WO2017152648A1 - 车辆急转弯的识别方法及装置 - Google Patents

车辆急转弯的识别方法及装置 Download PDF

Info

Publication number
WO2017152648A1
WO2017152648A1 PCT/CN2016/105538 CN2016105538W WO2017152648A1 WO 2017152648 A1 WO2017152648 A1 WO 2017152648A1 CN 2016105538 W CN2016105538 W CN 2016105538W WO 2017152648 A1 WO2017152648 A1 WO 2017152648A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
plane
lateral acceleration
sharp turn
gravity
Prior art date
Application number
PCT/CN2016/105538
Other languages
English (en)
French (fr)
Inventor
刘均
李磊
张伟
杨勇
Original Assignee
深圳市元征科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳市元征科技股份有限公司 filed Critical 深圳市元征科技股份有限公司
Publication of WO2017152648A1 publication Critical patent/WO2017152648A1/zh

Links

Images

Classifications

    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W40/00Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • B60W40/08Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to drivers or passengers
    • B60W40/09Driving style or behaviour
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W40/00Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • B60W40/10Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to vehicle motion
    • B60W40/109Lateral acceleration

Definitions

  • the invention relates to the technical field of automobiles, and in particular to a method and a device for identifying a sharp turn of a vehicle.
  • assisted driving and driving behavior analysis is a very important part.
  • Automated analysis of driving behavior helps to identify and prevent many factors that affect traffic safety and cause traffic accidents.
  • the identification of driving sharp turns is an important part of the analysis of the driver's driving behavior. This analysis can remind the driver to pay attention to avoid fatigue driving, lack of concentration and other issues, while many insurance companies also make sharp turns.
  • the analysis of driving behavior establishes the insurance model.
  • Driving a sharp turn is one of the driving behaviors. There is no clear regulation in traffic regulations. Usually, driving a sharp turn means turning at a sharp turn of the road at a speed exceeding the prescribed speed during driving. This line speed may be The feeling of the driver and the driver of the car is uncomfortable. The most serious hazard of driving a sharp turn is that the lateral acceleration of the car is too large, causing traffic accidents such as skidding or even overturning of the vehicle.
  • angular acceleration instantaneous angular velocity ⁇ 9 ° / s
  • angular difference five
  • the absolute difference between the driving direction angles in the second is greater than 45°; the turning angle is between 30 and 45° in three seconds (including 30°, excluding 45°), and the speed is reduced by less than 30% in three seconds, and the average speed Greater than 50 mph; 4, three seconds turning angle between 45-60 ° (including 45 °, excluding 60 °), and speed reduction of less than 50% in three seconds, and the average speed is greater than 45 mph; 5, three seconds turn
  • the angle is between 60-90° (including 60), and the speed is reduced by less than 75% in three seconds, and the average speed is greater than 30 mph.
  • the super-high part of the road surface is increased, and when these corners are turned, it is allowed to have a higher angular velocity than the flat road (or Said line speed); for example, at 36 mph, the lateral force with a turning angle of 90° in three seconds is only about 1/4 of the same turn (90°) with a speed of 144 mph.
  • the turning angle in three seconds only needs to be 22.5°, which will have the same lateral acceleration. This means that if it is from a safety angle (guarantee the car does not roll over, not side) Sliding) and the comfort feeling of the driver and passengers, the definition of the above sharp turn has serious drawbacks.
  • the main object of the present invention is to provide a method and a device for identifying a sharp turn of a vehicle, and to solve the technical problem of using a gyroscope to detect a sharp turning error of a vehicle by detecting an angular velocity change and having a high cost.
  • the present invention provides a method for identifying a sharp turn of a vehicle, comprising:
  • the step of acquiring a plane parameter of the gravity direction and the forward direction of the vehicle to form a plane comprises:
  • a plane parameter of a plane of gravity and a direction of advancement of the vehicle is obtained by a multiple linear regression fitting method.
  • the step of acquiring a plane parameter of the gravity direction and the forward direction of the vehicle to form a plane comprises:
  • a planar parameter of a plane is formed by acquiring a preset gravity direction and a forward direction of the vehicle from a server or a local according to information of the vehicle.
  • the invention also provides a method for identifying a sharp turn of a vehicle, comprising:
  • the step of determining whether the vehicle has a sharp turn according to the lateral acceleration comprises:
  • the step of obtaining the lateral acceleration of the vehicle by using a preset algorithm according to the monitoring data comprises:
  • the step of acquiring a plane parameter of the gravity direction and the forward direction of the vehicle to form a plane comprises:
  • a plane parameter of a plane of gravity and a direction of advancement of the vehicle is obtained by a multiple linear regression fitting method.
  • the step of acquiring a plane parameter of the gravity direction and the forward direction of the vehicle to form a plane comprises:
  • a planar parameter of a plane is formed by acquiring a preset gravity direction and a forward direction of the vehicle from a server or a local according to information of the vehicle.
  • the present invention further provides an identification device for a sharp turn of a vehicle, wherein the device for identifying a sharp turn of the vehicle includes:
  • a data acquisition module configured to acquire monitoring data returned by an acceleration sensor carried by the vehicle
  • a lateral acceleration acquisition module configured to calculate, according to the monitoring data, a lateral acceleration of the vehicle by using a preset algorithm, where a direction of the lateral acceleration is a normal vector of a plane formed by a gravity direction and a forward direction of the vehicle Consistent in direction;
  • a judging module configured to determine, according to the lateral acceleration, whether the vehicle has a sharp turn.
  • the determining module comprises:
  • a lateral acceleration determining module configured to determine whether the lateral acceleration is greater than a preset threshold, and if yes, determine that the vehicle has a sharp turn; otherwise, determine that the vehicle has not made a sharp turn;
  • the lateral acceleration acquisition module comprises:
  • a plane parameter obtaining unit configured to acquire a plane parameter of a plane of the gravity direction and the forward direction of the vehicle
  • a lateral acceleration obtaining unit configured to obtain a lateral acceleration of the vehicle by using a preset algorithm according to the plane parameter and the monitoring data.
  • the plane parameter obtaining unit includes:
  • a component data acquiring unit configured to acquire, according to the monitoring data, three-axis component data of the noise-reduced acceleration sensor
  • the plane parameter unit is configured to determine, according to the three-axis component data of the noise-reduced acceleration sensor, a plane parameter of a plane of gravity and a direction of advancement of the vehicle by a multiple linear regression fitting method.
  • the plane parameter obtaining unit includes:
  • a preset plane parameter acquiring unit configured to acquire, according to information of the vehicle, a preset plane parameter of the planar direction of the gravity direction and the forward direction of the vehicle from a server or a locality.
  • the method and device for identifying a sharp turn of a vehicle obtains the lateral acceleration of the vehicle through the monitoring data returned by the acceleration sensor, and simultaneously judges the lateral acceleration to determine whether the preset threshold value is exceeded and determines
  • the method for whether the vehicle has a sharp turn achieves a reduction in vehicle cost, and can accurately and timely determine that the vehicle has made a sharp turn when the vehicle makes a sharp turn.
  • FIG. 1 is a schematic flow chart of a first embodiment of a method for identifying a sharp turn of a vehicle according to the present invention
  • FIG. 2 is a schematic flow chart of a second embodiment of a method for identifying a sharp turn of a vehicle according to the present invention
  • FIG. 3 is a schematic diagram of functional modules of a first embodiment of a device for sharply turning a vehicle according to the present invention
  • FIG. 4 is a schematic diagram of functional modules of a second embodiment of the device for sharply turning a vehicle according to the present invention.
  • FIG. 5 is a schematic diagram of functional modules of a third embodiment of the device for sharply turning a vehicle according to the present invention.
  • FIG. 6 is a schematic diagram of functional modules of a fourth embodiment of the device for sharply turning a vehicle according to the present invention.
  • the main solution of the embodiment of the present invention is: acquiring monitoring data returned by the acceleration sensor carried by the vehicle; calculating, according to the monitoring data, a lateral acceleration of the vehicle by using a preset algorithm, the direction of the lateral acceleration and the The direction of the normal vector of the plane formed by the direction of gravity of the vehicle and the direction of advancement are consistent; and whether the vehicle has a sharp turn is determined according to the lateral acceleration.
  • the present invention provides a solution for using the acceleration sensor to monitor the lateral acceleration of the vehicle so that the identification of whether the vehicle has a sharp turn is more accurate and at a lower cost.
  • the method for identifying the sharp turn of the vehicle using the monitoring data returned by the acceleration sensor has the following problems: 1. How to obtain the vehicle coordinate system vector; 2. How to obtain the sensor chip coordinate system vector; 3. The gravity component on the slope For the influence of each axis, consider removing various external forces (mainly gravity and friction, eliminating the gravity and friction factors, and the rest of the acceleration is the actual vehicle motion acceleration) acting on each axis data, considering the actual vehicle motion acceleration (acceleration sensor display) The data does not necessarily reflect the vehicle motion state); 4, considering the acceleration sensor instantaneous value noise is large, need to be filtered; 5, lateral acceleration judgment threshold determination should consider the relationship between turning radius and speed, how to determine the judgment threshold; The problem of the accuracy of the projection of the sensor chip coordinate system to the vehicle coordinate system; 7, the side acceleration accuracy test problem of the side separation; 8. The value of the time value of the input data of the sharp turn model.
  • the plane equation perpendicular to the lateral acceleration component vector can be obtained.
  • the plane equation consisting of the direction of gravity and the direction of travel of the vehicle has the same direction of the plane normal vector and the direction of lateral acceleration. According to this scheme, the problems described in the above 1 and 2 can be avoided;
  • the embodiment of the present invention solves the instantaneous noise problem of the sensor signal by fitting the solution method
  • the embodiment of the present invention performs projection according to the spatial geometric principle, so there is no problem described in 6 and 7;
  • the direction of the plane normal vector is taken as the direction of the lateral acceleration, and the two-dimensional direction of the plane points to the direction of gravity and the direction of the vehicle respectively.
  • the plane equation based on the sensor coordinate system is first determined by calibration, so the separation plane can be considered. The method of the acceleration vector.
  • the plane equation is obtained in a fitting manner.
  • the fitted data can be obtained directly from the sensor data.
  • the plane is "flat" enough and perpendicular to gravity
  • data on the acceleration of gravity and the acceleration of the linear motion of the vehicle can be obtained to fit the direction of gravity.
  • the equation of the plane in which the straight line advances is obtained.
  • x, y, and z represent the three-axis components of the acceleration sensor, respectively.
  • the relationship between x, y, and z conforms to the binary linear regression relationship.
  • the multivariate linear regression fitting method is considered to find the plane parameters (ie, the plane normal vector), that is, the multiple linear regression parameter estimation.
  • the most commonly used method for estimating multiple linear regression parameters is the least squares method (OLS).
  • the multi-linear regression method based on least squares method can be used to obtain the plane parameters of the gravity direction and the forward direction, that is, the direction vector of the lateral acceleration and the vector composed of the plane parameters are coincident.
  • This plane parameter vector is fixed and does not change with whether it is a slope plane. It measures the plane equation parameters perpendicular to the plane where the car is traveling.
  • the vector composed of this equation parameter is the direction of the lateral acceleration vector.
  • the lateral acceleration direction vector and the sensor output data are known.
  • the projection value of the sensor data in the lateral acceleration direction vector that is, the separated lateral acceleration value, can be calculated.
  • the lateral acceleration reflects the magnitude of the centripetal force, according to the rigid body motion formula:
  • the turning radius is set to a threshold to determine whether it is a sharp turn.
  • the turning radius is set to a constant value, that is, a and w*w exhibit a positive linear relationship, that is, the amount of turning angle is reflected in the inner acceleration direction per unit time, and thus can be used as the set lateral acceleration threshold. According to the real-time size of the lateral acceleration and the threshold value, the sharp turn is judged.
  • a first embodiment of a method for identifying a sharp turn of a vehicle according to the present invention includes:
  • step S100 the monitoring data returned by the acceleration sensor mounted on the vehicle is acquired.
  • the monitoring data returned by the acceleration sensor mounted on the vehicle is acquired, and the vehicle may be equipped with a plurality of acceleration sensors, and the weighted average of the monitoring data returned by the plurality of acceleration sensors is obtained to obtain more accurate monitoring data.
  • Step S200 calculating, according to the monitoring data, a lateral acceleration of the vehicle by using a preset algorithm.
  • the vehicle lateral acceleration vector is separated by spatial conversion to the vehicle coordinate system, and the direction of the lateral acceleration is consistent with the normal vector direction of the plane formed by the gravity direction and the forward direction of the vehicle.
  • Step S300 determining whether the vehicle has a sharp turn according to the lateral acceleration.
  • a possible implementation manner is: determining whether the lateral acceleration of the vehicle is greater than a preset threshold, and if so, determining the vehicle A sharp turn occurs, otherwise, it is determined that the vehicle does not have a sharp turn; another possible implementation manner is: determining whether the increment of the lateral acceleration is greater than a preset incremental threshold according to the lateral acceleration, and if so, Then, it is determined that the vehicle has a sharp turn, otherwise, it is determined that the vehicle has not made a sharp turn.
  • the present embodiment by acquiring the monitoring data returned by the acceleration sensor mounted on the vehicle, Calculating a lateral acceleration of the vehicle, identifying whether the vehicle has a sharp turn according to the lateral acceleration, so that the identification of the sharp turn of the vehicle is more accurate and lower cost, and is intuitive from the driver and the driver Aspects of the perception identify the sharp turns of the vehicle, making it more versatile.
  • the step S300 is to determine whether the vehicle has a sharp turn according to the lateral acceleration.
  • S301 Determine whether the lateral acceleration is greater than a preset threshold. If yes, determine that the vehicle has a sharp turn; otherwise, determine that the vehicle has not made a sharp turn.
  • Determining whether the lateral acceleration is greater than a preset threshold according to the obtained lateral acceleration of the vehicle and if yes, determining that the vehicle has a sharp turn; otherwise, determining that the vehicle does not occur Sharp turn; In specific implementation, it is determined by multiple test tests that the preset threshold value ranges from 3.5G to 4.5G.
  • Step S302 determining whether the increment of the lateral acceleration is greater than a preset increment threshold according to the lateral acceleration, and if yes, determining that the vehicle has a sharp turn; otherwise, determining that the vehicle has not made a sharp turn.
  • the acceleration data when the vehicle is driving on a slope, according to the characteristics of the acceleration chip, the acceleration data includes the effect of gravity, but the actual acceleration of the vehicle on the slope should be the sum of gravity, friction, and power.
  • the acceleration sensor can't reflect the effect of friction (the friction acts on the car instead of the chip), that is to say, when calculating the lateral acceleration, the gravity has been included, but the friction is not taken into account, so consider two Scenes:
  • centripetal acceleration friction centripetal component - gravity component + dynamic component
  • centripetal acceleration sensor output value dynamic component - gravity component
  • centripetal acceleration friction centripetal component + gravity component + dynamic component
  • centripetal acceleration sensor output value dynamic component + gravity component
  • the acceleration is small (actually smaller gcos(c)*u, u is generally 0.6, c is the slope angle, the larger the angle, the smaller the gap), that is to say, a fixed one is measured on the slope according to the plane.
  • the threshold must have a larger acceleration than the actual one to determine a sharp turn, but in fact a sharp turn has already occurred. Therefore, judging from the plane-only threshold will result in a sharp turn but no report.
  • This threshold is not related to the slope, so its threshold design should be consistent with the plane acceleration increment threshold.
  • This threshold setting is also measured based on multiple tests. This threshold is an additional condition for determining the condition on the plane, and the accuracy of the judgment can be further increased.
  • the lateral acceleration of the vehicle includes:
  • Step S210 acquiring a plane parameter of the plane of gravity and the direction of advancement of the vehicle.
  • Step S220 calculating, according to the plane parameter and the monitoring data, a lateral acceleration of the vehicle by using a preset algorithm.
  • the random error should satisfy the standard Gaussian model according to the chip information provided by the chip manufacturer without considering the mechanical problem of the sensor;
  • the multivariate linear regression model based on the least squares method is suitable for the regression plane equation in the direction of gravity and forward direction.
  • the following is a derivation of the least squares multiple linear regression:
  • Q is respectively Find the first-order partial derivative and make it equal to zero, namely:
  • the parameters are x and z, and y is the direction of lateral acceleration, which is consistent with the binary linear regression model.
  • the expression of the OLS estimator of the binary linear regression model is derived.
  • acceleration sensor real-time data A (x, y, z), plane parameter vector Lateral acceleration
  • the step S210, acquiring a plane parameter of the gravity direction and the forward direction of the vehicle to form a plane includes: S211. Acquire a planar parameter of a plane from a server or a local acquisition of a preset gravity direction and a forward direction of the vehicle according to the information of the vehicle.
  • Presetting the gravity direction and the forward direction of the vehicle to form a plane parameter of the plane Presetting the gravity direction and the forward direction of the vehicle to form a plane parameter of the plane, and storing the preset plane parameter corresponding to the vehicle information in a server or the vehicle local, when needed, according to the vehicle
  • the information is obtained from the server or locally to preset the gravity direction and the forward direction of the vehicle to form a plane parameter of the plane.
  • a possible specific implementation manner is that the vehicle is tested by the automobile manufacturer, and the measured plane parameters are saved to the server or the vehicle locality, and when the vehicle starts, the vehicle is automatically read and the vehicle is automatically read. Plane parameters for information matching.
  • testing steps of the vehicle by the automobile manufacturer include: 1. collecting data of the vehicle being stationary for 5 seconds; 2. collecting 5-10 seconds of data of the straight line of the vehicle; 3. calling the library of least square plane parameter estimation. Calculate the plane parameters directly and save them.
  • the plane parameters corresponding to the vehicle information are directly acquired when used, and the vehicle is urgently increased. Turning the applicability of the turn to avoid the need for the user to perform the plane parameter operation.
  • the device for identifying a sharp turn of a vehicle includes:
  • the data acquisition module 100 is configured to acquire monitoring data returned by the acceleration sensor mounted on the vehicle.
  • the monitoring data returned by the acceleration sensor mounted on the vehicle is acquired, and the vehicle may be equipped with a plurality of acceleration sensors, and the weighted average of the monitoring data returned by the plurality of acceleration sensors is obtained to obtain more accurate monitoring data.
  • the lateral acceleration acquisition module 200 is configured to obtain a lateral acceleration of the vehicle by using a preset algorithm according to the monitoring data.
  • the vehicle lateral acceleration vector is separated by spatial conversion to the vehicle coordinate system, and the direction of the lateral acceleration is consistent with the normal vector direction of the plane formed by the gravity direction and the forward direction of the vehicle.
  • the determining module 300 is configured to determine, according to the lateral acceleration, whether the vehicle has a sharp turn.
  • a possible implementation manner is: determining whether the lateral acceleration of the vehicle is greater than a preset threshold, and if so, determining the vehicle A sharp turn occurs, otherwise, it is determined that the vehicle does not have a sharp turn; another possible implementation manner is: determining whether the increment of the lateral acceleration is greater than a preset incremental threshold according to the lateral acceleration, and if so, Then, it is determined that the vehicle has a sharp turn, otherwise, it is determined that the vehicle has not made a sharp turn.
  • the embodiment by obtaining the monitoring data returned by the acceleration sensor mounted on the vehicle, calculating the lateral acceleration of the vehicle, and identifying whether the vehicle has a sharp turn according to the lateral acceleration, so that The identification of sharp turns of the vehicle is more accurate and less costly, and the sharp turn of the vehicle is recognized from the perspective of the driver and the driver, making the scope of application wider.
  • the determining module 300 includes: a lateral acceleration determining module 301, configured to determine Whether the lateral acceleration is greater than a preset threshold, and if so, determining that the vehicle is in an emergency Turning, otherwise, it is determined that the vehicle has not made a sharp turn.
  • Determining whether the lateral acceleration is greater than a preset threshold according to the obtained lateral acceleration of the vehicle and if yes, determining that the vehicle has a sharp turn; otherwise, determining that the vehicle does not occur Sharp turn; In specific implementation, it is determined by multiple test tests that the preset threshold value ranges from 3.5G to 4.5G.
  • the lateral acceleration determining module 301 is further configured to determine, according to the lateral acceleration, whether the increment of the lateral acceleration is greater than a preset incremental threshold, and if yes, determine that the vehicle has a sharp turn; otherwise, determine the vehicle No sharp turns occurred.
  • the acceleration data when the vehicle is driving on a slope, according to the characteristics of the acceleration chip, the acceleration data includes the effect of gravity, but the acceleration of the actual driving of the vehicle on the slope should be the sum of gravity, friction, and power.
  • the acceleration sensor can't reflect the effect of friction (the friction acts on the car instead of the chip), that is to say, when calculating the lateral acceleration, the gravity has been included, but the friction is not taken into account, so consider two Scenes:
  • centripetal acceleration friction centripetal component - gravity component + dynamic component
  • centripetal acceleration sensor output value dynamic component - gravity component
  • centripetal acceleration friction centripetal component + gravity component + dynamic component
  • centripetal acceleration sensor output value dynamic component + gravity component
  • the acceleration is small (actually smaller gcos(c)*u, u is generally 0.6, c is the slope angle, the larger the angle, the smaller the gap), that is to say, a fixed one is measured on the slope according to the plane.
  • the threshold must have a larger acceleration than the actual one to determine a sharp turn, but in fact a sharp turn has already occurred. So root According to the plane threshold only, it will cause a sharp turn but no report.
  • This threshold is not related to the slope, so its threshold design should be consistent with the plane acceleration increment threshold.
  • This threshold setting is also measured based on multiple tests. This threshold is an additional condition for determining the condition on the plane, and the accuracy of the judgment can be further increased.
  • a third embodiment of the device for identifying a sharp turn of a vehicle according to the present invention based on the embodiment shown in FIG. 3, the lateral acceleration acquiring module 200 includes:
  • the plane parameter acquiring unit 210 is configured to acquire a plane parameter of the plane of gravity and the direction of advancement of the vehicle.
  • the lateral acceleration obtaining unit 220 is configured to obtain a lateral acceleration of the vehicle by using a preset algorithm according to the plane parameter and the monitoring data.
  • the random error should satisfy the standard Gaussian model according to the chip information provided by the chip manufacturer without considering the mechanical problem of the sensor;
  • the multivariate linear regression model based on the least squares method is suitable for the regression plane equation in the direction of gravity and forward direction.
  • the following is a derivation of the least squares multiple linear regression:
  • Q is respectively Find the first-order partial derivative and make it equal to zero, namely:
  • the parameters are x and z, and y is the direction of lateral acceleration, which is consistent with the binary linear regression model.
  • the expression of the OLS estimator of the binary linear regression model is derived.
  • acceleration sensor real-time data A (x, y, z), plane parameter vector Lateral acceleration
  • the plane parameter obtaining unit 210 includes: a preset plane parameter acquiring unit 211,
  • the plane parameters of the plane are formed by acquiring the preset gravity direction and the forward direction of the vehicle from the server or the local according to the information of the vehicle.
  • Presetting the gravity direction and the forward direction of the vehicle to form a plane parameter of the plane Presetting the gravity direction and the forward direction of the vehicle to form a plane parameter of the plane, and storing the preset plane parameter corresponding to the vehicle information in a server or the vehicle local, when needed, according to the vehicle
  • the information is obtained from the server or locally to preset the gravity direction and the forward direction of the vehicle to form a plane parameter of the plane.
  • a possible specific implementation manner is that the vehicle is tested by the automobile manufacturer, and the measured plane parameters are saved to the server or the vehicle locality, and when the vehicle starts, the vehicle is automatically read and the vehicle is automatically read. Plane parameters for information matching.
  • testing steps of the vehicle by the automobile manufacturer include: 1. collecting data of the vehicle being stationary for 5 seconds; 2. collecting 5-10 seconds of data of the straight line of the vehicle; 3. calling the library of least square plane parameter estimation. Calculate the plane parameters directly and save them.
  • the plane parameters corresponding to the vehicle information are directly acquired when used, and the vehicle is urgently increased. Turning the applicability of the turn to avoid the need for the user to perform the plane parameter operation.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Mathematical Physics (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
  • Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)

Abstract

一种车辆急转弯的识别方法,包括以下步骤:获取车辆搭载的加速度传感器返回的监测数据;根据监测数据通过预设算法计算获得车辆的侧向加速度,侧向加速度的方向与车辆的重力方向和前进方向组成的平面的法向量方向一致;根据侧向加速度判断车辆是否发生急转弯。还公开了一种车辆急转弯的识别装置。该车辆急转弯的识别方法和装置采用加速度传感器返回的监测数据获得侧向加速度,根据侧向加速度对车辆急转弯进行识别,使对车辆急转弯的识别更为准确且成本更低。

Description

车辆急转弯的识别方法及装置 技术领域
本发明涉及汽车技术领域,尤其涉及车辆急转弯的识别方法及装置。
背景技术
在智能交通系统中,辅助驾驶、驾驶行为分析是非常重要的一个部分。对驾驶行为的自动分析有助于发现和预防影响交通安全、造成交通事故的很多因素。其中,驾驶急转弯的识别是对司机的驾驶行为进行分析的一个重要部分,这种分析结果可以提醒驾驶员注意,避免疲劳驾驶,精力不集中等问题,同时很多保险公司也通过对驾驶急转弯等驾驶行为的分析建立投保模型。
驾驶急转弯是驾驶行为中的一种,该行为在交通法规中没有明确的规定,通常驾驶急转弯是指驾驶过程中用超过规定的速度在道路的急转弯处转弯,这种线速度可能会导致汽车的驾乘人员不舒适的感觉。驾驶急转弯最严重的危害是汽车横向加速度过大,造成车辆的侧滑甚至翻倒等交通事故。
根据现有的陀螺仪和加速度传感器检测角速度变化来识别急转弯是目前检测和识别急转弯的有效手段,其判定指标:1、角加速度:瞬间角速度≥9°/s;2、角度差:五秒内行驶方向角度差绝对值大于45°;3、三秒内转弯角度在30—45°之间(包括30°,不包括45°),且三秒内速度减少小于30%,且平均速度大于50mph;4、三秒内转弯角度在45—60°之间(包括45°,不包括60°),且三秒内速度减少小于50%,且平均速度大于45mph;5、三秒内转弯角度在60—90°之间(包括60),且三秒内速度减少小于75%,且平均速度大于30mph。
但是该方法存在以下缺陷:1、陀螺仪价格昂贵,且对于低功耗的嵌入式产品来说,陀螺仪芯片耗费较大的功率,造成发热以及资源消耗严重;2、陀螺仪具有在长时间积累误差较严重,会对根据角速度变化判断急转弯有影响;3、根据完全依赖角速度这样的标准,而没有考虑到实际司机和乘客的主观感觉,没有从直接引起驾驶安的因素——横向加速度的角度来定义驾驶急转弯,其直接的后果就是直接干扰由于驾驶急转弯导致的安全问题严重程度的判定。比如,在国内外一些公路的转弯地带为了安全和舒适考虑增加了路面的超高部分,在这些弯道转弯的时候是允许比平曲路面有跟高的角速度(或者 说线速度)的;再比如,在36mph的时候三秒内转弯角度在90°的横向力只有用144mph的速度做同样的转弯(90°)的1/4左右。换句话说如果从横向加速度的角度看,144mph的速度下其实三秒内转弯角度仅仅只需要转22.5°就会有同样的横向加速度,这说明如果从安全角度(保证汽车不侧翻,不侧滑)和驾乘人员舒适感觉方面来看,以上急转弯的定义存在严重弊端。
发明内容
本发明的主要目的在于提供一种车辆急转弯的识别方法及装置,旨在解决采用陀螺仪通过检测角速度变化识别车辆急转弯误差大,且成本较高的技术问题。
为实现上述目的,本发明提供一种车辆急转弯的识别方法,包括:
获取车辆搭载的加速度传感器返回的监测数据;
获取所述车辆的重力方向和前进方向组成平面的平面参数;
根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度,所述侧向加速度的方向与所述车辆的重力方向和前进方向组成的平面的法向量方向一致;
判断所述侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;
或根据所述侧向加速度判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
优选地,所述获取所述车辆的重力方向和前进方向组成平面的平面参数的步骤包括:
根据所述监测数据获取降噪后的加速度传感器的三轴分量数据;
根据所述降噪后的加速度传感器的三轴分量数据,通过多元线性回归拟合方法求出所述车辆的重力方向和前进方向组成平面的平面参数。
优选地,所述获取所述车辆的重力方向和前进方向组成平面的平面参数的步骤包括:
根据所述车辆的信息从服务器或本地获取预置的所述车辆的重力方向和前进方向组成平面的平面参数。
本发明还提供了一种车辆急转弯的识别方法,包括:
获取车辆搭载的加速度传感器返回的监测数据;
根据所述监测数据通过预设算法计算获得所述车辆的侧向加速度,所述侧向加速度的方向与所述车辆的重力方向和前进方向组成的平面的法向量方向一致;
根据所述侧向加速度判断所述车辆是否发生急转弯。
优选地,所述根据所述侧向加速度判断所述车辆是否发生急转弯的步骤包括:
判断所述侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;
或根据所述侧向加速度判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
优选地,所述根据所述监测数据通过预设算法计算获得所述车辆的侧向加速度的步骤包括:
获取所述车辆的重力方向和前进方向组成平面的平面参数;
根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度。
优选地,所述获取所述车辆的重力方向和前进方向组成平面的平面参数的步骤包括:
根据所述监测数据获取降噪后的加速度传感器的三轴分量数据;
根据所述降噪后的加速度传感器的三轴分量数据,通过多元线性回归拟合方法求出所述车辆的重力方向和前进方向组成平面的平面参数。
优选地,所述获取所述车辆的重力方向和前进方向组成平面的平面参数的步骤包括:
根据所述车辆的信息从服务器或本地获取预置的所述车辆的重力方向和前进方向组成平面的平面参数。
此外,为实现上述目的,本发明还提供一种车辆急转弯的识别装置,所述车辆急转弯的识别装置包括:
数据获取模块,用于获取车辆搭载的加速度传感器返回的监测数据;
侧向加速度获取模块,用于根据所述监测数据通过预设算法计算获得所述车辆的侧向加速度,所述侧向加速度的方向与所述车辆的重力方向和前进方向组成的平面的法向量方向一致;
判断模块,用于根据所述侧向加速度判断所述车辆是否发生急转弯。
优选地,所述判断模块包括:
侧向加速度判断模块,用于判断所述侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;
还用于根据所述侧向加速度判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
优选地,所述侧向加速度获取模块包括:
平面参数获取单元,用于获取所述车辆的重力方向和前进方向组成平面的平面参数;
侧向加速度获得单元,用于根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度。
优选地,所述平面参数获取单元包括:
分量数据获取单元,用于根据所述监测数据获取降噪后的加速度传感器的三轴分量数据;
求解平面参数单元,用于根据所述降噪后的加速度传感器的三轴分量数据,通过多元线性回归拟合方法求出所述车辆的重力方向和前进方向组成平面的平面参数。
优选地,所述平面参数获取单元包括:
预置平面参数获取单元,用于根据所述车辆的信息从服务器或本地获取预置的所述车辆的重力方向和前进方向组成平面的平面参数。
本发明实施例提出的一种车辆急转弯的识别方法及装置,通过加速度传感器返回的监测数据,计算获得车辆的侧向加速度,同时对侧向加速度进行判断,判断是否超过预设阀值从而确定所述车辆是否发生急转弯的方法,实现了对降低车辆成本,且当车辆发生急转弯时能够准确及时的判断出所述车辆已经发生急转弯。
附图说明
图1为本发明车辆急转弯的识别方法第一实施例的流程示意图;
图2为本发明车辆急转弯的识别方法第二实施例的流程示意图;
图3为本发明车辆急转弯的识别装置第一实施例的功能模块示意图;
图4为本发明车辆急转弯的识别装置第二实施例的功能模块示意图;
图5为本发明车辆急转弯的识别装置第三实施例的功能模块示意图;
图6为本发明车辆急转弯的识别装置第四实施例的功能模块示意图。
具体实施方式
应当理解,此处所描述的具体实施例仅仅用以解释本发明,并不用于限定本发明。
本发明实施例的主要解决方案是:获取车辆搭载的加速度传感器返回的监测数据;根据所述监测数据通过预设算法计算获得所述车辆的侧向加速度,所述侧向加速度的方向与所述车辆的重力方向和前进方向组成的平面的法向量方向一致;根据所述侧向加速度判断所述车辆是否发生急转弯。
由于现有技术采用陀螺仪检测角速度的方式判断车辆是否发生急转弯方法,成本高且由于陀螺仪作为惯性器件存在长时间使用后监测数据误差变大,存在对车辆是否发生急转弯判断不准确的问题。
本发明提供一种解决方案,使用加速度传感器监测车辆侧向加速度的方法,使得对车辆是否发生急转弯的识别更为准确,同时成本更低。
需要说明的,使用加速度传感器返回的监测数据对车辆急转弯进行识别的方法,存在如下问题:1、如何获取车辆坐标系向量;2、如何获取传感器芯片坐标系向量;3、斜坡上因为重力分量对各轴的影响,考虑去除各种外力(主要是重力和摩擦力,消除重力和摩擦力因素其余的加速度就是实际车辆运动加速度)对各轴数据作用,考虑实际车辆运动加速度(加速度传感器显示的数据并不一定反应车辆运动状态);4、考虑加速度传感器瞬时值噪声较大,需要进行滤波处理;5、侧向加速度判断阈值确定应该考虑转弯半径、速度的关系,如何确定判断阈值;6、传感器芯片坐标系向车辆坐标系投影准确性问题;7、侧分离出来的侧向加速度准确性测试问题;8、急转弯模型输入数据时间duration取值问题。
作为一种优选地的方法,针对上述1和2所述的问题,本发明实施例根据直接求侧向加速度向量方向的方法,可以求出与侧向加速度分量向量垂直的平面方程,即求出在理想状况下,重力方向和车辆前进方向组成的平面方程,其平面法向量方向和侧向加速度方向一致,根据此方案可以避免上述1和2所述的问题;
针对上述3-8所述的问题,因为在斜坡上重力会反映在加速度数据里面, 而摩擦力不会反映在加速度数据里面,所以本发明实施例在投影之前,去除重力数据的影响;
针对上述4所述的问题,本发明实施例通过拟合求解的方法解决传感器信号的瞬时噪声问题;
考虑到上述5所述的问题,在测定加速度阈值时,应该考虑车辆速度和半径的影响(a=v*v/r),可以考虑在同一个半径下测定侧向加速度阈值;
针对6和7所述的传感器数据向车辆坐标系投影准确性测定的问题,本发明实施例根据空间几何原理进行投影,因此不存在6和7所述的问题;
针对于8所述的急转弯模型输入数据时间duration取值的问题,在计算侧向加速度时,持续时间越长,在这段时间转弯值应该越大,因此本发明实施例通过多次试验预设合理的duration取值。
在理想状况下,我们可以分离出一个特定的平面和基于这个平面的法向量。这里把平面法向量的方向作为侧向加速度的方向,而平面二维方向分别指向重力方向和车辆前进方向,通过校准先确定在基于传感器坐标系下的平面方程,所以可以考虑分离平面求出侧向加速度向量的方法。
根据上述4所述的问题,以一种拟合的方式求出平面方程。拟合的数据可以直接从传感器数据获取,在理想状况下(平面足够“平”,并与重力垂直),可以获取关于在重力加速度和车辆直线运动加速度情况下的数据,用来拟合重力方向和直线前进方向所在平面的方程。
设定重力方向和直线前进方向所在平面的一般平面方程式为:
y=β01x+β2z
其中x、y、z分别代表加速度传感器的三轴分量。显而易见,x、y、z之间关系符合二元线性回归关系。考虑多元线性回归拟合方法求出平面参数(即平面法向量),也即多元线性回归参数估计。多元线性回归参数估计目前最常用的方法是最小二乘法(OLS)。
利用基于最小二乘法多元线性回归方法可以得出重力方向和前进方向所在平面参数,也就是侧向加速度的方向矢量和平面参数组成的矢量是重合的。这种平面参数矢量是固定的,不随着是否是坡度平面而变化,它衡量了与汽车行驶所在地面垂直的平面方程参数,这个方程参数组成的向量就是侧向加速度矢量方向。
在传感器芯片坐标系下,已知侧向加速度方向矢量和传感器输出数据, 根据一个矢量向另一个投影方法,可以算出传感器数据在侧向加速度方向矢量的投影值也就是分离出来的侧向加速度值。
已知汽车在任何路面的分离的侧向加速度值,则可以对急转弯进行判断方法定义。在理论上,侧向加速度反映了向心力的大小,根据刚体运动公式:
Figure PCTCN2016105538-appb-000001
可以看出在转弯半径一定的情况下,侧向加速度越大,角速度越大,角速度反映了单位时间转弯角度,可以设定侧向加速度一个阈值来判定是否是急转弯。在实际应用中,把转弯半径设置成一个定常值,也就是a与w*w呈现正线性关系,即在单位时间内侧向加速度大小反映转弯角度程度,从而可以用来作为设定侧向加速度阈值、根据侧向加速度实时大小与阈值比较来判断急转弯。
参照图1,为本发明车辆急转弯的识别方法的第一实施例,所述车辆急转弯的识别方法包括:
步骤S100,获取车辆搭载的加速度传感器返回的监测数据。
车辆行驶时,获取车辆所搭载的加速度传感器返回的监测数据,所述车辆可以搭载多个加速度传感器,通过获取多个加速度传感器返回的监测数据进行加权平均后获得更为准确的监测数据。
步骤S200,根据所述监测数据通过预设算法计算获得所述车辆的侧向加速度。
根据所述监测数据通过空间转换到车辆坐标系下,分离出车辆侧向加速度向量,所述侧向加速度的方向与所述车辆的重力方向和前进方向组成的平面的法向量方向一致。
步骤S300,根据所述侧向加速度判断所述车辆是否发生急转弯。
根据所述车辆侧向加速度对车辆是否发生急转弯进行判断,可以预见的,一种可能的实施方式为:判断所述车辆的侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;另一种可能的实施方式为:根据所述侧向加速度判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
在本实施例中,通过获取车辆所搭载的加速度传感器返回的监测数据, 计算获得所述车辆的侧向加速度,根据所述侧向加速度对所述车辆是否发生急转弯进行识别,使得对所述车辆急转弯的识别更为准确且成本更低,同时从驾乘人员直观感受的方面对车辆急转弯进行识别,使得适用范围更广。
进一步的,为本发明车辆急转弯的识别方法的第二实施例,基于上述图1所示的实施例,所述步骤S300,根据所述侧向加速度判断所述车辆是否发生急转弯包括:步骤S301,判断所述侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
根据获取的所述车辆的侧向加速度与预设阀值进行比较,判断所述侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;具体实施时,通过多次试验测试确定,所述预设阀值取值范围在3.5G-4.5G之间较为合适。
步骤S302,根据所述侧向加速度判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
根据获取的所述车辆的侧向加速度的增量与预设阀值进行比较,判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;具体实施时,通过多次试验测试确定,所述预设阀值选取0.8G较为合适。
针对本步骤,具体实施时,所述车辆在斜坡上行驶时,根据加速度芯片特性,加速度数据包含了重力的作用,但在斜坡上车辆实际行驶的加速度应该是重力、摩擦力、动力作用之和,而加速度传感器无法反映出摩擦力的作用(摩擦力作用于车而不是芯片),也就是说在计算侧向加速度时,重力作用已经包含进去了但是摩擦力作用并未考虑进去,所以考虑两个场景:
1、车辆下坡突然急转弯:此时向心加速度实际数据值=摩擦力向心分量-重力分量+动力分量,但向心加速度传感器输出值=动力分量-重力分量。也就是用传感器输出值实时分离出来的向心加速度偏小(小于在平面的情况下);
2、车辆上坡突然急转弯:此时向心加速度实际数据值=摩擦力向心分量+重力分量+动力分量,但向心加速度传感器输出值=动力分量+重力分量。也就是用传感器输出值实时分离出来的向心加速度仍然偏小(小于在平面的情况下)。
所以斜坡上测定的阈值应该小于在平面上测定的加速度阈值。因为A=(x,y,z)实际包含重力作用,但是并没有包含摩擦的作用,摩擦力的作用起正作用;相对于平面而言,A变小了,斜坡上实时计算出来的侧向加速度偏小(实际小了gcos(c)*u,u一般取0.6,c为斜坡角度,角度越大,差距越小),也就是说在斜坡上按照平面上测出来的给定一个固定的阈值必须要有比实际的大一些的加速度才能判定为急转弯,但是实际上早已经发生急转弯了。所以根据只平面阈值来判断会造成已发生急转弯但没报告的情况。鉴于此,我们考虑增加一个在单位时间内侧向加速度增量阈值来判断在斜坡上急转弯判定情况。这个增量阈值是跟坡度没有关系的,所以其阈值设计应该和平面加速度增量阈值一致。这个阈值的设定也是根据多次试验测量出来的。这个阈值作为在平面上判定条件的附加条件,可以进一步增加判断的准确性。
在本实施例中,通过获得的所述车辆的侧向加速度与预设阀值进行对比或通过所述车辆的侧向加速度的增量与预设增量阀值进行对比,判断所述车辆是否发生急转弯,使得对于所述车辆急转弯的识别更准确,且在所述车辆在斜坡上行驶时,依然能够对所述车辆是否发生急转弯做出准确判断。
进一步的,参照图2,为本发明车辆急转弯的识别方法的第三实施例,基于上述图1所示的实施例,所述步骤S200,根据所述监测数据通过预设算法计算获得所述车辆的侧向加速度包括:
步骤S210,获取所述车辆的重力方向和前进方向组成平面的平面参数。
根据所述监测数据获取降噪后的加速度传感器的三轴分量数据,根据所述降噪后的加速度传感器的三轴分量数据,通过多元线性回归拟合方法求出所述车辆的重力方向和前进方向组成平面的平面参数。
步骤S220,根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度。
根据所述平面参数和所述监测数据通过预设的多元线性回归模型算法计算获得所述车辆的侧向加速度。本实施例具体实施时,采用多元线性回归模型利用普通最小二乘法(OLS)对参数进行估计时,有如下假定:
1:零均值假定:E(μi)=0,i=1,2,…,n;
2:同方差假定(μ的方差为同一常数)Var(μi)=E(μi 2)=σ2,(i=1,2,…,n);
3:无自相关性:Cov(μi,μj)=E(μiμj)=0,(i≠j,i,j=1,2,…,n);
4:随机误差项μ与解释变量X不相关(这个假定自动成立): Cov(Xji,μi)=0,(j=1,2,…,k,i=1,2,…,n);
5:随机误差项μ服从均值为零,方差为σ2的正态分布:
Figure PCTCN2016105538-appb-000002
6:解释变量之间不存在多重共线性:rank(X)=k+1≤n;
根据上述六个假定,在本设计中平面方程y=β01x+β2z,现在分析其对六个假定满足情况:
(1)根据假定1,随机误差项在样本足够多的时候,在不考虑传感器机械问题情况下,根据芯片厂商提供的芯片信息,随机误差应该满足标准高斯模型,其期望满足:E(μi)=0,i=1,2,…,n,n取1;
(2)根据假定2,在样本足够多的时候,在不考虑传感器机械问题情况下,根据芯片厂商提供的芯片信息,随机误差应该满足标准高斯模型,其方差满足:Var(μi)=E(μi 2)=σ2,(i=1,2,…,n),n=1;
(3)根据假定3,在样本足够多的时候,在不考虑传感器机械问题情况下,根据芯片厂商提供的芯片信息,随机误差间是自不相关;
(4)假定4自动成立;
(5)在样本足够多的时候,在不考虑传感器机械问题情况下,根据芯片厂商提供的芯片信息,随机误差应该满足标准高斯模型;
(6)变量之间x和z是相互垂直的,也就不存在多重共线性问题。
根据上述多元线性回归模型的假定判断,基于最小二乘法多元线性回归模型是适用于在求重力与前进方向的回归平面方程的。下面对最小二乘法多元线性回归推导:
对于含有k个解释变量的多元线性回归模型:
Yi=β01X1i2X2i+…+βkXkii    (i=1,2,…,n),
Figure PCTCN2016105538-appb-000003
分别作为参数
Figure PCTCN2016105538-appb-000004
的估计量,得样本回归方程为:
Figure PCTCN2016105538-appb-000005
观测值Yi与回归值
Figure PCTCN2016105538-appb-000006
的残差ei为:
Figure PCTCN2016105538-appb-000007
由最小二乘法可知
Figure PCTCN2016105538-appb-000008
应使全部观测值Yi与回归值
Figure PCTCN2016105538-appb-000009
的残差ei的平方和最小,即使:
Figure PCTCN2016105538-appb-000010
公式1
取得最小值。根据多元函数的极值原理,Q分别对
Figure PCTCN2016105538-appb-000011
求一阶偏导,并令其等于零,即:
Figure PCTCN2016105538-appb-000012
即:
Figure PCTCN2016105538-appb-000013
化简得下列方程组:
Figure PCTCN2016105538-appb-000014
上述(k+1)个方程称为正规方程,其矩阵形式为:
Figure PCTCN2016105538-appb-000015
因为:
Figure PCTCN2016105538-appb-000016
Figure PCTCN2016105538-appb-000017
为估计值向量,样本回归模型
Figure PCTCN2016105538-appb-000018
两边同乘样本观测值矩阵X 的转置矩阵Xt,则有:
Figure PCTCN2016105538-appb-000019
得正规方程组:
Figure PCTCN2016105538-appb-000020
由假定(6),R(X)=k+1,XtX为(k+1)阶方阵,所以XtX满秩,XtX的逆矩阵(XtX)-1存在。因而:
Figure PCTCN2016105538-appb-000021
则为向量β的OLS估计量。
根据最小二乘法的多元线性回归模型,参数为x和z,y为侧向加速度方向,符合二元线性回归模型。根据上面推导,导出二元线性回归模型的OLS估计量的表达式。由公式1得二元线性回归模型为:Yi=β01X1i2X2ii;为了计算的方便,先将模型中心化:
Figure PCTCN2016105538-appb-000022
Lpq=Σxpixqi,(p,q=1,2)LjY=Σxjiyi,(j=1,2)
Figure PCTCN2016105538-appb-000023
Figure PCTCN2016105538-appb-000024
则二元回归模型改写为中心化模型:Yi=α01x1i2x2ii,记:
Figure PCTCN2016105538-appb-000025
Figure PCTCN2016105538-appb-000026
将Lpq=Σxpixqi,(p,q=1,2)代入得:
Figure PCTCN2016105538-appb-000027
因为
Figure PCTCN2016105538-appb-000028
则:
Figure PCTCN2016105538-appb-000029
由公式2得:
Figure PCTCN2016105538-appb-000030
公式3
其中:
Figure PCTCN2016105538-appb-000031
由公式3可知:
Figure PCTCN2016105538-appb-000032
得:
Figure PCTCN2016105538-appb-000033
向量
Figure PCTCN2016105538-appb-000034
即为重力和前进所在平面的平面参数,同时也是平面法向量,侧向加速度矢量方向和向量
Figure PCTCN2016105538-appb-000035
方向一致。根据同坐标系下,传感器数据在a上投影方法,得出侧向加速度的值。
假设加速度传感器实时数据A=(x,y,z),平面参数向量
Figure PCTCN2016105538-appb-000036
则侧向加速度
Figure PCTCN2016105538-appb-000037
本实施例中,通过获取所述车辆所搭载的加速度传感器返回的监测数据的三轴分量数据,并通过多元线性回归拟合方法计算求出所述车辆的重力方向和前进方向组成平面的平面参数,并根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度,为识别所述车辆是否急转弯提供了车辆侧向加速度的值,使得能够准确的识别所述车辆是否发生急转弯。
进一步的,本发明车辆急转弯的识别方法的第四实施例,基于上述图3所述的实施例,所述步骤S210,获取所述车辆的重力方向和前进方向组成平面的平面参数包括:步骤S211,根据所述车辆的信息从服务器或本地获取预置的所述车辆的重力方向和前进方向组成平面的平面参数。
预置所述车辆的重力方向和前进方向组成平面的平面参数,将所述预置的平面参数与所述车辆信息相对应保存在服务器或者所述车辆本地,在需要使用时,根据所述车辆的信息从服务器或本地获取预置的所述车辆的重力方向和前进方向组成平面的平面参数。
具体实施时,一种可能的具体实施方式为由汽车生产厂商对车辆进行测试,并将测得的所述平面参数保存到服务器或者所述车辆本地,车辆启动时,自动读取与所述车辆信息匹配的平面参数。
进一步的,所述由汽车生产厂商对车辆进行测试步骤包括:1、采集汽车静止5秒的数据;2、采集汽车直线前进的5-10秒数据;3、调用最小二乘平面参数估计的库直接计算出平面参数并保存。
在本实施例中,通过提前预置所述车辆对应的平面参数,并将所述平面参数保存在服务器或本地,需要使用时直接获取与所述车辆信息相对应的平面参数,增加了车辆急转弯识别适用性,避免用户需要进行获取平面参数操作。
参照图3,为本发明车辆急转弯的识别装置的第一实施例,所述车辆急转弯的识别装置包括:
数据获取模块100,用于获取车辆搭载的加速度传感器返回的监测数据。
车辆行驶时,获取车辆所搭载的加速度传感器返回的监测数据,所述车辆可以搭载多个加速度传感器,通过获取多个加速度传感器返回的监测数据进行加权平均后获得更为准确的监测数据。
侧向加速度获取模块200,用于根据所述监测数据通过预设算法计算获得所述车辆的侧向加速度。
根据所述监测数据通过空间转换到车辆坐标系下,分离出车辆侧向加速度向量,所述侧向加速度的方向与所述车辆的重力方向和前进方向组成的平面的法向量方向一致。
判断模块300,用于根据所述侧向加速度判断所述车辆是否发生急转弯。
根据所述车辆侧向加速度对车辆是否发生急转弯进行判断,可以预见的,一种可能的实施方式为:判断所述车辆的侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;另一种可能的实施方式为:根据所述侧向加速度判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
在本实施例中,通过获取车辆所搭载的加速度传感器返回的监测数据,计算获得所述车辆的侧向加速度,根据所述侧向加速度对所述车辆是否发生急转弯进行识别,使得对所述车辆急转弯的识别更为准确且成本更低,同时从驾乘人员直观感受的方面对车辆急转弯进行识别,使得适用范围更广。
进一步的,参照图4,为本发明车辆急转弯的识别装置的第二实施例,基于上述图3所示的实施例,所述判断模块300包括:侧向加速度判断模块301,用于判断所述侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急 转弯,否则,判定所述车辆未发生急转弯。
根据获取的所述车辆的侧向加速度与预设阀值进行比较,判断所述侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;具体实施时,通过多次试验测试确定,所述预设阀值取值范围在3.5G-4.5G之间较为合适。
侧向加速度判断模块301还用于根据所述侧向加速度判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
根据获取的所述车辆的侧向加速度的增量与预设阀值进行比较,判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;具体实施时,通过多次试验测试确定,所述预设阀值选取0.8G较为合适。
针对本模块,具体实施时,所述车辆在斜坡上行驶时,根据加速度芯片特性,加速度数据包含了重力的作用,但在斜坡上车辆实际行驶的加速度应该是重力、摩擦力、动力作用之和,而加速度传感器无法反映出摩擦力的作用(摩擦力作用于车而不是芯片),也就是说在计算侧向加速度时,重力作用已经包含进去了但是摩擦力作用并未考虑进去,所以考虑两个场景:
1、车辆下坡突然急转弯:此时向心加速度实际数据值=摩擦力向心分量-重力分量+动力分量,但向心加速度传感器输出值=动力分量-重力分量。也就是用传感器输出值实时分离出来的向心加速度偏小(小于在平面的情况下);
2、车辆上坡突然急转弯:此时向心加速度实际数据值=摩擦力向心分量+重力分量+动力分量,但向心加速度传感器输出值=动力分量+重力分量。也就是用传感器输出值实时分离出来的向心加速度仍然偏小(小于在平面的情况下)。
所以斜坡上测定的阈值应该小于在平面上测定的加速度阈值。因为A=(x,y,z)实际包含重力作用,但是并没有包含摩擦的作用,摩擦力的作用起正作用;相对于平面而言,A变小了,斜坡上实时计算出来的侧向加速度偏小(实际小了gcos(c)*u,u一般取0.6,c为斜坡角度,角度越大,差距越小),也就是说在斜坡上按照平面上测出来的给定一个固定的阈值必须要有比实际的大一些的加速度才能判定为急转弯,但是实际上早已经发生急转弯了。所以根 据只平面阈值来判断会造成已发生急转弯但没报告的情况。鉴于此,我们考虑增加一个在单位时间内侧向加速度增量阈值来判断在斜坡上急转弯判定情况。这个增量阈值是跟坡度没有关系的,所以其阈值设计应该和平面加速度增量阈值一致。这个阈值的设定也是根据多次试验测量出来的。这个阈值作为在平面上判定条件的附加条件,可以进一步增加判断的准确性。
在本实施例中,通过获得的所述车辆的侧向加速度与预设阀值进行对比或通过所述车辆的侧向加速度的增量与预设增量阀值进行对比,判断所述车辆是否发生急转弯,使得对于所述车辆急转弯的识别更准确,且在所述车辆在斜坡上行驶时,依然能够对所述车辆是否发生急转弯做出准确判断。
进一步的,参照图5,为本发明车辆急转弯的识别装置的第三实施例,基于上述图3所示的实施例,所述侧向加速度获取模块200包括:
平面参数获取单元210,用于获取所述车辆的重力方向和前进方向组成平面的平面参数。
根据所述监测数据获取降噪后的加速度传感器的三轴分量数据,根据所述降噪后的加速度传感器的三轴分量数据,通过多元线性回归拟合方法求出所述车辆的重力方向和前进方向组成平面的平面参数。
侧向加速度获得单元220,用于根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度。
根据所述平面参数和所述监测数据通过预设的多元线性回归模型算法计算获得所述车辆的侧向加速度。
本实施例具体实施时,采用多元线性回归模型利用普通最小二乘法(OLS)对参数进行估计时,有如下假定:
1:零均值假定:E(μi)=0,i=1,2,…,n;
2:同方差假定(μ的方差为同一常数)Var(μi)=E(μi 2)=σ2,(i=1,2,…,n);
3:无自相关性:Cov(μi,μj)=E(μiμj)=0,(i≠j,i,j=1,2,…,n);
4:随机误差项μ与解释变量X不相关(这个假定自动成立):Cov(Xji,μi)=0,(j=1,2,…,k,i=1,2,…,n);
5:随机误差项μ服从均值为零,方差为σ2的正态分布:
Figure PCTCN2016105538-appb-000038
6:解释变量之间不存在多重共线性:rank(X)=k+1≤n;
根据上述六个假定,在本设计中平面方程y=β01x+β2z,现在分析其对 六个假定满足情况:
(1)根据假定1,随机误差项在样本足够多的时候,在不考虑传感器机械问题情况下,根据芯片厂商提供的芯片信息,随机误差应该满足标准高斯模型,其期望满足:E(μi)=0,i=1,2,…,n,n取1;
(2)根据假定2,在样本足够多的时候,在不考虑传感器机械问题情况下,根据芯片厂商提供的芯片信息,随机误差应该满足标准高斯模型,其方差满足:Var(μi)=E(μi 2)=σ2,(i=1,2,…,n),n=1;
(3)根据假定3,在样本足够多的时候,在不考虑传感器机械问题情况下,根据芯片厂商提供的芯片信息,随机误差间是自不相关;
(4)假定4自动成立;
(5)在样本足够多的时候,在不考虑传感器机械问题情况下,根据芯片厂商提供的芯片信息,随机误差应该满足标准高斯模型;
(6)变量之间x和z是相互垂直的,也就不存在多重共线性问题。
根据上述多元线性回归模型的假定判断,基于最小二乘法多元线性回归模型是适用于在求重力与前进方向的回归平面方程的。下面对最小二乘法多元线性回归推导:
对于含有k个解释变量的多元线性回归模型:
Yi=β01X1i2X2i+…+βkXkii    (i=1,2,…,n),
Figure PCTCN2016105538-appb-000039
分别作为参数
Figure PCTCN2016105538-appb-000040
的估计量,得样本回归方程为:
Figure PCTCN2016105538-appb-000041
观测值Yi与回归值
Figure PCTCN2016105538-appb-000042
的残差ei为:
Figure PCTCN2016105538-appb-000043
由最小二乘法可知
Figure PCTCN2016105538-appb-000044
应使全部观测值Yi与回归值
Figure PCTCN2016105538-appb-000045
的残差ei的平方和最小,即使:
Figure PCTCN2016105538-appb-000046
公式1
取得最小值。根据多元函数的极值原理,Q分别对
Figure PCTCN2016105538-appb-000047
求一阶偏导,并令其等于零,即:
Figure PCTCN2016105538-appb-000048
即:
Figure PCTCN2016105538-appb-000049
化简得下列方程组:
Figure PCTCN2016105538-appb-000050
上述(k+1)个方程称为正规方程,其矩阵形式为:
Figure PCTCN2016105538-appb-000051
因为:
Figure PCTCN2016105538-appb-000052
Figure PCTCN2016105538-appb-000053
为估计值向量,样本回归模型
Figure PCTCN2016105538-appb-000054
两边同乘样本观测值矩阵X的转置矩阵Xt,则有:
Figure PCTCN2016105538-appb-000055
得正规方程组:
Figure PCTCN2016105538-appb-000056
由假定6,R(X)=k+1,XtX为(k+1)阶方阵,所以XtX满秩,XtX的逆矩阵(XtX)-1存在。因而:
Figure PCTCN2016105538-appb-000057
则为向量β的OLS估计量。
根据最小二乘法的多元线性回归模型,参数为x和z,y为侧向加速度方向,符合二元线性回归模型。根据上面推导,导出二元线性回归模型的OLS估计量的表达式。由公式1得二元线性回归模型为:Yi=β01X1i2X2ii;为了计算的方便,先将模型中心化:
Figure PCTCN2016105538-appb-000058
Lpq=Σxpixqi,(p,q=1,2)  LjY=Σxjiyi,(j=1,2)
Figure PCTCN2016105538-appb-000059
Figure PCTCN2016105538-appb-000060
则二元回归模型改写为中心化模型:Yi=α01x1i2x2ii,记:
Figure PCTCN2016105538-appb-000061
将Lpq=Σxpixqi,(p,q=1,2)代入得:
Figure PCTCN2016105538-appb-000062
因为
Figure PCTCN2016105538-appb-000063
则:
Figure PCTCN2016105538-appb-000064
由公式2得:
Figure PCTCN2016105538-appb-000065
其中:
Figure PCTCN2016105538-appb-000066
由公式3可知:
Figure PCTCN2016105538-appb-000067
得:
Figure PCTCN2016105538-appb-000068
向量
Figure PCTCN2016105538-appb-000069
即为重力和前进所在平面的平面参数,同时也是平面法向量,侧向加速度矢量方向和向量
Figure PCTCN2016105538-appb-000070
方向一致。根据同坐标系下,传感器数据在a上投影方法,得出侧向加速度的值。
假设加速度传感器实时数据A=(x,y,z),平面参数向量
Figure PCTCN2016105538-appb-000071
则侧向加速度
Figure PCTCN2016105538-appb-000072
本实施例中,通过获取所述车辆所搭载的加速度传感器返回的监测数据的三轴分量数据,并通过多元线性回归拟合方法计算求出所述车辆的重力方向和前进方向组成平面的平面参数,并根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度,为识别所述车辆是否急转弯提供了车辆侧向加速度的值,使得能够准确的识别所述车辆是否发生急转弯。
进一步的,参照图6,为本发明车辆急转弯的识别装置的第四实施例,基于上述图5所述的实施例,所述平面参数获取单元210包括:预置平面参数获取单元211,用于根据所述车辆的信息从服务器或本地获取预置的所述车辆的重力方向和前进方向组成平面的平面参数。
预置所述车辆的重力方向和前进方向组成平面的平面参数,将所述预置的平面参数与所述车辆信息相对应保存在服务器或者所述车辆本地,在需要使用时,根据所述车辆的信息从服务器或本地获取预置的所述车辆的重力方向和前进方向组成平面的平面参数。
具体实施时,一种可能的具体实施方式为由汽车生产厂商对车辆进行测试,并将测得的所述平面参数保存到服务器或者所述车辆本地,车辆启动时,自动读取与所述车辆信息匹配的平面参数。
进一步的,所述由汽车生产厂商对车辆进行测试步骤包括:1、采集汽车静止5秒的数据;2、采集汽车直线前进的5-10秒数据;3、调用最小二乘平面参数估计的库直接计算出平面参数并保存。
在本实施例中,通过提前预置所述车辆对应的平面参数,并将所述平面参数保存在服务器或本地,需要使用时直接获取与所述车辆信息相对应的平面参数,增加了车辆急转弯识别适用性,避免用户需要进行获取平面参数操作。

Claims (14)

  1. 一种车辆急转弯的识别方法,其特征在于,所述车辆急转弯的识别方法包括以下步骤:
    获取车辆搭载的加速度传感器返回的监测数据;
    获取所述车辆的重力方向和前进方向组成平面的平面参数;
    根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度,所述侧向加速度的方向与所述车辆的重力方向和前进方向组成的平面的法向量方向一致;
    判断所述侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;
    或根据所述侧向加速度判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
  2. 如权利要求1所述的方法,其特征在于,所述获取所述车辆的重力方向和前进方向组成平面的平面参数的步骤包括:
    根据所述监测数据获取降噪后的加速度传感器的三轴分量数据;
    根据所述降噪后的加速度传感器的三轴分量数据,通过多元线性回归拟合方法求出所述车辆的重力方向和前进方向组成平面的平面参数。
  3. 如权利要求1所述的方法,其特征在于,所述获取所述车辆的重力方向和前进方向组成平面的平面参数的步骤包括:
    根据所述车辆的信息从服务器或本地获取预置的所述车辆的重力方向和前进方向组成平面的平面参数。
  4. 一种车辆急转弯的识别方法,其特征在于,所述车辆急转弯的识别方法包括以下步骤:
    获取车辆搭载的加速度传感器返回的监测数据;
    根据所述监测数据通过预设算法计算获得所述车辆的侧向加速度,所述侧向加速度的方向与所述车辆的重力方向和前进方向组成的平面的法向量方向一致;
    根据所述侧向加速度判断所述车辆是否发生急转弯。
  5. 如权利要求4所述的方法,其特征在于,所述根据所述侧向加速度判 断所述车辆是否发生急转弯的步骤包括:
    判断所述侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;
    或根据所述侧向加速度判断所述侧向加速度的增量是否大于预设增量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
  6. 如权利要求4所述的方法,其特征在于,所述根据所述监测数据通过预设算法计算获得所述车辆的侧向加速度的步骤包括:
    获取所述车辆的重力方向和前进方向组成平面的平面参数;
    根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度。
  7. 如权利要求6所述的方法,其特征在于,所述获取所述车辆的重力方向和前进方向组成平面的平面参数的步骤包括:
    根据所述监测数据获取降噪后的加速度传感器的三轴分量数据;
    根据所述降噪后的加速度传感器的三轴分量数据,通过多元线性回归拟合方法求出所述车辆的重力方向和前进方向组成平面的平面参数。
  8. 如权利要求6所述的方法,其特征在于,所述获取所述车辆的重力方向和前进方向组成平面的平面参数的步骤包括:
    根据所述车辆的信息从服务器或本地获取预置的所述车辆的重力方向和前进方向组成平面的平面参数。
  9. 一种车辆急转弯的识别装置,其特征在于,所述车辆急转弯的识别装置包括:
    数据获取模块,用于获取车辆搭载的加速度传感器返回的监测数据;
    侧向加速度获取模块,用于根据所述监测数据通过预设算法计算获得所述车辆的侧向加速度,所述侧向加速度的方向与所述车辆的重力方向和前进方向组成的平面的法向量方向一致;
    判断模块,用于根据所述侧向加速度判断所述车辆是否发生急转弯。
  10. 如权利要求9所述的装置,其特征在于,所述判断模块包括:
    侧向加速度判断模块,用于判断所述侧向加速度是否大于预设阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯;
    还用于根据所述侧向加速度判断所述侧向加速度的增量是否大于预设增 量阀值,若是,则判定所述车辆发生急转弯,否则,判定所述车辆未发生急转弯。
  11. 如权利要求9所述的装置,其特张在于,所述侧向加速度获取模块包括:
    平面参数获取单元,用于获取所述车辆的重力方向和前进方向组成平面的平面参数;
    侧向加速度获得单元,用于根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度。
  12. 如权利要求10所述的装置,其特张在于,所述侧向加速度获取模块包括:
    平面参数获取单元,用于获取所述车辆的重力方向和前进方向组成平面的平面参数;
    侧向加速度获得单元,用于根据所述平面参数和所述监测数据通过预设算法计算获得所述车辆的侧向加速度。
  13. 如权利要求11所述的装置,其特征在于,所述平面参数获取单元包括:
    分量数据获取单元,用于根据所述监测数据获取降噪后的加速度传感器的三轴分量数据;
    求解平面参数单元,用于根据所述降噪后的加速度传感器的三轴分量数据,通过多元线性回归拟合方法求出所述车辆的重力方向和前进方向组成平面的平面参数。
  14. 如权利要求11所述的装置,其特征在于,所述平面参数获取单元包括:
    预置平面参数获取单元,用于根据所述车辆的信息从服务器或本地获取预置的所述车辆的重力方向和前进方向组成平面的平面参数。
PCT/CN2016/105538 2016-03-10 2016-11-12 车辆急转弯的识别方法及装置 WO2017152648A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610139142.5 2016-03-10
CN201610139142.5A CN105799709B (zh) 2016-03-10 2016-03-10 车辆急转弯的识别方法及装置

Publications (1)

Publication Number Publication Date
WO2017152648A1 true WO2017152648A1 (zh) 2017-09-14

Family

ID=56468122

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/105538 WO2017152648A1 (zh) 2016-03-10 2016-11-12 车辆急转弯的识别方法及装置

Country Status (2)

Country Link
CN (1) CN105799709B (zh)
WO (1) WO2017152648A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111310125A (zh) * 2020-02-14 2020-06-19 上海本安仪表系统有限公司 一种车辆急加速急减速和急转弯的判断装置及其判断方法

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105799709B (zh) * 2016-03-10 2018-10-26 深圳市元征科技股份有限公司 车辆急转弯的识别方法及装置
CN107909678A (zh) * 2017-11-29 2018-04-13 思建科技有限公司 一种行车风险评价方法及系统
CN108068823A (zh) * 2017-12-06 2018-05-25 上海评驾科技有限公司 一种车辆驾驶行为检测方法
CN110143200A (zh) * 2019-04-01 2019-08-20 深圳市元征科技股份有限公司 一种车辆状态数据获取方法、装置、车载设备及存储介质
CN110733510B (zh) * 2019-11-22 2020-09-29 辽宁工业大学 一种用于车辆急转弯识别的检测方法
CN112776714A (zh) * 2021-01-27 2021-05-11 吉林云帆智能工程有限公司 一种用于轨道车辆的转弯监测预警方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050200088A1 (en) * 2004-03-09 2005-09-15 Mamoru Sawada Vehicle stability control system
US20120173040A1 (en) * 2009-09-24 2012-07-05 Toyota Jidosha Kabushiki Kaisha Device for estimating turning characteristic of vehicle
JP5494814B2 (ja) * 2010-09-29 2014-05-21 トヨタ自動車株式会社 車両の制御装置
CN104354699A (zh) * 2014-10-08 2015-02-18 北京远特科技有限公司 一种基于obd终端检测驾驶行为信息的方法和装置
CN105109490A (zh) * 2015-09-22 2015-12-02 厦门雅迅网络股份有限公司 一种基于三轴加速度传感器判断车辆急转弯的方法
CN105799709A (zh) * 2016-03-10 2016-07-27 深圳市元征科技股份有限公司 车辆急转弯的识别方法及装置

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100870091B1 (ko) * 2007-05-11 2008-11-25 팅크웨어(주) 센서를 이용한 회전상태 판단 방법 및 장치
CN103770644A (zh) * 2014-01-20 2014-05-07 深圳市元征科技股份有限公司 一种驾驶行为的数据获取方法和系统

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050200088A1 (en) * 2004-03-09 2005-09-15 Mamoru Sawada Vehicle stability control system
US20120173040A1 (en) * 2009-09-24 2012-07-05 Toyota Jidosha Kabushiki Kaisha Device for estimating turning characteristic of vehicle
JP5494814B2 (ja) * 2010-09-29 2014-05-21 トヨタ自動車株式会社 車両の制御装置
CN104354699A (zh) * 2014-10-08 2015-02-18 北京远特科技有限公司 一种基于obd终端检测驾驶行为信息的方法和装置
CN105109490A (zh) * 2015-09-22 2015-12-02 厦门雅迅网络股份有限公司 一种基于三轴加速度传感器判断车辆急转弯的方法
CN105799709A (zh) * 2016-03-10 2016-07-27 深圳市元征科技股份有限公司 车辆急转弯的识别方法及装置

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111310125A (zh) * 2020-02-14 2020-06-19 上海本安仪表系统有限公司 一种车辆急加速急减速和急转弯的判断装置及其判断方法
CN111310125B (zh) * 2020-02-14 2023-06-09 上海本安仪表系统有限公司 一种车辆急加速急减速和急转弯的判断方法

Also Published As

Publication number Publication date
CN105799709B (zh) 2018-10-26
CN105799709A (zh) 2016-07-27

Similar Documents

Publication Publication Date Title
WO2017152648A1 (zh) 车辆急转弯的识别方法及装置
CN106043299B (zh) 车辆控制装置
EP2828781B1 (en) A system and a method for improved car prognosis
Peng et al. Multi-parameter prediction of drivers' lane-changing behaviour with neural network model
JP6329569B2 (ja) 道路を走行中の車両の異常運転挙動を評価するシステムおよび方法
KR101069409B1 (ko) 차량 운행에 관련된 안전지수 산출 시스템 및 그 방법
US9082285B2 (en) System and method for determining vehicle driving status information-based driving concentration
KR102239992B1 (ko) 차량 충돌 경고 방법 및 장치
US20050128092A1 (en) Method and computer program for identification of inattentiveness by the driver of a vehicle
CN105109490B (zh) 一种基于三轴加速度传感器判断车辆急转弯的方法
WO2014080507A1 (ja) 車両状態判定装置、車両状態判定方法及び運転操作診断装置
US10249112B2 (en) Vehicle state monitoring apparatus, vehicle state monitoring system, and vehicle state monitoring method
EP2472493B1 (en) Driver state assessment device
RU2679299C2 (ru) Система и способ для обнаружения опасного вождения и компьютер транспортного средства
JP2019112049A (ja) 陸上車両の運転者の運転スタイルを認識するための方法及び対応する装置
JP6061782B2 (ja) 異音検知装置及びプログラム
US8994654B2 (en) Apparatus and method for controlling gaze tracking
US20200175474A1 (en) Information processing system, program, and control method
KR20140097669A (ko) 운전 패턴 인식에 기반하여 보험료를 결정하는 장치 및 방법
KR101776568B1 (ko) 차량 쏠림 검사 시스템 및 그 방법
CN105523084A (zh) 一种基于三轴加速度传感器检测车辆转弯角度的方法
Wahlström et al. Risk assessment of vehicle cornering events in GNSS data driven insurance telematics
JP2019046012A (ja) 運転者状態判定装置
CN109808641A (zh) 防盗告警方法及装置、电子设备和可读存储介质
US20210366274A1 (en) Method and device for predicting the trajectory of a traffic participant, and sensor system

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16893291

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 16893291

Country of ref document: EP

Kind code of ref document: A1