# US 2017/0254895 A1 · Detecting long objects by sensor fusion

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2017/0254895 A1
- Document ID: US20170254895A1
- Status: Published application
- Kind: A1
- Application number: 15/057,437
- Filed: 2016-03-01
- Published: 2017-09-07
- Earliest claimed: 2016-03-01
- Assignees: GM Global Technology Operations LLC
- Applicants: GM Global Technology Operations LLC
- Inventors: WEI TONG; SHUQING ZENG; JINSONG WANG; WENDE ZHANG
- Web page: https://patentgrep.com/patent/US20170254895A1

## CPC

- G01S 13/867: {Combination of radar systems with cameras}
- G01S 7/41: using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S 13/878: {Combination of several spaced transmitters or receivers of known location for determining the position of a transponder or a reflector (G01S13/874 takes precedence)}
- G01S 13/931: of land vehicles
- G01S 2013/9322: {using additional data, e.g. driver condition, road state or weather data}
- G01S 2013/93271: {in the front of the vehicles}
- G01S 2013/93272: {in the back of the vehicles}
- G01S 2013/93274: {on the side of the vehicles}
- G01S: RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES

## Abstract

A system and method are provided for detecting and identifying elongated objects relative to a host vehicle. The method includes detecting objects relative to the host vehicle using a plurality of object detection devices, identifying patterns in detection data that correspond to an elongated object, wherein the detection data includes data fused from at least two of the plurality of object detection devices, determining initial object parameter estimates for the elongated object using each of the plurality of object detection devices, calculating object parameter estimates for the elongated object by fusing the initial object parameter estimates from each of the plurality of object detection devices, and determining an object type classification for the elongated object by fusing the initial object parameter estimates from each of the plurality of object detection devices.

## Claims

### Claim 1

A method for detecting and identifying elongated objects relative to a host vehicle, the method comprising the steps of:
detecting objects relative to the host vehicle using a plurality of object detection devices;
identifying patterns in detection data that correspond to an elongated object, wherein the detection data includes data fused from at least two of the plurality of object detection devices;
determining initial object parameter estimates for the elongated object using each of the plurality of object detection devices;
calculating object parameter estimates for the elongated object by fusing the initial object parameter estimates from each of the plurality of object detection devices; and
determining an object type classification for the elongated object by fusing the initial object parameter estimates from each of the plurality of object detection devices.

### Claim 2 (depends on 2)

The method of claim 2, wherein the plurality of object detection devices includes at least two object detection devices including a radar detection device and an imaging system.

### Claim 3 (depends on 2)

The method of claim 2, wherein the step of identifying patterns includes identifying radar signal patterns in radar detection data generated by radar sensors.

### Claim 4 (depends on 3)

The method of claim 3, wherein identifying distinct radar signal patterns includes clustering an initial set of radar measurements into at least one group according to one or more metric(s).

### Claim 5 (depends on 4)

The method of claim 4, wherein the one or metrics includes a location and signal strength for each radar measurement in the radar detection data.

### Claim 6 (depends on 5)

The method of claim 5, wherein the one or more metric(s) further include a metric derived from captured image data from the imaging system.

### Claim 7 (depends on 6)

The method of claim 6, wherein the clustering step further includes mapping radar detection data to the captured image data to determine corresponding image pixel data.

### Claim 8 (depends on 4)

The method of claim 4, wherein identifying patterns in the detection data further includes determining, for each of the at least one group, a distribution of the radar measurements in a first dimension, and a concentration of radar measurement locations along a second dimension, which is perpendicular to the first dimension.

### Claim 9 (depends on 4)

The method of claim 4, wherein identifying patterns in the detection data further includes applying a linear regression model to the radar measurements in each of the at least one group that satisfy the one or more metric(s).

### Claim 10 (depends on 4)

The method of claim 4, wherein identifying patterns in the detection data further includes sorting each radar measurement in each of the at least one group according to an azimuth angle measured relative to the host vehicle.

### Claim 11 (depends on 10)

The method of claim 10, further including determining if a range rate associated with each of the sorted radar measurements changes gradually with respect to one another according to a range rate threshold.

### Claim 12 (depends on 11)

The method of claim 11, further including estimating a relative speed for a group of radar measurements that have a gradual change in range rate according to the range rate threshold, wherein estimating the relative speed includes applying a linear regression speed model to the group of radar measurements.

### Claim 13 (depends on 2)

The method of claim 2, wherein detection data generated by the imaging system includes one or more frame(s) of image data relating to captured images, and wherein the step of identifying patterns includes identifying regions of interest in each frame and tracking points of interest in each of the regions of interest from a previous frame to a current frame.

### Claim 14 (depends on 1)

The method of claim 1, wherein calculating the object parameter estimates further includes determining a parameter weight for each of the initial object parameter estimates computed by each of the plurality of object detection devices, and summing weighted object parameters for each of the plurality of object detection devices.

### Claim 15 (depends on 1)

The method of claim 1, wherein determining the object type classification for the elongated object includes determining a maximum probability of a particular class of objects as a function of the initial object parameter estimates from each of the plurality of object detection devices.

### Claim 16

A method for detecting and identifying elongated objects relative to a host vehicle, the method comprising the steps of:
obtaining detection data relating to objects detected using a plurality of object detection devices;
fusing the detection data at a first processing level by referencing complementary detection data from each of the plurality of object detection devices;
classifying a detected object as an elongated object at each of the plurality of object detection devices by identifying patterns in the fused detection data that correspond to an elongated object;
calculating initial object parameter estimates for the elongated object using each of the plurality of object detection devices;
fusing the detection data at a second processing level by:
weighting the initial object parameter estimates for each of the plurality of object detection devices and summing the weighted initial object parameter estimates to determine object parameter estimates for the elongated object, and
calculating a maximum probability of a particular class of objects as a function of the initial object parameter estimates from each of the plurality of object detection devices to determine an object type classification for the elongated object.

### Claim 17 (depends on 16)

The method of claim 16, wherein obtaining detection data includes radar measurements from a radar detection device and captured image data from an imaging system, and wherein classifying the detected object as an elongated object includes using the radar detection device to apply position and velocity linear regression models to groups of radar measurements clustered according to a location, a signal strength, or both, of the radar measurements and corresponding image pixel data from the captured image data, and using the imaging system to identify regions of interest in each frame of image data and tracking points of interest in each of the regions of interest from a previous frame to a current frame.

### Claim 18

A system for detecting and identifying elongated objects relative to a host vehicle, the system comprising:
at least two object detection devices configured to:
obtain detection data relating to objects detected in a field-of-view relating to each of the at least two object detection devices;
fuse the detection data by referencing complementary detection data from each of the at least two object detection devices;
classify a detected object as an elongated object by identifying patterns in the fused detection data that correspond to an elongated object; and
calculate initial object parameter estimates for the elongated object using each of the at least two object detection devices;
at least one vehicle control module configured to:
receive the initial object parameter estimates from each of the at least two object detection devices and fuse the detection data to determine object parameter estimates and to classify the type of elongated object.

### Claim 19 (depends on 18)

The system claim 18, wherein the at least two object detection devices include a radar detection device and an imaging system, and wherein the at least one vehicle control module is a fusion module.

### Claim 20 (depends on 18)

The system of claim 18, wherein the at least one vehicle control module is configured to fuse the detection data by:
weighting the initial object parameter estimates for each of the plurality of object detection devices and summing the weighted initial object parameter estimates to determine object parameter estimates for the elongated object, and
calculating a maximum probability of a particular class of objects as a function of the initial object parameter estimates from each of the plurality of object detection devices to determine an object type classification for the elongated object.
