DE102014212410B4 - Method and sensor for determining a vehicle's own speed over the ground - Google Patents

Abstract

Method for determining a vehicle's own speed (Vx) over the ground, an object (6, 8, 10) located in the vicinity of the vehicle (2) being detected in several measuring cycles, with a check being made to determine whether the detected object is (6, 8, 10) is a stationary object (6, 8) or a non-stationary object (10) and in each measurement cycle a measurement data set (MI, MII, ..., Mn) with a relative speed value (Vri; Vr2) and an angle value (α1, α2) of the stationary object (6, 8) is determined in relation to the vehicle (2), the measurement data sets (MI, MII, ..., Mn) of the stationary object determined in each measurement cycle (6, 8) are assigned to clusters (C1, I; C2, I) by a cluster analysis, whereby one or more value hypotheses (H1, I; H2, I) of the Own speed (Vx) of the vehicle (2) can be determined, the own speed (Vx) of the vehicle (2) from the one or more value hypothes en (H1, I; H2, I) is determined, characterized in that to determine the vehicle's own speed (Vx) from the one or more value hypotheses (H1, I; H2, I) the one or more value hypotheses (H1, I; H2 , I) be filtered with a Kalman filter.

Description

The present invention relates to a method and a sensor for determining an airspeed of a vehicle over the ground as well as a computing unit and a computer program for its implementation.

State of the art

A vehicle's own speed over the ground can be measured using wheel speed sensors, optical sensors, GPS-based sensors or Doppler radars. However, wheel speed sensors are susceptible to slippage, optical sensors can get dirty, GPS signals are not available in a tunnel, for example, and Doppler radars are expensive, take up a lot of installation space and radar signals are falsified by the ground along the route.

According to the DE 10 2012 200 139 A1 For example, to determine the speed of a vehicle, an object in the vicinity can be detected with a frequency-modulated radar, such as an FMCW radar (frequencymodulated-continiouos-wave), and a relative speed of the detected object in relation to the vehicle can be measured. A differentiation can be made between self-moving and stationary objects. In order to determine the exact position of objects in a world coordinate system, in addition to the radial distance between the objects and the vehicle, direction information (azimuthal angle) of the objects in question in relation to the direction of travel can also be measured. By means of a plausibility check, the vehicle's own speed can then be determined via the relative speed to the detected objects, in particular to the stationary objects. However, this method is sensitive to measurement artifacts based on falsified radar signals.

There is therefore a need to further improve the determination of the speed over the ground.

From the DE 102 52 323 A1 a method for determining the proper movement of a vehicle is known, the speed of the vehicle being determined on the basis of a relative speed to a stationary object.

From the US 2002/0107637 A1 a device for monitoring a vehicle environment is known which is set up to detect objects.

From the JP 2010-43960 A a system and a method for detecting the movement of a vehicle are known. A relative speed to stationary objects is determined by means of radar sensors.

From the DE 10 2012 224 338 A1 a device and a method for estimating a speed of a vehicle are known, the speed of the vehicle being determined from detected speeds of objects in the vicinity of the vehicle that are detected by a sensor.

Disclosure of the invention

According to the invention, a method and a sensor for determining a vehicle's own speed over the ground and a computing unit with the features of the independent claims are proposed. Advantageous refinements are the subject matter of the subclaims and the description below.

Within the scope of the invention, an object located in the vicinity of the vehicle is recorded in a number of measurement cycles. It is then checked whether the detected object is a stationary object or a non-stationary object. For a stationary object, a measurement data set with a relative speed value and an angle value of the stationary object in relation to the vehicle is determined in each measurement cycle, in particular by means of a relative speed and angle determination device. The measurement data records of the stationary object determined in each measurement cycle are then assigned to clusters by a cluster analysis, with value hypotheses of the vehicle's own speed being determined by evaluating the clusters.

A cluster analysis is understood to mean a process in which similarity structures can be discovered in (large) databases using a clustering algorithm (also: cluster analysis). The values of the measurement data records are interpreted as random variables. They are usually represented in the form of vectors as points in a vector space, the dimensions of which form the properties of the object. Areas in which points accumulate (point cloud) are called clusters. In scatter diagrams, the distances between the points or the variance within a cluster serve as so-called proximity measures, which express the similarity or difference between the objects. Objects of a heterogeneous total amount (described by different values of the measurement data sets) are therefore combined with the help of the cluster analysis to form clusters which are as homogeneous as possible, ie the differences in the values of the measurement data sets are as small as possible. By evaluating the clusters, a value hypothesis of the vehicle's own speed becomes certainly. For this purpose, the respective clusters are analyzed, for example with an algorithm that determines a center point, for example a point cloud. The value corresponding to the center point is then selected as the value for the vehicle's own speed. The measurement accuracy of the determination of the vehicle's own speed over the ground is thereby improved through the evaluation of measurement data obtained in several measurement cycles and evaluated by cluster formation. The cluster formation ensures that measurement value artifacts are excluded from further processing of the measurement data and thus do not falsify the measurement result. This increases the measurement accuracy.

According to one embodiment, the measured measurement data records have a distance value of the object from the vehicle. From the relative speed value and the angle value, the vehicle's own speed in the direction of travel can be determined in a simple manner, e.g. after a transformation of the measured values of the measured data records in polar coordinates into Cartesian coordinates. The accuracy of the determination of the vehicle's own speed in the direction of travel can be further increased by taking the distance value into account.

According to the invention, the value hypotheses are filtered with a Kalman filter. A Kalman filter is understood to mean a method that is used to estimate models with parameters that vary over time. This is a recursive estimation method. By filtering with a Kalman filter, for example, interference caused by measuring devices can be removed from the measured values. The Kalman filter gates out incorrect measured values and also allows a series of measured values to be continued by predicting further measured values in the event of failure of the measuring devices. Thus the measuring accuracy is increased again.

According to a further embodiment, the value hypotheses are tracked with a tracking algorithm. For example, an α / β / γ filter or an algorithm according to the Monte Carlo method can be used as the tracking algorithm. When tracking the value hypothesis, information about the course of the value hypothesis and, on the other hand, deviations (relative error data) are extracted from a stream of value hypotheses. Interferences arise from random technical measurement errors or from unavoidable physical measurement noise. The aim of tracking the value hypothesis is to map the changes in the value hypothesis in order to merge several value hypotheses.

According to a further embodiment, an indicator is determined from each of the value hypotheses. The indicator allows a selection of a value hypothesis. For example, the indicators can be a quality value, such as a quadratic error deviation of tracked and merged value hypotheses.

According to a further embodiment, a value hypothesis is selected on the basis of the indicator. Thus, an indicator is used whose quality value indicates that this value hypothesis is to be preferred over other value hypotheses.

According to a further embodiment, the measurement data records are transmitted via an interface of a sensor of the vehicle. For example, the measurement data sets can be transferred to a human machine interface (HMI) (such as a monitor, PC, laptop, SmartDevice (tablet or phone)) via a wired connection or a wireless connection in a bus system of the vehicle (e.g. CAN). For this purpose, the sensor can have a Bluetooth interface with which a data transferring connection, e.g. a wireless data transferring connection, with e.g. a CAN bus system of the vehicle (in particular via a so-called CAN2Bluetooth module) can be established. The sensor with the interface can be supplied with electrical energy via an on-board network of the vehicle or has its own energy source, such as a battery.

A computing unit according to the invention, e.g. in a sensor, is set up, in particular in terms of programming, to carry out a method according to the invention. A sensor according to the invention has such a computing unit and a relative speed and angle determination device, in particular having a radar device.

The implementation of the method in the form of software is also advantageous, since this results in particularly low costs, in particular if an executing control device is still used for other tasks and is therefore available anyway. Suitable data carriers for providing the computer program are, in particular, floppy disks, hard disks, flash memories, EEPROMs, CD-ROMs, DVDs, etc. A program can also be downloaded via computer networks (Internet, intranet, etc.).

Further advantages and embodiments of the invention emerge from the description and the accompanying drawing.

It goes without saying that the features mentioned above and those yet to be explained below not only in the respectively specified combination, but also in other combinations or can be used on their own without departing from the scope of the present invention.

The invention is shown schematically in the drawing using an exemplary embodiment and is described in detail below with reference to the drawing.

Figure list

• 1 shows a speed measurement scenario of a vehicle in a schematic representation.
• 2 shows a measurement and data transmission arrangement in a schematic representation.
• 3 shows an evaluation process in a schematic representation.

Embodiment (s) of the invention

1 shows a speed measurement scenario in which an airspeed Vx of a vehicle 2 is measured in the direction of travel.

With the vehicle 2 In the present exemplary embodiment, it is a railroad locomotive that moves over the ground in the direction of a railroad 4th moved at the airspeed Vx. It can be the same with the vehicle 2 also be a motor vehicle or utility vehicle that is moving analogously over ground in the direction of a road at its own speed Vx.

For speed measurement of the vehicle's own speed Vx 2 instructs the vehicle 2 a sensor 12th on. The sensor 12th is present exemplary embodiment by means of a bracket (not in 1 shown) on the front of the vehicle 2 attached and is powered by a power supply unit (not in 1 shown) supplied with electrical energy. Thus, the sensor can 12th Can also be used to measure the speed of other vehicles as it is not firmly attached to the vehicle 2 is connected, but is designed to be portable. Alternatively, the sensor 2 firmly in the vehicle 2 be integrated and via an on-board network of the vehicle 2 be supplied with electrical energy.

In the present exemplary embodiment, the sensor 12th an FMCW radar (not in 1 shown) as a relative speed and angle determination device and an evaluation unit (not in 1 shown). An FMCW radar (from English frequency modulated continuous wave radar) is a modulated continuous wave radar, the transmitter of which sends out a frequency modulated signal, for example. It is therefore a continuous wave radar in which a transmitter of the continuous wave radar works continuously for the duration of a measurement process.

With the FMCW radar, an area is located in front of the vehicle in the direction of travel 2 scanned. In the present exemplary embodiment, there are three objects in this area 6th , 8th , 10 . Of these three objects 6th , 8th , 10 are the two objects 6th , 8th stationary while the third object 10 in the direction of the arrow from the railroad track 4th moved away at an airspeed Vy.

The 2 shows a preferred embodiment of the sensor 12th in detail. The sensor 12th has the holder designed here as a magnetic holder 44 with a permanent magnet, the power supply unit designed here as a battery 38 to supply the sensor 12th and its components, here the FMCW radar as a relative speed and angle determination device 40 and the evaluation unit 42 for evaluating the radar signals of the FMCW radar 40 on.

Furthermore, the 2 that the sensor 12th an interface 14th having. With the interface 14th can establish a data transmission connection with a bus system 16 of the vehicle 2 , e.g. a CAN bus. The data-transmitting connection can be wired or wireless, for example in accordance with Bluetooth. The bus system then points accordingly 16 of the vehicle 2 an interface 36 in the form, for example, of a Bluetooth adapter to establish the data-transferring connection. The bus system also has a human-machine interface (HMI) 18th on.

With the human-machine interface 18th it can be a monitor, a PC, a tablet computer or a mobile phone such as a smartphone.

In operation, the FMCW radar 40 Measurement data sets MI, MII, ..., Mn for the evaluation unit 42 transfer.

In the present exemplary embodiment, the evaluation unit 42 an arithmetic unit which is designed to determine the airspeed Vx by evaluating the measurement data sets MI, MII, ..., Mn. Alternatively, a control unit of the vehicle can also be used 2 carry out the evaluation of the measurement data records MI, MII, ..., Mn partially or completely. To this end, the evaluation unit 42 or the control unit has hardware and / or software modules.

In the present exemplary embodiment, the evaluation unit 42 to determine the airspeed Vx on the following modules, the function of which will be explained in more detail later: a measurement data processing module 22nd , a cluster module 24 , a hypothesis module 26th , a Kalman filter module 28 , a tracking module 30th , an indicator module 32 and a determination module 34 .

The 3 shows an evaluation process in a schematic representation.

The first step will be using the FMCW radar 40 the objects 6th , 8th , 10 detected. By evaluating, for example, frequency shifts in the radar signals 12th it is still determined whether the objects 6th , 8th , 10 (like the two objects 6th , 8th ) are stationary or not (like the object 10 ). Radar signals based on non-stationary objects are discarded or filtered out, while radar signals based on stationary objects are further processed as described below.

From the two stationary objects 6th , 8th be with the sensor 12th In the present exemplary embodiment, a relative speed _{value V r1} , a distance _{value A 1} and an angle value α _{1 of} the first object 6th and a relative speed _{value V r2} , a distance _{value A 2} and an angle value α _{2 of} the second object 8th in relation to the vehicle 2 measured. For example, the current alignment of the receiving antenna of the sensor is used to measure the angle values α ₁ , α ₂ 12th evaluated or the distribution of signal intensities determined by different antennas of the sensor 12th be received. In the present exemplary embodiment, these values are recorded in polar coordinates.

The relative speed values V _r1 , V _r2 , the distance values A _1, A ₂ and the angle values α ₁ , α ₂ form a first measurement data set MI.

mit

Vx,n

Zwischenwert

Vrx,n

Relativgeschwindigkeitswert

Ax,n

Abstandswert

αx,n

Winkel des Objekts zur Fahrtrichtung

φ̇n

Eigenrotation (Winkelgeschwindigkeit) des Fahrzeugs um die Vertikale

x

Index Objekt

n

Index Zyklus

In a further step, the measurement data processing module 22nd in a first measurement cycle I the first measurement data record MI of the two stationary objects 6th , 8th with the respective relative speed values Vr _{1, I} , Vr _{2, I,} the distance values A _{1, I} , A _{2, I} and the angle values α _{1, I} , α _{2, I} in relation to the vehicle 2 evaluated by an intermediate value V _{1, I} , V _{2, I} each for the two stationary objects 6th , 8th according to the following formula to determine: $$V_{x,n}=V{r}_{x,n}\cdot \text{cos}\left({\alpha }_{x,n}\right)-{\mathrm{A.}}_{x,n}\cdot {\dot{\phi }}_n\cdot \text{sin}\left({\alpha }_{x,n}\right)$$

With

Vx, n

Intermediate value

Vrx, n

Relative speed value

Ax, n

Distance value

αx, n

Angle of the object to the direction of travel

φ̇n

Self-rotation (angular speed) of the vehicle around the vertical

x

Index object

n

Index cycle

Alternatively, the intermediate values V _{x, n} can be determined without taking into account the distance values A _{x, n} , but with less accuracy using the following formula: $$V_{x,n}\approx V{r}_{x,n}\ast \text{cos}\left({\alpha }_{x,n}\right)$$

The intermediate _{values V 1, I} , V _{2, I} determined in this way are then added to the first measurement data set MI in the present exemplary embodiment.

In the further measuring cycles II, ..., Mn, further measurement data sets MII, ..., Mn are processed analogously.

Thus, the first measurement data record MI and the further measurement data records MII, ... Mn include the intermediate values V _{1, I} , V _{1, II} , ..., V _{1, n} and V _{2, I} , V _{2, II,.} .., V _{2, n} or in general V _{x, n} also values that characterize the measuring process and are therefore also of importance for the measuring accuracy, such as the respective distance values A _{1, I} , A _{1, II} , ..., A _{1, n} and A _{2, I} , A _{2, II} , ..., A _{2, n} , or generally A _{x, n} and / or the respective angle values α _{1, I} , α _{1, II} , ... , α _{1, n} and α _{2, I} , α _{2, II} , ..., α _{2, n} or in general α _{x, n} .

The measurement data sets MI, MII, ..., Mn are then in a further step with the cluster module 24 clustered according to a clustering method. For this purpose, the data of the measurement data records MI, MII,..., Mn are represented in the form of vectors as points in a vector space. After a number of measuring cycles I, II, ..., n, areas crystallize in which points accumulate (point cloud). After the first measurement cycle I, these areas form the clusters C _{1, I} , C _{2, I.}

With each measurement cycle I, II, ..., n, further measurement data sets MI, MII, ..., Mn are provided, which are clustered using the clustering method and corresponding clusters C _{1, I} , C _{1, II} , ..., C _{1, n} and C _{2, I} , C _{2, II} , ..., C _{2, n} or, in general, C _{x, n} .

In the present exemplary embodiment, measurement data based on the first object _{are in the cluster C 1, I} 6th and in the cluster C _{2, I} measurement data based on the second object 8th summarized. In other words, the clustering is unambiguous in the present exemplary embodiment.

In a further step, the hypothesis module 26th the cluster _{C1, I,} C _2, the first measurement cycle I evaluated _I in order for each cluster C _{1, I} C _{2, I} is a value hypothesis H _{1, I,} H _{2, I} based on the intrinsic speed Vx on the data of first measurement cycle I. To determine the value hypotheses H _{1, I} , H _{2, I} , for example, the center points of the point clouds of clusters C _{1, I} , C _{2, I are} determined. The center points of the respective point clouds of clusters C _{1, I} , C _{2, I are then} selected as value hypotheses H _{1, I} , H _{2, I.}

With each measurement cycle I, II, ..., n further clusters C _{1, I} , C _{1, II} , ..., C _{1, n} and C _{2, I} , C _{2, II} , ..., C _{2 , n} , based on the further value hypotheses H _{1, I} , H _{1, II} , ..., H _{1, n} and H _{2, I} , _{H2, II,.} .., H _{2, n} or generally H _{x, n are} formed.

In a further step, the value hypotheses H _{1, I} , H _{1, II} , ..., H _{1, n} and H _{2, I} , _{H2, II} , ..., H _{2, n} with the Kalman filter module 28 filtered with a Kalman filter.

The Kalman filter is a recursive filter, the value hypotheses H _{1, I} ; H _{1, II} , ..., H _{1, n} and H _{2, I} , H _{2, II} , ..., H _{2, n of} the further measuring cycles I, II, ...., n are continuously supplied, and _{the Kalman-filtered value hypotheses H 1, I} ', H _{1, II} ', ..., H _{1, n} '' and H _{2, I} ', H _{2, II} with each measurement cycle I, II, ..., n ', ..., H _{2, n} ' or generally H _{x, n} 'provides.

Filtering with the Kalman filter removes interference from the value hypotheses H _{1, I} , H _{1, II} , ..., H _{1, n} and H _{2, I} , _{H2, II} , ..., H _{2, n} which, for example, on measurement errors of the sensor 12th are based. The Kalman filter hides incorrect value hypotheses and also allows a series of value hypotheses to be continued by predicting further value hypotheses in the event of a temporary failure of the sensor 12th .

In a further step, the tracking module 30th the Kalman-filtered value hypotheses H _{1, I '} , H _{1, II} ', ..., H _{1, n} 'and H _{2, I} ', H _{2, II} ', ..., H _{2 with a tracking algorithm , n} 'tracked. In the present exemplary embodiment, the tracking of the value hypotheses H _{1, I} ', H _{1, II} ', ..., H _{1, n} 'and H _{2, I} ', H _{2, IL} ', ..., H _{2, n} 'tracked with increasing number of measurement cycles. In addition, deviations from the stream of value hypotheses can be removed. The tracking creates an image of the temporal change in the value hypotheses H _{1, I} ', H _{1, II} ', ..., H _{1, n} 'and H _{2, I} ', H _{2, II} ', ..., H _{2, n} '. An α / β / γ filter or an algorithm according to the Monte Carlo method can be used as the tracking algorithm.

In a further step, the indicator module 32 one indicator each from the tracked value hypotheses H _{1, I} ', H _{1, II} ', ..., H _{1, n} 'and H _{2, I} ', H _{2, II} ', ..., H _{2, n'} certainly. The indicator can be a figure of merit, such as a least square. However, other methods for regression analysis can also be used in addition or as an alternative.

In a further step, the determination module 34 the value selected from the value hypotheses H _{1, I} ', H _{1, II} ', ..., H _{1, n} 'and H _{2, I} ', H _{2, II} ', ..., H _{2, n} ', whose indicator shows that one value hypothesis is preferable to other value hypotheses. So shows 3 that the value hypotheses H _{1, I} ', H _{1, II} ', ..., H _{1, n} 'form a lower value hypothesis series R1, while the upper value hypotheses H _{2, I} ', H _{2, II} ', ... , H _{2, n} 'form an upper series of value hypotheses R2. Also shows 3 that the lower value hypothesis series R1 converges faster than the upper value hypothesis series R2. Correspondingly, the error square of the lower value hypothesis series R1 is smaller than the error square of the upper value hypothesis series R2. Therefore, the value hypothesis H _{1, n} 'is used as the value for the airspeed Vx of the vehicle 2 selected and as a measured value at the interface 14th issued ..

In the further course, ie when the vehicle continues to move 2 along the railroad 4th , the distance _{value A 1} and the angle value α _{1 of} the first object change 6th as well as the distance _{value A 2} and the angle value α _{2 of} the second object 8th in relation to the vehicle 2 . While in the in 1 the scenario shown is the first object 6th is in a position in which, for example, the measuring accuracy of the sensor 12th is optimal based on the distance _{value A 1} and the angle value α ₁ , is at another point in the vehicle 2 along the railroad 4th the second object 8th in a position in which the measurement accuracy of the sensor 12th is optimal due to the distance _{value A 2} and the angle value α _2.

One ride of the vehicle 2 along the railroad 4th thus leads to changing measurement accuracies and thus errors. The error squares for the two series of value hypotheses R1, R2 therefore change. Consequently, for example, the upper value hypothesis series R2 then converges faster than the lower value hypothesis series R1.

The measurement accuracy of the measurement of the vehicle's own speed over the ground is thus improved by evaluating measurement data obtained in several measurement cycles and evaluated by cluster formation.

Claims (13)

Method for determining a vehicle's own speed (Vx) over the ground, with an object (6, 8, 10) located in the vicinity of the vehicle (2) being recorded in several measuring cycles, with a check being made as to whether the detected object (6, 8, 10) is a stationary object (6, 8) or a non-stationary object (10) and in each measurement cycle a measurement data set (MI, MII, ..., Mn) with a relative speed value (Vri; Vr ₂ ) and an angle value (α ₁ , α ₂ ) of the stationary object (6, 8) is determined in relation to the vehicle (2), the measurement data sets (MI, MII, ..., Mn) determined in each measurement cycle. of the stationary object (6, 8) by a cluster analysis Clusters (C _{1, I} ; C _{2, I} ) are assigned, whereby by evaluating the clusters (C _{1, I} ; C _{2, I} ) one or more value hypotheses (H _{1, I} ; H _{2, I} ) of the airspeed ( Vx) of the vehicle (2) can be determined, the vehicle's own speed (Vx) being determined from the one or more value hypotheses (H _{1, I} ; H _{2, I} ), characterized in that for determining the own speed (Vx) of the vehicle (2) from the one or more value hypotheses (H _{1, I} ; H _{2, I} ) the one or more value hypotheses (H _{1, I} ; H _{2, I} ) are filtered with a Kalman filter. Procedure according to Claim 1 , in which the measured data records (MI, MII, ..., Mn) have a distance value (A ₁ ; A ₂ ) of the stationary object (6, 8) from the vehicle (2). Procedure according to Claim 1 or 2 , in which to determine the vehicle's own speed (Vx) from the one or more value hypotheses (H _{1, I} ; H _{2, I} ) the one or more value hypotheses (H _{1, I} ; H _{2, I} ) with can be tracked using a tracking algorithm. Method according to one of the preceding claims, in which to determine the vehicle's own speed (Vx) from the one or more value hypotheses (H _{1, I} ; H _{2, I} ) from the one or more value hypotheses (H _{1, I} ; H _{2, I} ) one indicator each is determined. Procedure according to Claim 4 , in which to determine the vehicle's own speed (Vx) from the one or more value hypotheses (H _{1, I} ; H _{2, I} ) using the indicator, one of the one or more value hypotheses (H _{1, I} ; H _{2 , I} ) is selected. Method according to one of the preceding claims, in which the measurement data records (MI, MII, ..., Mn) and / or the determined vehicle speed (Vx) of the vehicle (2) via an interface (14) of a sensor (12) of the vehicle ( 2) are transferred. Computing unit which is set up to carry out a method according to one of the preceding claims. Sensor (12) for determining an airspeed (Vx) of a vehicle (2) above ground, with a relative speed and angle determination device (40) and a computing unit (42) Claim 7 . Sensor (12) Claim 8 , with a holder (44) for detachable and reconnectable fastening of the sensor (12) to a vehicle (2). Sensor (12) Claim 8 or 9 , with an energy supply unit (38) for supplying the sensor (12) and / or its components with operating energy. Sensor (12) Claim 8 , 9 or 10 , with an interface (14) for the transmission of data to a vehicle (2) and / or a man-machine interface (18). Computer program that causes a computing unit to implement a method according to one of the Claims 1 to 6th to be carried out when it is executed on the processing unit. Machine-readable storage medium with a computer program stored thereon Claim 12 .

Images