Method for assisting a driver in driving a vehicle, a driver assistance system, a computer software program product and vehicle

Patent metadata

Field	Value
Publication number	US-9669830-B2
Application number	US-201514720101-A
Country	US
Kind code	B2
Filing date	May 22, 2015
Priority date	May 30, 2014
Publication date	Jun 6, 2017
Grant date	Jun 6, 2017

Abstract

The invention regards to a method for assisting a driver in driving a vehicle, comprising the steps of producing sensor data by at least one sensor physically sensing the environment of a host vehicle and/or obtaining data conveying information about the environment of a host vehicle, generating a plurality of representation segments each segment being a portion of an entire area of representation of the environment of the host vehicle at a particular point in time wherein a relative position of the portion of such representation segment with respect to a current position of the host vehicle corresponds to a possible position of the host vehicle at that particular point in time, combining the representation segments to a spatio-temporal representation of the environment of the host vehicle; evaluating the spatio-temporal representation and outputting an assistance signal on the basis of an evaluation result.

First claim

The invention claimed is:

1. A method for assisting a driver in driving a host vehicle, comprising the steps of:
performing at least one of producing sensor data by at least one sensor physically sensing the environment of the host vehicle and obtaining data conveying information about the environment of the host vehicle;
generating, based on at least one of the sensor data and the information about the environment of the host vehicle, for each of different points in time, including present and future, a representation segment to provide timewise and spatially adjacent representation segments, each of the representation segments being a portion of a grid of an entire area of representation of the environment of the host vehicle at a particular point in time, wherein a relative position of the portion of at least one of the representation segments with respect to a current position of the host vehicle corresponds to a possible position of the host vehicle at the particular point in time and the grid comprises units forming surface of the environment, wherein information relevant for driving the host vehicle is assigned to at least a plurality of the units and the representation segments are generated by prediction of semantic labels of the plurality of units of the representation segment for the particular point in time, or by cutting out at least one of the representation segments from an entire area representation having the predicted semantic labels of the plurality of units;
connecting the timewise and spatially adjacent representation segments to form a single two-dimensional spatio-temporal representation of the environment of the host vehicle that comprises a plurality of timewise and spatially consecutive segments representing relevant spatial portions of the environment at their respective relevant time;
performing an evaluation process on behavior of at least one of the host vehicle and other traffic participants based on the spatio-temporal representation; and
outputting an assistance signal on the basis of an evaluation result to control an actuator or a warning signal of the host vehicle.

2. The method according to claim 1 , wherein
the representation segments have a predetermined shape and a corresponding discrete point in time is determined based on the host vehicle's velocity.

3. The method according to claim 2 , wherein the discrete point in time corresponding to each representation segment is estimated by one of:
an extrapolation of a current dynamics of the vehicle,
mapping of one or multiple stereotyped behavior options of the host vehicle onto the entire area of the representation,
a probabilistic or deterministic summation of all possible behaviors of the host vehicle, and
an explicit prediction of the most likely executed future behavior of the host vehicle.

4. The method according to claim 1 , wherein
information about the environment of the host vehicle obtained from a plurality of data sources is associated with its corresponding unit of the representation by at least one of:
generating one label for each unit the label corresponding to information only from a prioritized data source,
associating information from each data source to one label respectively and combining these labels to a label vector,
using a multidimensional distribution of a probability of a number of labels, and
creating a new label from a plurality of labels corresponding to individual information.

5. The method according to claim 4 , wherein labels include information about at least one of the following semantic characteristics:
road, host vehicle lane, adjacent lane, opposing lane, incoming road, outgoing road, bicycle lane, walkway, entrance, pothole, offroad, ice, snow, emergency lane, highway entrance, highway exit, occupied road, static occupied, static vehicle, dynamic vehicle, safety area around/next to particular parts of traffic participants/structures, stopping area, one-way street, parking space, intersection area, historically dangerous area, puddle, zebra crossing, active zebra crossing, traversable obstacle, road debris, animal crossing area, free area, unknown area, pedestrian zone, bus stop, bus lane, traffic island, cobblestone, slippery road, slanted road, speed bump, wet road, tram rails, gravel.

6. The method according to claim 1 , wherein
in the step of performing the evaluation process a plurality of possible future movement behaviors of a traffic participant are compared and its future movement behavior is predicted.

7. The method according to claim 6 , wherein the prediction is conducted for at least one of the host vehicle and other traffic participants.

8. The method according to claim 6 , wherein the predicted future movement behavior of other traffic participants is used for the generation of the representation segments.

9. The method according to claim 6 , wherein the possible future movement behaviors are evaluated individually for each of the representation segments or are evaluated as representation segment independent trajectories.

10. The method according to claim 1 , wherein in the performing of the evaluation process the size of the host vehicle or the size of the host vehicle plus an additional safety area is taken into consideration.

11. A computer software program product embodied on a non-transitory computer-readable medium, said product performing a process, when executed on a computer, the process comprising:
performing at least one of producing sensor data by at least one sensor physically sensing the environment of the host vehicle and obtaining data conveying information about the environment of the host vehicle,
generating, based on at least one of the sensor data and the information about the environment of the host vehicle, for each of different points in time, including present and future, a representation segment to provide timewise and spatially adjacent representation segments each of the representation segments being a portion of a grid of an entire area of representation of the environment of the host vehicle at a particular point in time, wherein a relative position of the portion of at least one of the representation segments with respect to a current position of the host vehicle corresponds to a possible position of the host vehicle at the particular point in time and the grid comprises units forming surface of the environment, wherein information relevant for driving the host vehicle is assigned to at least a plurality of the units and the representation segments are generated by prediction of semantic labels of the plurality of units of the representation segments for the particular point in time, or by cutting out at least one of the representation segments from an entire area representation having the predicted semantic labels of the plurality of units,
connecting the timewise and spatially adjacent representation segments to form a single two-dimensional spatio-temporal representation of the environment of the host vehicle that comprises a plurality of timewise and spatially consecutive segments representing relevant spatial portions of the environment at their respective relevant time;
performing an evaluation process on behavior of at least one of the host vehicle and other traffic participants based on the spatio-temporal representation; and
outputting an assistance signal on the basis of an evaluation result to control an actuator or a warning signal of the host vehicle.

Assignees

• Honda Res Inst Europe Gmbh

Inventors

• Weisswange Thomas H

Classifications

• G01S7/24 Primary

  the display being orientated or displaced in accordance with movement of object carrying the transmitting and receiving apparatus, e.g. true-motion radar · CPC title

• G01S13/931

  of land vehicles · CPC title

• G08G1/167

  Driving aids for lane monitoring, lane changing, e.g. blind spot detection · CPC title

• G08G1/165

  for passive traffic, e.g. including static obstacles, trees · CPC title

• G08G1/166

  for active traffic, e.g. moving vehicles, pedestrians, bikes · CPC title