Enhanced object detection and motion estimation for a vehicle environment detection system

1. A vehicle environment detection system that comprises, at least one detection device and at least one processing unit, where the vehicle environment detection system is arranged to detect at least two feature points at objects outside a vehicle, where each of the at least two feature points constitutes a retrievable point that has a position (x 1 , y 1 ; x 2 , y 2 ) on the object, the processing unit is arranged to:

determine the positions (x 1 , y 1 ; x 2 , y 2 ) and resulting velocities (v r1 , v r2 ) for each one of the at least two feature points for multiple frames by use of a feature point tracker; and

determine a reference position (x 0 , y 0 ), a corresponding reference velocity (v r0 ) and a reference angular velocity (ω), constituting a common motion state for the object, by use of results from the feature point tracker;

where the feature point tracker is constituted by a tracking algorithm which is arranged to track multiple of the feature points and which comprises temporal filtering;

wherein the vehicle environment detection system is arranged to calculate the common motion state by solving a linear equation system:

[

v

x

⁢

⁢

1

v

y

⁢

⁢

1

v

x

⁢

⁢

2

v

y

⁢

⁢

2

⋮

v

xN

v

yN

]

=

[

-

y

1

+

y

0

1

0

x

1

-

x

0

0

1

-

y

2

+

y

0

1

0

x

2

-

x

0

0

1

⋮

-

y

N

+

y

0

1

0

x

N

-

x

0

0

1

]

⁡

[

ω

v

x

⁢

⁢

0

v

y

⁢

⁢

0

]

where v x1 . . . v xN are determined velocity x-components for each of the at least two feature points, v y1 . . . v yN are determined velocity y-components for each of the at least two feature points, x 1 . . . x N is an x-coordinate for each of the feature point positions, y 1 . . . y N is a y-coordinate for each feature of the at least two point positions, v x0 is a velocity x-component for a reference velocity vector (v r0 ), v y0 is a velocity y-component for the reference velocity vector (v r0 ), x 0 is a reference position x-coordinate and y 0 is a reference position y-coordinate, where all the coordinates are determined relative a coordinate system having an x-axis and a y-axis.

2. The vehicle environment detection system according to claim 1 , wherein the feature point tracker comprises a Kalman filter.

3. The vehicle environment detection system according to claim 1 , further comprising, the processing unit is arranged to transform the reference velocity vector (v r0 ) to desired positions.

4. The vehicle environment detection system according to claim 3 , further comprising, the processing unit is arranged to transform the reference velocity vector (v r0 ) to the positions of the feature points, and to correct the resulting velocities (v r1 , v r2 ) of the feature points.

5. The vehicle environment detection system according to claim 1 , wherein the vehicle environment detection system comprises a radar sensor arrangement that is arranged to provide radial detected Doppler velocity components (v D1 , v D2 ) and corresponding azimuth detection angles (θ 1 , θ 2 ), where the processing unit is arranged to calculate a velocity vector (v Sr ) of the object at the radar sensor arrangement position (x S , y S ) by use of the detected Doppler velocity components (v D1 , v D2 ) and corresponding azimuth detection angles (θ 1 , θ 2 ), where the processing unit is arranged to apply the velocity vector (v Sr ) to the linear equation system such that the calculation accuracy of the velocity vector (v r0 ) and yaw rate (ω) is improved.

6. A method for determining a common motion state for at least one object outside a vehicle, where the method comprises the steps of:

detecting at least two feature points at objects outside a vehicle, where each of the at least two feature point constitutes a retrievable point that has a fixed position (x 1 , y 1 ; x 2 , y 2 ) on the object;

determining the positions (x 1 , y 1 ; x 2 , y 2 ) and resulting velocities (v r1 , v r2 ) for each one of the at least two feature points for multiple frames by using a feature point tracker; and

determining a reference position (x 0 , y 0 ), a corresponding reference velocity (v r0 ) and reference angular velocity (ω), constituting the common motion state for the object, using the results from the feature point tracker;

where the feature point tracker is constituted by a tracking algorithm which is arranged to track multiple features and which comprises temporal filtering;

calculating the common motion state by solving a linear equation system:

[

v

x

⁢

⁢

1

v

y

⁢

⁢

1

v

x

⁢

⁢

2

v

y

⁢

⁢

2

⋮

v

xN

v

yN

]

=

[

-

y

1

+

y

0

1

0

x

1

-

x

0

0

1

-

y

2

+

y

0

1

0

x

2

-

x

0

0

1

⋮

-

y

N

+

y

0

1

0

x

N

-

x

0

0

1

]

⁡

[

ω

v

x

⁢

⁢

0

v

y

⁢

⁢

0

]

where v x1 . . . v xN are determined velocity x-components for each of the at least two feature points, v y1 . . . v yN are determined velocity y-components for each of the at least two feature points, x 1 . . . x N is an x-coordinate for each of the feature point positions, y 1 . . . y N is a y-coordinate for each feature of the at least two point positions, v x0 is a velocity x-component for a reference velocity vector (v r0 ), v y0 is a velocity y-component for the reference velocity vector (v r0 ), x 0 is a reference position x-coordinate and y 0 is a reference position y-coordinate, where all the coordinates are determined relative a coordinate system having an x-axis and a y-axis.

7. The method according to claim 6 , wherein the method further comprises transforming the reference velocity vector (v r0 ) to desired positions.