DE10307848A1 - Determination of the movement situation of a traffic object within the vicinity of a vehicle by measuring relative movement values and inputting the values to a computer mathematical model - Google Patents

Abstract

Method whereby relative values between the vehicle and the traffic object are measured to enable condition values of the traffic object to be determined in a computer unit (4). The computer unit implements a mathematical model using the vehicle acceleration (aFz), the distance to the object (dobj) and at least one speed value as inputs in order to determine values describing the longitudinal dynamics of the traffic object. The invention also relates to a corresponding arrangement for implementation of the inventive method.

Description

The invention relates to a Method and device for determining the state of motion a traffic object in the vicinity of a vehicle after the The preamble of claim 1 or 13.

From the publication DE 197 43 726 A1 It is known, with the aid of a position data acquisition device, which is carried in a moving vehicle, as well as an image recording device to determine the relative distance and the relative speed to a vehicle in front and accelerate or decelerate when exceeding or falling below predetermined limits the vehicle or a warning signal to create. For a high degree of safety when using this distance control system, especially in the case that is driven with a small distance, in addition to the relative distance and the relative speed and the acceleration of the vehicle in front must be known to ensure that in dynamic driving maneuvers of preceding vehicle (for example, sudden, strong braking) all safety-related criteria in the following vehicle can be met.

The value of the acceleration can be used in systems such as DE 197 43 726 A1 described in principle also be determined mathematically by means of a numerical differentiation of the measured relative velocity. However, the numerical differentiation increases the noise and the signal noise, with the result that the obtained acceleration value must be low-pass filtered in order to eliminate or reduce the resulting from the differentiation disturbances. This method is described for example in WO 95/14939 A1, from which a Kalman filtering of the relative distance, the relative speed and the relative acceleration is known.

However, the problem is that different sensor systems for the Measurement of the speeds or accelerations used Need to become. Here you can temporal phase shifts between the different sensor systems occur, which are reinforced by the numerical differentiation. The phase shifts represent an additional source of error that the values for the speed and the acceleration worsen and the Impair driving safety can.

Starting from this state of the art the invention is based on the problem, the state of motion of a traffic object in the vicinity of a vehicle with high accuracy.

This problem is inventively with the Characteristics of claim 1 or 13 solved. The subclaims give appropriate training on.

In the method according to the invention for determining the state of motion of a traffic object in The surroundings of a vehicle are measured as the acceleration of the vehicle, the relative distance between the vehicle and the traffic object and at least a speed variable - especially the intrinsic speed of the vehicle and / or the relative speed between vehicle and traffic object - measured or determined. With knowledge of these sizes can subsequently in a computing unit based on a mathematical model, in particular by means of an observer model, as a state variable at least one the longitudinal dynamics the traffic object descriptive size such. its driving speed and / or its acceleration are calculated as absolute values. With With the help of a suitable observer model, the measured data will be displayed fused the different measuring equipment and it can do that different temporal measuring behavior of the measuring devices over the be corrected dynamic observer model.

The descriptive of the longitudinal dynamics of the traffic object Size - in particular Driving speed or acceleration of the traffic object - be in contrast to known from the prior art method no longer algebraic by adding the own velocity or Acceleration of the vehicle with an associated relative value calculated between vehicle and traffic object, but in the observer model issued as absolute sizes. This results in the advantage that no phase shifts from different measuring arrangements into the calculated state variables of Incorporate traffic object, which is more precise results for the state variables of Traffic object can be achieved. In particular by numerical Differentiation generated disturbances like numerical noise can be avoided.

Basically, it is enough, the airspeed as well as the vehicle's own acceleration and the relative distance to the vehicle Traffic object to know, with the help of the observer model the Driving speed and the acceleration of the traffic object to determine. Such a system is observable.

Instead of the vehicle's own speed, the relative speed between the vehicle and the moving traffic object can also be measured or determined and the mathematical Be model of care. Furthermore, both the airspeed and the relative velocity can be included in the observer model. In both alternative and cumulative formulations, the system is observable.

In the mathematical model of a Observers enter an observer or gain matrix, which determines the quality of the approximation of the observer state to the system state determines determined. This gain matrix can according to a first appropriate execution by Eigenwertvorgabe the observer model are set, where through an appropriate choice of eigenvalues the observer model can be kept stable. Alternatively, the gain matrix of the observer model also determined using a Kalman filter become; therefor can considering of prescribing matrices for Measurement noise as well as for System noise recursive relationships known in the art for determining the gain matrix be used.

The knowledge of the complete system state of the traffic object, which is in particular a preceding vehicle or a subsequent vehicle may be in driver assistance systems be used, such that the driver on the current driving situation is informed and / or the driving condition of their own vehicle consideration the driving condition of the traffic object influenced or according to predetermined optimization criteria can be adjusted. In such driver assistance systems act For example, and emergency braking systems or proximity control systems, in which vehicles drive a short distance behind each other.

The inventive device for carrying out the Method includes a measuring device on which at least one's own Acceleration of the vehicle, the relative distance between the vehicle and traffic object as well as at least one speed variable can be. Furthermore, the device is a computing unit in which a mathematical observer model is stored is in which considering the measured sizes the Driving speed and the acceleration of the traffic object to be calculated.

Further advantages and expedient designs are the further claims, the figure description and the drawings. Show it:

1 3 is a schematic block diagram for determining the state of motion of a traffic object which moves relative to a vehicle.

2 a block diagram of a device in a modified embodiment,

3 a block diagram of yet another device in a modified embodiment.

In the 1 illustrated device 10 is arranged in a vehicle and is used to determine the state of motion of a traffic object in the vicinity of a vehicle.

With the help of the device 10 For example, it is possible to determine the state of motion of the traffic object, which can in particular move relative to the vehicle and which is, for example, a preceding vehicle or a following vehicle. With knowledge of the state of motion of the traffic object driver assistance systems can be used and adjusted according to predetermined control or optimization criteria or it can be displayed to the driver, the current traffic situation.

The device 10 includes a plurality of measuring devices 20 , which is a vehicle speed measuring device 1 for determining the vehicle's own vehicle speed v _FZ to a driving acceleration measuring device 2 for determining the acceleration a _{Fz of} the vehicle as well as a distance detection device 3 for determining the relative _distance d _obj between the vehicle and the traffic _object . The of measuring equipment 1 . 2 and 3 supplied measuring signals 301 (Acceleration a _Fz ), 302 (Airspeed v _FZ ) and 303 (Relative _distance d _obj ) become a computing unit 4 fed, which is also part of the facility 10 and in which the supplied signals are processed in accordance with a stored observer model for determining the system states driving speed v _obj and acceleration a _{obj of} the traffic _object . The acceleration a _obj and the velocity v _obj are signals 305 respectively. 306 before and can be further processed in the vehicle to control various aggregates of the vehicle.

beschrieben werden, wobei mit x ^ der Schätzvektor für den Zustandsvektor x, mit u ein Eingangsvektor, mit y ein Ausgangsvektor, mit A die Systemmatrix des mathematischen Beobachtermodells, mit B die Eingangsmatrix, mit C die Ausgangsmatrix und mit K eine Verstärkungsmatrix bezeichnet wird.That in the arithmetic unit 4 filed, the method or the device underlying mathematical observer model (Luenberger or identity observer), with the help of the sought state variables speed and acceleration of the traffic object to be determined, can with the relationships

where x is the estimate vector for the state vector x, u is an input vector, y is an output vector, A is the system matrix of the mathematical observer model, B is the input matrix, C is the output matrix, and K is a gain matrix.

wobei a_obj die Beschleunigung und v_obj die Geschwindigkeit des Verkehrsobjekts, v_Fz Eigengeschwindigkeit des Fahrzeugs und d_obj den Relativabstand zwischen Fahrzeug und Verkehrsobjekt bezeichnet. Bei einem in dieser Weise aufgebauten Zustandsvektor nimmt die Systemmatrix A die Dimension 4x4 ein und besitzt folgenden Aufbau:

The state vector x to be estimated is expediently constructed as a 4x1 vector and comprises the elements

where a _{obj denotes} the acceleration and v _obj the speed of the traffic _object , v _Fz own vehicle speed and d _obj the relative _distance between the vehicle and the traffic _object . In a state vector constructed in this way, the system matrix A occupies the dimension 4x4 and has the following structure:

The input matrix B has the dimension 4x1:

The input vector u has the dimension 1x1 and is identical to the measured - or derived from measured variables - acceleration a _{Fz of} the vehicle: u = a Fz ,

The output or measuring vector y comprises the relative _distance d _obj and at least one of the speed _variables v _Fz (vehicle's own speed) and v _{rel, obj} (relative speed between vehicle and traffic _object ). It is thereby possible to calculate the estimated state vector k on the basis of measured values for the acceleration a _{Fz of} the vehicle, for the relative _distance d _obj and for at least one speed variable.

In the facility 1 assigned embodiment, the 2x1 output vector y has the two variables relative _distance d _obj and _airspeed v _Fz :

sowie folgender Aufbau für die Verstärkungsmatrix K mit der Dimension 4x2:

This results in the following structure for the output matrix C with the dimension 2x4:

and the following structure for the gain matrix K with the dimension 4x2:

In the device according to 2 be the measured variables, the vehicle acceleration a _Fz in the driving acceleration measuring device 2 and the relative distance d _obj in the distance recognizer 3 generated. Furthermore, the distance recognizer measures 3 also the relative velocity v _{rel, obj} between the vehicle and the traffic _object , for example using radar systems and utilizing the Doppler effect. The relative velocity v _{rel, obj} is used as a measurement signal 304 the arithmetic unit 4 fed for further processing. On a driving speed measuring device can in the embodiment according to 2 be waived. In the arithmetic unit 4 In turn, according to the stored observer model, the speed d _obj and the acceleration a _{obj of} the traffic _{object are} calculated.

In the facility 2 assigned embodiment, the 2x1 output vector y has the two variables relative _distance d _obj and relative velocity v _{rel, obj}

The starting matrix C with dimension 2x4 has the following structure:

The gain matrix K with the dimension 4x2 has the same construction as in the first embodiment:

The furnishings after 3 represents a combination of the two above-described facilities from the 1 and 2 dar. The measuring device 20 includes the vehicle speed measuring device 1 for measuring the airspeed v _FZ the driving acceleration measuring device 2 for measuring the acceleration a _{Fz of} the vehicle and the distance detection device 3 for measuring both the relative _distance d _obj and the relative velocity v _{rel, obj} between the vehicle and the traffic _object . The measuring signals 301 to 304 become the arithmetic unit 4 fed, in turn calculated from the stored observer model speed v _obj and acceleration a _{obj of} the traffic _object as absolute values.

In the facility 3 The 3x1 output vector y has three elements: the relative _distance d _obj , the own velocity v _Fz and the relative velocity v _{rel, obj}

Accordingly, the output matrix C now has the dimension 3x4 with the following structure:

The gain matrix K now has the dimension 4x3 with the following structure:

In all three versions according to the 1 . 2 and 3 One obtains an observable system, where the elements of the gain matrix K can be determined either by eigenvalue prediction or by means of a Kalman filter.

In the eigenvalue prediction method, the eigenvalues become β ₁ , β ₂ , β ₃ , β _{4 of} the matrix F = A - KC given and calculated from the resulting equations of determination, the elements of the gain matrix K. In the case of a Kalman filter, the gain matrix K becomes the recursive relations K k = P k C T (CP k C T + R) -1 P k = (E - K k C) P k P k = AP k 1 A T + Q where R is a given covariance matrix for measurement noise, Q is a given covariance matrix for system noise, and E is the unit matrix.

Optionally, in the above-described embodiments, the driving acceleration measuring device 2 be dispensed with for measuring the acceleration a _{Fz of} the vehicle. In this case, the driving acceleration a _Fz can be determined by numerical differentiation of the airspeed v _FZ .

Claims (13)

Method for determining the state of motion of a traffic object in the surroundings of a vehicle, in which state variables of the vehicle or difference values between the vehicle and the traffic object are measured as measured variables and stored in a computing unit ( 4 ) State variables of the traffic object are determined, characterized in that the acceleration (a _Fz ) of the vehicle, the relative _distance (d _obj ) between the vehicle and traffic _object and at least one speed _variable (a _Fz , v _{rel, obj} ) is measured or determined and from it in the arithmetic unit ( 4 ) is calculated in a mathematical model as a state variable at least one of the longitudinal dynamics of the traffic object descriptive size. A method according to claim 1, characterized in that as the longitudinal dynamics of the traffic object descriptive size of the _vehicle speed (v _obj ) and / or the acceleration (a _obj ) of the traffic _object are used. Method according to claim 1 or 2, characterized in that at least one of the following state variables is used in the mathematical model: - speed variable (v _Fz , v _{rel, obj} ) and / or - relative _distance (d _obj ) between vehicle and traffic _object and / or Driving speed (v _obj ) of the traffic _object and / or - acceleration (a _obj ) of the traffic _object . Method according to one of claims 1 to 3, characterized in that as a speed variable, the intrinsic speed (v _Fz ) of the vehicle is measured. Method according to one of claims 1 to 4, characterized in that is measured as the speed _variable, the relative speed (v _{rel, obj} ) between the vehicle and the traffic _object . Method according to one of claims 1 to 5, characterized in that as mathematical Model a state observer, in particular Kalman filter is used. Method according to one of claims 1 to 6, characterized in that in the mathematical observer model, the state variables of the traffic object according to the context

where x ^ an estimated state vector u an input vector y an output vector A a system matrix of the mathematical observer model, B an input matrix of the mathematical observer model, C an output matrix of the mathematical observer model K a gain matrix, the state vector (x) to be estimated as According to 4x1 vector

with a _obj acceleration of the traffic _object v _obj velocity of the traffic _object v _Fz own velocity of the vehicle d _obj relative _distance between vehicle and traffic _object is constructed, the system matrix (A) as a 4x4 matrix

is constructed, the input matrix (B) has the dimension 4x1 and according to

is constructed, the input vector (u) has the dimension 1x1 and with the measured or calculated acceleration (a _Fz ) of the vehicle according to u = a Fz is identical and the output vector (y) _comprises the relative _distance (d _obj ) and at least one velocity _variable (v _Fz , v _{rel, obj} ). Method according to Claim 7, characterized in that the output vector (y) _comprises the relative _distance (d _obj ) and, as speed _variable, the vehicle's own speed (v _Fz ) and the shape

and that the output matrix is constructed as a 2x4 matrix.

A method according to claim 7, characterized in that the output vector (y) _comprises the relative _distance (d _obj ) and as a speed _variable, the relative speed (v _{rel, obj} ) between the vehicle and the traffic _object and the shape

and that the starting matrix (C) is constructed as a 2x4 matrix:

A method according to claim 7, characterized in that the output vector (y) _comprises the relative _distance (d _obj ) and as speed _{variables the airspeed} (v _Fz ) of the vehicle and the relative speed (v _{rel, obj} ) between the vehicle and traffic _object and the shape

and that the starting matrix (C) is constructed as a 3 × 4 matrix:

Method according to one of claims 7 to 10, characterized in that the gain matrix (K) by specifying the eigenvalues (β ₁ , β ₂ , β ₃ , β ₄ ) of the matrix F = A - KC is determined. Method according to one of claims 7 to 10, characterized in that the gain matrix (K) using a Kalman filter from the recursive relationships K k = P k C T (CP k C T + R) -1 P k = (E - K k C) P k P k = AP k 1 A T + Q is determined, where R is a predetermined covariance matrix for measurement noise Q a predetermined covariance matrix for system noise E is the unit matrix. Device for carrying out the method according to one of Claims 1 to 12, having a measuring device ( 20 ) for measuring the vehicle 's own acceleration (a _Fz ), the relative _distance (v _obj ) between the vehicle and the traffic _object and at least one speed _variable (v _Fz , v _{rel, obj} ) for determining the state of motion of a traffic _object in the vicinity of the vehicle, with a Arithmetic unit ( 4 ), in which state variables of the traffic object can be determined, characterized in that in the arithmetic unit ( 4 ) can be calculated using a mathematical model as a state variable at least one of the longitudinal dynamics of the traffic object descriptive size.