感廳與輪胎速度感測器，接著 側向加速規、㉝向加速規、麵肖速度❹指、橫擺角速度 感測器與輪胎速度感測H所測得之物理量細至簡化公式來 決定暫制則與敎麵肖，_這_值與兩個數值運 算濾波器來估測出側傾角以判斷翻覆事件。 但上述專利皆利用四至五種感測器方能估算出車輛的 側傾角’而後藉由經驗加權函數或是簡化公式來預測車輛侧 傾角，並以此物理量來宣告是否翻覆，'然而所預測之車輛側 傾角皆無考慮道路狀況之影響。 【發明内容】 繁於以上的問題’本發明所欲解決之問題在於提供一種 車輛動態預_統與方法，藉以_三域測器所得之物理 1294376 車輛於未 斷車輛是 蝴型來得出 心，糟由車輛於未來時間之側傾角來判 货f翻覆。 本發明揭露it上述Γ種車輛域_統的技術問題， 來時門之、Μ 她__、統，_以_—車輛於未 估、ρίΙ 崎為’此車輛動態預測系統包含-感測部、— L與—預測器’其中感測部由下航件組成： …-第-感測器係測量車輛之質心的縱向速度，並產生— 7向速度信號；—第二感測器係測量車輛之質心的側向加速 X JL產生側向加速度信號及一第三感測器係測量車輛之 懸掛彈簧之長度變化，並產生—長度變化信號。 …二估測器包含—由一車輛動態模型所建構之非線性狀 心Ιτ、’且接观向速度錢、側向加速度錢與長度變 化信號’並經由麵性狀驗察ϋ而估_車輛之一側傾 角、-側傾角速度、—側傾角加速度、—俯仰角、—俯仰角 速度、-俯仰角加速度、—紐肖、—橫擺肢度、一橫擺 角加速度、-縱向位移、一縱向加速度、—側向位移、一侧 向速度、-垂直位移、—垂直速度、—垂直加速度與一輪胎 角速度。而預測器包含—車輛動態模型，且接受縱向速度信 號、側向加速度信號、長度變化信號、側傾角、側傾角速度、 側傾角加速度、俯仰角、俯仰角速度、俯仰角加速度、橫擺 1294376 角、橫速度、侧加速度、縱向位移、縱向加速度、 於i向位私側向速度、垂直位移、垂直速度、垂直加速度盘 輪胎角速度，而預測車胁未來時間之運動行為。” 車輛_侧她^咖，翻義角或俯 :角，亦或同時接收兩者，而估算側傾角或俯仰角之變化趨 勢而判斷車婦未料暇__域，域算出未來車輛 。其中侧傾角或俯仰角之變化趨勢係為側傾角或 角疋否為發散函數’而_車輛於未來時間是否翻覆狀 二=斷賴增估算讀理量嘴出未來車輛 ^且第二感測器用以測量車輛之質心的側向加速度，可 訊=里車輛之質心的橫擺角速度代替，並產生-橫擺角速度 ’亦可使估測器產生相同的作用。另亦可用測量車翻之 =的檢擺_侧向加速度，並產生—觀角速度 冋樣可使細指產生彳目_侧。 於·為解决上述—種車輛賴制方法的技術問題，本發明 之運種車_<%糊方法，制以細—車輛於未來時間 動仃為’此車柄動態預測方法包含下列步驟： 測量+^車輛之質心的縱向速度，並產生—縱向速度信號； 測足兩之貝〜的側向加速度，並產生一侧向加速度信號； 1里車柄之輯彈簧之長度變化，並產生-長度變化信號； I294376 接受縱向速度信號、側向加速度信號與長度變化信號，而估 測出車輛之一側傾角、一側傾角速度、一側傾角加速度、/ 俯仰角、一俯仰角速度、一俯仰角加速度、一橫擺角、一橫 擺角速度、一橫擺角加速度、一縱向位移、一縱向加速度、 一側向位移、一侧向速度、一垂直位移、一垂直速度、一燊 直加速度與一輪胎角速度；及接受縱向速度信號、側向加速 度信號、長度變化信號、側傾角、側傾角速度、側傾角加速 度、俯仰角、俯仰角速度、俯仰角加速度、橫擺角、橫擺角 速度、橫擺角加速度、縱向位移、縱向加速度、側向位移、 侧向速度、垂直位移、垂直速度、垂直加速度與輪胎角速度， 而預測車輛於未來時間之運動行為。 此車辅動態預測方法更包含接收侧傾角，而估算側傾角 之變化趨勢而判斷車輛於未來時間是否翻覆。 、 其中側傾角之變化趨勢的步驟係為估算側傾角或 角是否為發散函數的步驟。 ' 樹此車辅動態預測方法更包含接收俯仰角，而估算 之艾化趨勢而觸車祕未來_是否翻覆。 其中俯仰角之變化㈣的步驟係為 發散函數的步驟。 价角疋否為 此車翻動態預 其中測量車輛 測方法更包含估异該車輛之翻覆位置 之質心的側向加速度，可用測量車輛之質 1294376 心的橫擺角速度代替，並產生一橫擺角速度訊號，亦可估算 出相同的結果。另亦可用測量車輛之質心的橫擺角代替測量 f心的側向加速度，而產生一橫擺角速度信號，同樣可估測 鼻出相同的結果。 因此，本發明之一種車輛動態預測系統與方法，可利用 • 三個感測出的物理量，並經由估測器與預測器可以預測車輛 ^ 於未來時間之執跡，更進而可預測車輛於行進時是否會翻 覆’因此，經由估測器的設置而可使必須利用量測之物理量 大幅減少，而降低感測器的成本，另—方面，由於感測器的 j幅減少，使得因感測器失效或產生錯誤信號而造成估測錯 誤之情況亦可大大地降低，更增加預_穩定性與精確性。 而且由於習知技術中，其估算之法測係為經驗公式，因此， 办斤估算之值大都枝轉料，穌發明係顧完整車輛 1 模型作為估算之法則’因此’所估算之值可非常準確。而且 ， 本發明另將道路狀況所產生之因鱗慮在内，可更增加其估 '算之準確性。 9 ^ 有關本發明㈣徵與實作，㈣合圖轉最佳實施例詳 細說明如下。 【實施方式】 請翏閱「第1圖」，所示為本發明之組合方塊示意圖。 本發明係—種車姉態侧系統，_以觸-車輛之運動 1294376 行為’此車輛動態預測系統包含一感測部、—估測器3〇〇與 一預測器500。其中感測部係由下列元件組成： 一第一感測器110係測量車輛之質心的縱向速度，並產 生一縱向速度信號·’ 一第二感測器130係測量車輛之質心的 侧向加速度，並產生一側向加速度信號及一第三感測器15〇 係測量車輛之懸掛彈簧之長度變化，並產生一長度變化信 號。其中，第一感測器110可為縱向速度感測器，而第二感 測器130可為側向加速規，而第三感測器15〇可為懸掛距離 感測器。 而估測器300係包含非線性狀態觀察器，係由完整 車輛模型320為基礎所建立，在此利用擴增卡曼濾波器 (Extended Kalman Filter)為一例子說明，此為利用系統動 態方程式、系統誤差和觀測誤差的統計與隨機特性、及初始 條件等信息對觀測數據進行處理，從而得到系統狀態變數之 最小誤差估測的一種演算法則。當估測器3〇〇接受縱向速度 L號、側向加速度k$虎與長度變化信號，而經由非線性狀態 觀祭器310而估測出車輛之一側傾角、一側傾角速度、一側 傾角加速度、一俯仰角、一俯仰角速度、一俯仰角加速度、 一橫擺角、一橫擺角速度、一橫擺角加速度、一縱向位移、 一縱向加速度、一側向位移、一侧向速度、一垂直位移、一 垂直速度、一垂直加速度與一輪胎角速度。 1294376 而預測Is 5GG係包含-完整車輛模型32()，且接受縱命 速度15说、侧向加速度信號 '長度變化信號、侧傾角、側傾 角速度、側傾角加速度、俯仰角、俯仰祕度、俯仰角加速 度、橫擺角、橫擺角速度、橫擺角加速度、縱向位移、縱向 加速度、幾娜、側向速度、垂直娜、垂直速度、垂直 加速度與輪胎角速度’而細彳車輛於未來時間之運動行為。 另一道路狀況（road condition) 400被以“外界干擾” 的方式輸入系統，藉由將道路狀況加入至完整車輛模型32〇 透過選取適當的感測部’道路狀況侧對車輛動態的影響將 可被估測器300正確估出。 曰 -駕駛者操作行為（steeringmaneuve;r) 等訊息傳 輸至估測II ，以協助伽m 3⑼錢估測出其它的物理 量。 在估測器300中，我們可以採用以完整車輛模型32〇， 如「第1圖」’絲礎所建立之擴增卡曼滤波器。 完整車輛模型咖包含了車麵六個自由度、懸掛系統位BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle dynamic prediction system and method, and more particularly to using a physical quantity of three measurements, and using a numerical algorithm to estimate one. The associated physical quantities, and using a numerical algorithm, one predictor accepts these physical quantities to predict the dynamic behavior of the vehicle. [Prior Art] For the driving condition of the vehicle, the more advanced vehicle electronic system will be equipped with a vehicle overturn warning device to predict the environment in which the vehicle is under and the driving situation, which may cause the vehicle to overturn, and may be timely. Give a warning or just give it a heart. Related Art, such as U.S. Patent Publication No. 6002974, utilizes a roll angle sensor, a pitch rate sensor, a longitudinal acceleration gauge, a lateral acceleration gauge, and a vertical acceleration gauge, and utilizes an augmented Kalman filter (Extended Kalman) Filter) to estimate the current roll angle and pitch angle, and then use the roll angle and roll angle to apply the empirical weighting function to predict the future roll angle, and similarly use the pitch angle and pitch angle to apply the empirical weighting function to predict the future. The pitch angle. For example, Patent No. 61,912-5 uses a roll angle sense, a sense of pitch sensitivity, a longitudinal acceleration gauge, a lateral acceleration gauge, and a vertical acceleration gauge, and an amplifying Karman filter to estimate the current roll angle and pitch angle. The estimated lateral acceleration and longitudinal velocity are then applied to the empirical formula to derive the deviation value produced by the lion, and the deviation 5 1294376 value and the roll angle are applied to the empirical weighting function to predict the future roll angle, and the deviation The value and pitch angle are applied to the empirical weighting function to predict the future pitch angle. For example, the sensor of No. 6556908 is selected from a lateral acceleration gauge, a lateral acceleration gauge, a yaw angle sensor, a roll angular velocity sensor and a tire speed sensor, and is applied via lateral acceleration and roll angular velocity. The empirical weighting function is used to calculate the past and current roll angle velocities, and the past and current slashes are used to calculate the surface gradient of the lock correction by the other empirical weighting function. The more flexible side is to judge the overturning event. The sensor of No. 6631317 includes a sharp acceleration gauge, a lateral acceleration gauge, a yaw angle sensing H, a side speed sensing hall and a tire speed sensor, and then a lateral acceleration gauge, a 33-way acceleration gauge, and a surface. The oscillating speed, the yaw rate sensor and the tire speed sensing H measure the physical quantity to the simplified formula to determine the temporary and the surface, _ this value and two numerical calculation filters to estimate The roll angle is derived to determine the overturning event. However, the above patents use four to five sensors to estimate the roll angle of the vehicle' and then predict the vehicle roll angle by empirical weighting function or simplified formula, and use this physical quantity to announce whether to overturn, 'but predicted The vehicle roll angle is not affected by the road conditions. SUMMARY OF THE INVENTION The problem to be solved by the present invention is to provide a vehicle dynamic pre-system and method. The physical body 1294376 obtained by the _ three-domain detector is a butterfly type to obtain a heart. The bad thing is that the vehicle will be overturned by the side angle of the future time. The invention discloses the technical problem of the above-mentioned vehicle domain, the time of the door, the Μ __, the system, the _ the vehicle is not estimated, the ρίΙ is the vehicle dynamic prediction system including - the sensing part , - L and - predictor 'where the sensing part consists of the lower navigation parts: ... - the first - sensor measures the longitudinal velocity of the vehicle's centroid and produces - 7-way speed signal; - the second sensor system The lateral acceleration of the centroid of the vehicle is measured. X JL generates a lateral acceleration signal and a third sensor measures the length change of the suspension spring of the vehicle and generates a length change signal. ...the second estimator consists of a non-linear heart-shaped τ, 'and the velocity of the money, the lateral acceleration money and the length change signal' constructed by a vehicle dynamic model and estimated by the surface trait inspection _ vehicle One side inclination angle, - roll angle speed, - roll angle acceleration, - pitch angle, - pitch rate, - pitch angle acceleration, - Newshaw, - yaw limb, one yaw rate acceleration, - longitudinal displacement, one longitudinal acceleration - lateral displacement, lateral velocity, - vertical displacement, - vertical velocity, - vertical acceleration and a tire angular velocity. The predictor includes a vehicle dynamic model and accepts longitudinal velocity signals, lateral acceleration signals, length variation signals, roll angles, roll angular velocities, roll angular accelerations, pitch angles, pitch angular velocities, pitch angular accelerations, yaw 1294376 angles, The lateral velocity, the lateral acceleration, the longitudinal displacement, the longitudinal acceleration, the i-bit private lateral velocity, the vertical displacement, the vertical velocity, and the vertical acceleration disk tire angular velocity, and predict the motion behavior of the vehicle in the future. "Vehicle_side her ^ca, angling or bowing: angle, or both, and estimating the trend of the roll angle or pitch angle to determine the vehicle __ domain, the domain calculates the future vehicle. The trend of the roll angle or the pitch angle is whether the roll angle or angle 为 is a divergence function' _ whether the vehicle overturns in the future time 2 = the lag increase is estimated to read the future out of the vehicle ^ and the second sensor is used Measuring the lateral acceleration of the vehicle's center of mass can be replaced by the yaw rate of the center of mass of the vehicle, and the yaw rate can also cause the estimator to have the same effect. The detection of the _ lateral acceleration, and the generation of the angular velocity 可使 可使 可使 可使 可使 细 细 于 于 于 于 于 于 于 于 于 于 于 于 于 为 为 为 为 为 为 为 为 为 为 为 为 为 为 为 为 为 为 为 为 为 为 为 为Paste method, made fine - vehicle in the future time as 'this handle dynamic prediction method includes the following steps: measuring +^ vehicle's centroid longitudinal speed, and generating - longitudinal speed signal; measuring the two shells ~ Lateral acceleration and side acceleration Signal; 1 length of the handle of the spring changes, and produces a - length change signal; I294376 accepts the longitudinal speed signal, the lateral acceleration signal and the length change signal, and estimates a vehicle roll angle, one side tilt speed, One side inclination acceleration, / pitch angle, one pitch angle speed, one pitch angle acceleration, one yaw angle, one yaw rate, one yaw rate acceleration, one longitudinal displacement, one longitudinal acceleration, one side displacement, one side direction Speed, a vertical displacement, a vertical speed, a straight acceleration and a tire angular velocity; and accept longitudinal velocity signals, lateral acceleration signals, length variation signals, roll angles, roll angular velocities, roll angular accelerations, pitch angles, pitch angular velocities , pitch acceleration, yaw angle, yaw rate, yaw rate acceleration, longitudinal displacement, longitudinal acceleration, lateral displacement, lateral velocity, vertical displacement, vertical velocity, vertical acceleration and tire angular velocity, while predicting the vehicle in the future The behavior of the vehicle is also included in the vehicle dynamic prediction method. The trend of the angle changes to determine whether the vehicle overturns in the future time. The step in which the trend of the roll angle changes is a step of estimating whether the roll angle or angle is a divergence function. 'The tree-assisted dynamic prediction method further includes receiving the pitch angle, And the estimated Aihua trend and the car secret future _ whether to overturn. The step of the pitch angle change (4) is the step of the divergence function. The valence angle 为此 No for this car turnover dynamic pre-measurement vehicle measurement method more includes estimation The lateral acceleration of the center of mass of the vehicle's overturned position can be measured by measuring the yaw rate of the vehicle's mass of 1294376 and generating a yaw rate signal, which can also estimate the same result. It can also be used to measure the vehicle's center of mass. Instead of measuring the lateral acceleration of the f-center, the yaw angle produces a yaw rate signal, which can also estimate the same result of the nose. Therefore, the vehicle dynamic prediction system and method of the present invention can utilize three sensing The physical quantity, and the estimator and predictor can predict the vehicle's performance in the future, and further predict the vehicle Whether it will be overturned' Therefore, the physical quantity of the measurement must be greatly reduced by the setting of the estimator, and the cost of the sensor is reduced. On the other hand, since the j-span of the sensor is reduced, the sensing is caused by the sensing. The failure of the device or the generation of an error signal can also greatly reduce the estimation error, and increase the stability and accuracy. Moreover, due to the conventional technique, the estimated method is an empirical formula. Therefore, the value of the estimate is mostly transferred. The invention is based on the complete vehicle 1 model as the rule of estimation. Therefore, the estimated value can be very accurate. Moreover, the present invention further increases the accuracy of the estimation of the road conditions caused by the road conditions. 9 ^ In relation to the invention (4) signing and implementation, (4) the best embodiment of the drawing is detailed as follows. [Embodiment] Please refer to "FIG. 1", which is a schematic block diagram of the present invention. The present invention is a vehicle-side state system, _acting-vehicle motion 1294376 behavior' This vehicle dynamic prediction system includes a sensing portion, an estimator 3A, and a predictor 500. The sensing portion is composed of the following components: a first sensor 110 measures the longitudinal velocity of the center of mass of the vehicle and generates a longitudinal velocity signal. A second sensor 130 measures the side of the center of mass of the vehicle. To the acceleration, and to generate a side acceleration signal and a third sensor 15 to measure the length change of the suspension spring of the vehicle, and generate a length change signal. The first sensor 110 can be a longitudinal speed sensor, and the second sensor 130 can be a lateral acceleration gauge, and the third sensor 15 can be a suspension distance sensor. The estimator 300 includes a nonlinear state observer, which is established based on the complete vehicle model 320. Here, an extended Kalman filter is used as an example. This is a system dynamic equation. The statistical and stochastic characteristics of the systematic and observed errors, and the initial conditions are used to process the observed data to obtain an algorithm for estimating the minimum error of the system state variables. When the estimator 3 receives the longitudinal velocity L number, the lateral acceleration k$ tiger and the length change signal, and estimates the vehicle's roll angle, the side inclination speed, and the side inclination angle via the nonlinear state spectator 310 Acceleration, a pitch angle, a pitch angle speed, a pitch angle acceleration, a yaw angle, a yaw rate, a yaw rate acceleration, a longitudinal displacement, a longitudinal acceleration, a lateral displacement, a lateral velocity, a Vertical displacement, a vertical velocity, a vertical acceleration, and a tire angular velocity. 1294376 and the prediction Is 5GG contains the complete vehicle model 32(), and accepts the vertical velocity 15, the lateral acceleration signal 'length change signal, the roll angle, the roll rate, the roll angle acceleration, the pitch angle, the pitch margin, Pitch acceleration, yaw angle, yaw rate, yaw rate acceleration, longitudinal displacement, longitudinal acceleration, Jona, lateral velocity, vertical Na, vertical velocity, vertical acceleration and tire angular velocity' Sports behavior. Another road condition 400 is entered into the system in an "outside interference" manner by adding road conditions to the complete vehicle model 32. By selecting the appropriate sensing portion, the road condition side will have an impact on vehicle dynamics. The estimator 300 is correctly estimated.曰 - Driver's operational behavior (steeringmaneuve; r) and other information is transmitted to Estimation II to assist gamma 3 (9) money to estimate other physical quantities. In the estimator 300, we can use an augmented Kalman filter built on a complete vehicle model 32, such as "Fig. 1". The complete vehicle model coffee contains six degrees of freedom in the car surface and the suspension system
移自由度、以及輪胎轉速自由度的車輛模型，其動態方程 可以表示如下式： X y=^^+Fr)~xe 12 1294376 Ιχώχ = Mx + (iy ~ ΙΖ y〇yCDzThe vehicle model of the degree of freedom of movement and the degree of freedom of the tire speed can be expressed as follows: X y=^^+Fr)~xe 12 1294376 Ιχώχ = Mx + (iy ~ ΙΖ y〇yCDz
Iy0y = My + (lz - Ιχ )〇^ζ〇)χ Ιζώζ = Μζ + (ΐχ - Iy \〇xcoyIy0y = My + (lz - Ιχ )〇^ζ〇)χ Ιζώζ = Μζ + (ΐχ - Iy \〇xcoy
Ht = Z-^cos^ + ^-^cos^cos^-^sin^sin^-i ώ，= — {^^Fai ^Tbrakei ^-Tmotori) wheel 其中“[/丨/〗-/2 -/2]r，從=[+夸令令}" ωχ -ψ-έ^ιηθ, ωγ = i9cos^ + ^cosi9sin^5 ωζ --θ^\ηφ +έοο^θζο^φ Fzi^K-Hi + D-Hi 上式參數的定義為： X’J’Z ·車輛之質心在縱向（longitudinal direction)、 侧向（lateral direction)與垂直(vertical direction)線性 方向上的位移量。 •在尤拉轉換中所使用的三個尤拉角（Euier angles) 〇 ·車幸兩沿|X，j；，z 三軸旋轉的角速度，其中a定義為 侧傾（r〇11)角速度、〜定義為俯仰（pitch)角速度、你-定 義為橫擺（yaw)角速度。 w/第輪胎上之有效三軸作用力，其中為非線性 輪胎模型之輪出。 ^ :第ζ·輪胎的縱向黏著力。Ht = Z-^cos^ + ^-^cos^cos^-^sin^sin^-i ώ,= — {^^Fai ^Tbrakei ^-Tmotori) wheel where "[/丨/〗-/2 -/ 2]r, from =[+callal order}" ωχ -ψ-έ^ιηθ, ωγ = i9cos^ + ^cosi9sin^5 ωζ --θ^\ηφ +έοο^θζο^φ Fzi^K-Hi + The D-Hi parameter is defined as: X'J'Z · The amount of displacement of the center of mass of the vehicle in the longitudinal direction, the lateral direction, and the vertical direction. The three Euler angles used in the conversion are the angular velocity of the three-axis rotation, where a is defined as the roll (r〇11) angular velocity, and ~ is defined as the pitch. (pitch) angular velocity, you - defined as yaw angular velocity. w / effective triaxial force on the tire, which is the wheel of the nonlinear tire model. ^ : Dimensional tire longitudinal adhesion.
Fw :模擬道路狀況所激發的力。 13 !294376 :以質心為轉動中心之轉動動量，是由輪胎上之有 效作用力所引起。 A:懸掛系統之彈黃長度變化量。 :第/輪胎的煞車力矩，藉由駕駛者踩煞車力道 算出。 道所 ·第’輪胎的引擎力矩，藉由駕駛者踩油門力 計算出。 以下為車輛常數： :車輛總重量。 車輛成何參數，此為車輛前軸兩輪所夾之長度。 峋：車輛幾何參數，此為車輛後軸兩輪所夾之長度。 △:車輛幾何參數，此為車輛質心到前軸之長度。 2 ·車輛幾何參數，此為車輛質心到後轴之長度。 •針對於質心之三軸轉動慣量。 :輪胎的轉動慣量。 尤·懸掛系統之彈簣彈性係數。 D ··懸掛系統之阻尼系統。 巧：輪胎半徑。 因此，由上述公式之觀察矩陣可知，把用以測量車 質心的側向加逮度，可用測量車輛之質心的橫擺角速度代 14 I294376 替’並產哇一 ^ 的作用， 訊號，亦可使估測11 _產生相同 可用、群f ^二感測1113G可為橫擺驗度_儀。另亦 ’、·里輪之質心的橫擺角代替側向加速度，並 ^角速度域，同樣可使估測_產生相同的作用，^ 感測ϋ 130可為量測橫擺角感測器。 另，車輛動態預測系統更包含一判斷器6〇〇，係接收側 j ’而估算側則之變化趨勢_斷車輛於未來時間是否 _ :其中_角之變化趨勢係為側傾角是否為發散函數， 而判斷車輛於未來時岐否翻覆。並將估測器3GG產生之訊 號與下-時序之駕駛者參數而再次提供☆下一時序之估_ _。若當判斷器6_斷車輛於未來時間會翻覆，則產生; 號至-防止系統，以自動調整車輛狀況，而避免車辅翻 覆。或是賴麟至魏11，肋警告驗者車輛於未來時 間會翻覆，祕駕駛者耕離車触況，㈣防止車輛翻 覆。並且’此满H _可祕算_車輛位置為起始 點’而估算未來車輛之翻覆位置。 —閱「第2圖」至「第5圖」’所示為本發明之模擬 數據圖。圖中實線為車輛在_中之模擬情形，在這些模擬 中，車輛行走的速度為每小時90公里並且在第4秒的時候作 急速轉彎的動作’這些模擬-共有兩種情況，而這些情況分 別是駕駛者處於不同角度的路面上行駛，從「第2圖」至「第 15 1294376 力。圖」便壬現其模擬結果，於圖中用實線來當作車輛在某— 駕駛行為下的_響應，虛線為估· 測與預測器 500的預測。 勒怨輿未來動態之間的差異，預測器 〜將不έ去做每伽桃_未來祕酬，而是會故意地 设计，隨著時間而輸出目前的車輛絲。除此之外，由三個 、'£、、' ° Ϊ10 130、150測得的量（質^之側向加速度與縱向 ^度以及_彈簧之長度變化）會故意地在5.1秒前進入並 估測器_且對車輛即時動態進行估算，並在5. i秒後 將感如110、130、15〇關掉，進行車輛動態的預測。其中 ^向4 5〜8秒採用固定的轉f纽’是模擬車輛在固定的 二入下（固定方向盤卿角度）的動態行為。若在5〜8秒中 訊^動態行為（如圖中虛線，沒有感測器110、130、150 辅在時域中的模擬（如圖中實線）相符合，即 = '們可此估測H 3⑽與_ 苹 —輛二 如「第2圖」盥「筮α闽 α — 角度、_^/、 係顯示車輛之方向盤 、购擺角對時_係圖以及車輛質心、之三軸方 ^位私對時_麵。此情⑽代 日_ 滑順地轉f，本發明之|減At 斜坦的路面上做 月之車輛動關測系統可成功地預測車輛 1294376 於未來時間之動態行為，雖 為舆預測器_輪出有=:8秒之後在車輛動態行 駕敬者改變了方向盤的角戶補^同’這是因為8秒之後’ 用度亚且如預期地預測器500盔法 然而美中不足的是車輛橫擺角並無法相當吻合。… 如弟4圖」與「第5圖」所示，係顯示車輛之方向盤 角度、麵角與橫擺角對時間關係圖以及車輛質心之三軸方 ^移對時間關係圖。此情況為車輛在斜坡上做滑獅轉 弓’此斜坡為沿者路面縱向軸偏-25細斜坡，本發明之車 輛動態預測系統依然是可作用的，且此情況是經過特殊設 。十’其所有的操作情況與第一種情況不同的是僅僅在於車輛 行敬於不關路Μ。此軌的重躲在鱗贿駛於斜坡 上’而此斜坡亦是引起車輛翻覆的要件，此外，在這個狀況 下，車辅橫擺角不同於其他情況’而可成功地被估測且預測 ^，這是因為側向加速度在斜坡上時含有車婦擺角此項動 恶，因此車輛横擺角可藉由側向加速度絲綱。而且可預 估車輛於接近七秒時翻覆。 、 請參閱「第6圖」、「第7圖」與「第8圖」，所示為本 么明之方法流程圖。如「第6圖」所示，本發明之_種車輛 動態預測方法，係用以·卜車輛於未糾間之運動行為， 此車輛動態預剛方法包含下列步驟·· 首先測量車輛之質心的縱向速度及側向加速度與車柄 17 !294376 之懸掛彈簧之長度變化’並產生一縱向速度信號、一側向加 速度信號與-長度變化信號（步驟900);接受縱向速度信 號、側向加速度信號與長度變化信號，而估測出車輛之一側 傾角、一側傾角速度、一側傾角加速度、一俯仰角、一俯仰 角速度、一俯仰角加速度、一橫擺角、一橫擺角速度、一橫 擺角加速度、-縱向位移、—縱向加速度、—側向位移、一 侧向速度、-垂直位移、—垂直速度、—垂直加速度與一輪 月。角速度（步驟9〇1)，接*縱向速度信號、侧向加速度信號、 長度變化錢、側傾角、侧則速度、侧傾角加速度俯°仰 角、俯仰角速度、俯仰角加速度、橫擺角、橫擺角速度、橫 擺角加速度、縱向位移、縱向加速度、側向位移、側向速度、 垂直位私、垂直逮度、垂直加速度與輪胎肖速度，而預測車 柄於未來時間之運動行為（步驟903)。接收側傾角，而估瞀 ®」所不，經由上述步驟_至步驟903後， 估算俯仰角之變化趨勢而判斷車輛於未來時 間疋否翻復（步驟907)。 _至步驟_後 」所不’經由上述步驟 (步驟_)。 且可預估車輛於未來時間的翻覆位置 /、中侧角與俯仰肖之變減勢的步 角與俯仰角是否為發散函數的步驟。 - 18 A294376 、、/、中測里車輛之質心的側向加速度，可用測量車輛之質 、也丄角速度代替’並產生-橫擺角速度訊號，亦可估算 ^車輛的所有相關物理量。另亦可㈣量車輛之質心的择擺 =代替測量質心_向加速度，而產生—橫翻速度信號， 5樣可估㈣^車輛的所有蝴物理量。 〆^然本發明簡述之較佳實施例揭露如上，然其並非用 、、定本么月’任何熟習相像技藝者，在不脫離本發明之 神和耗_，當可作些許之更動與稿，因此本發明之 ;=:r書所附之申請專刪所界定者為準。 「第1圖」_示本發明之組合方塊示意圖； 關=圖」係㈣㈣顺W位移對時間 第4圖」係顯示本發明車辅. 角對時間關係圖； 第5圖」係顯示本發明之車輛質心之， 丨之方向盤角度、側傾角與 横擺 關係圖；及 「弟6圖」、「第7 圖。 .軸方向位移對時間 圖」與「第8圖」係顯示本發明之方 法流 19 1294376 【主要元件符號說明】 110第一感測器 130第二感測器 150第三感測器 300估測器 310非線性狀態觀察器 320完整車輛模型 400道路狀況 410駕駛者操作行為 500預測器 600判斷器 700防止系統 20Fw: Simulates the force stimulated by road conditions. 13 !294376 : The rotational momentum with the center of mass as the center of rotation is caused by the effective force on the tire. A: The amount of change in the length of the suspension system. : The braking torque of the tire/tire is calculated by the driver's pedaling force. The engine torque of the first tire is calculated by the driver's stepping on the throttle force. The following are vehicle constants: : Total vehicle weight. What is the vehicle's parameters, this is the length of the two wheels of the front axle of the vehicle.峋: Vehicle geometry, this is the length of the two wheels of the rear axle of the vehicle. △: Vehicle geometry, this is the length of the vehicle's center of mass to the front axle. 2 · Vehicle geometry, this is the length of the vehicle's center of mass to the rear axle. • Three-axis moment of inertia for the center of mass. : The moment of inertia of the tire. The elastic coefficient of the magazine of the suspension system. D ··The damping system of the suspension system. Qiao: The radius of the tire. Therefore, from the observation matrix of the above formula, it can be known that the lateral acceleration of the car's centroid can be measured by measuring the yaw rate of the vehicle's center of mass and using 14 I294376 for the purpose of producing a whip. It can be made that the estimate 11 _ produces the same available, the group f ^ two senses 1113G can be a yaw test _ instrument. In addition, the yaw angle of the centroid of the inner wheel replaces the lateral acceleration, and the angular velocity domain can also make the estimation _ produce the same effect. ^ The sensing ϋ 130 can be a measuring yaw angle sensor. . In addition, the vehicle dynamic prediction system further includes a determiner 6〇〇, which is the receiving side j′ and the trend of the estimated side is _when the vehicle is in the future time _ : wherein the trend of the _ angle is whether the roll angle is a divergence function And judge whether the vehicle will be overturned in the future. The signal generated by the estimator 3GG and the driver parameter of the lower-time sequence are again provided ☆ the estimate of the next timing _ _. If the judger 6_ disconnected the vehicle will overturn in the future time, the number is generated to prevent the system from automatically adjusting the condition of the vehicle to avoid the car's auxiliary overturning. Or Lai Lin to Wei 11, the rib warning vehicle will be overturned in the future, the secret driver will plough away from the car, and (4) prevent the vehicle from overturning. And 'this full H _ secretly _ vehicle position is the starting point' and the position of the future vehicle is estimated. - The "Fig. 2" to "5th figure" are shown as simulation data maps of the present invention. The solid line in the figure is the simulation of the vehicle in _. In these simulations, the speed at which the vehicle walks is 90 km per hour and the sharp turn at the 4th hour. These simulations have two cases. The situation is that the driver is driving on a different angle of the road. From "2nd picture" to "15th 1294376 force. Figure", the simulation results are shown. In the figure, the solid line is used as the vehicle in a certain driving behavior. The lower _ response, the dashed line is the prediction of the estimation and predictor 500. Resent the difference between future dynamics, the predictor ~ will not go to do every gamma _ future secrets, but will deliberately design, output the current vehicle wire over time. In addition, the amount measured by three, '£,, ' ° Ϊ 10 130, 150 (the lateral acceleration and the longitudinal length of the mass ^ and the length of the spring) will deliberately enter before 5.1 seconds and The estimator _ is used to estimate the instantaneous dynamics of the vehicle, and after 5 seconds, the senses such as 110, 130, and 15 are turned off, and the vehicle dynamics are predicted. Where ^^4 to 8 seconds using a fixed turn-f button' is the dynamic behavior of the simulated vehicle under a fixed two-in (fixed steering wheel angle). If the dynamic behavior is in 5~8 seconds (the dotted line in the figure, there is no analog 110, 130, 150 in the time domain simulation (solid line in the figure), ie = 'we can estimate Measure H 3(10) and _ ping - vehicle 2 as "2" 盥 "筮α闽α - angle, _^/, showing the steering wheel of the vehicle, purchasing the angle of the traverse _ system and the vehicle's center of mass, the three axes The situation is (10) on behalf of the day _ smoothly turn f, the invention of the | reduction At the slope of the road on the moon, the vehicle dynamics measurement system can successfully predict the vehicle 1294376 in the future time Dynamic behavior, although the 舆 predictor _ wheel has = = 8 seconds after the vehicle dynamics in the vehicle, the driver changed the direction of the steering wheel to fill the same 'this is because after 8 seconds' used and expected as predictor The 500 helmet method, however, is that the yaw angle of the vehicle is not quite consistent.... As shown in Figure 4 and Figure 5, it shows the steering angle, face angle and yaw angle of the vehicle and the vehicle. The three-axis square of the centroid is shifted to the time relationship. This situation is for the vehicle to make a lion turn on the slope. The longitudinal axis deviation is 25-thick slope, the vehicle dynamic prediction system of the present invention is still applicable, and the situation is specially designed. The tenth operation of the tenth is different from the first case only in that the vehicle is respectful It’s not a road. The track is heavily hidden in the scales and bribes on the slopes. This slope is also a requirement for the vehicle to overturn. In addition, in this case, the vehicle's yaw angle is different from other cases' and can be successfully Estimated and predicted ^, because the lateral acceleration on the slope contains the driver's swing angle, so the vehicle yaw angle can be used by the lateral acceleration line. And the vehicle can be estimated to be close to seven seconds. Please refer to "Picture 6", "Picture 7" and "Figure 8" for the flow chart of the method of the present invention. As shown in Figure 6, the vehicle dynamics of the present invention The prediction method is used for the motion behavior of the vehicle in the uncorrected position. The vehicle dynamic pre-compaction method includes the following steps: First, measure the longitudinal speed and lateral acceleration of the center of mass of the vehicle and the suspension spring of the handle 17!294376 The length changes 'and produces Longitudinal speed signal, lateral acceleration signal and length change signal (step 900); receiving longitudinal velocity signal, lateral acceleration signal and length variation signal, and estimating one side roll angle, one side inclination speed, one side of the vehicle Inclination acceleration, a pitch angle, a pitch rate, a pitch angle acceleration, a yaw angle, a yaw rate, a yaw rate acceleration, a longitudinal displacement, a longitudinal acceleration, a lateral displacement, a lateral velocity, - Vertical displacement, - Vertical velocity, - Vertical acceleration and one month. Angular velocity (step 9〇1), connected to * longitudinal velocity signal, lateral acceleration signal, length change money, roll angle, side velocity, roll acceleration Elevation angle, pitch rate, pitch angle acceleration, yaw angle, yaw rate, yaw rate acceleration, longitudinal displacement, longitudinal acceleration, lateral displacement, lateral velocity, vertical position private, vertical catch, vertical acceleration and tire slewing speed And predicting the behavior of the handle in the future time (step 903). Receiving the roll angle, and estimating 瞀 」, after the above steps _ to 903, the trend of the change in the pitch angle is estimated to determine whether the vehicle will be retaliated in the future (step 907). _ to step _ after "not" via the above steps (step _). It is also possible to estimate whether the step angle and the pitch angle of the vehicle's overturning position / mid-side angle and pitch-down change are divergence functions. - 18 A294376 , , / , The lateral acceleration of the center of mass of the vehicle in the mid-measurement can be measured by measuring the mass of the vehicle and also the angular velocity instead of generating a yaw rate signal. It can also estimate all relevant physical quantities of the vehicle. Alternatively, (4) the vehicle's centroid selection = instead of measuring the centroid _ to the acceleration, and generating - the traverse speed signal, 5 can estimate (four) ^ all the physical quantities of the vehicle. The preferred embodiment of the present invention is disclosed above, but it is not intended to be used by any skilled artist, without departing from the spirit and expense of the present invention, when a slight change can be made. Therefore, the invention is defined in the following: "Fig. 1" - shows a schematic block diagram of the present invention; "off" = "system" (4) (four) VS shift to time, Fig. 4" shows the vehicle auxiliary of the present invention. Angle versus time diagram; Fig. 5 shows the present invention The vehicle's center of mass, the steering wheel angle, the roll angle and the yaw diagram; and the "Sixth Diagram", "The 7th Diagram. The Axis Direction Displacement vs. Time Chart" and "Figure 8" show the present invention. Method Flow 19 1294376 [Main Element Symbol Description] 110 First Sensor 130 Second Sensor 150 Third Sensor 300 Estimator 310 Nonlinear State Viewer 320 Complete Vehicle Model 400 Road Condition 410 Driver Operation Behavior 500 predictor 600 determiner 700 prevents system 20