U.S. patent application number 15/730572 was filed with the patent office on 2018-11-01 for calibration of laser sensors.
The applicant listed for this patent is SZ DJI TECHNOLOGY CO., LTD.. Invention is credited to Lu Ma, Kanzhi Wu.
Application Number | 20180313942 15/730572 |
Document ID | / |
Family ID | 63916534 |
Filed Date | 2018-11-01 |
United States Patent
Application |
20180313942 |
Kind Code |
A1 |
Wu; Kanzhi ; et al. |
November 1, 2018 |
CALIBRATION OF LASER SENSORS
Abstract
Automatic calibration of laser sensors carried by a mobile
platform, and associated systems and methods are disclosed herein.
A representative method includes determining an overlapping region
of point cloud data generated by laser sensors, comparing surface
features of the point clouds within the overlapping region, and
generating calibration rules based thereon.
Inventors: |
Wu; Kanzhi; (Shenzhen City,
CN) ; Ma; Lu; (Shenzhen City, CN) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
SZ DJI TECHNOLOGY CO., LTD. |
Shenzhen City |
|
CN |
|
|
Family ID: |
63916534 |
Appl. No.: |
15/730572 |
Filed: |
October 11, 2017 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
PCT/CN17/82584 |
Apr 28, 2017 |
|
|
|
15730572 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G01S 7/4972 20130101;
G01S 17/89 20130101; G01S 13/867 20130101; G01S 17/42 20130101;
G01S 7/4808 20130101; G01B 11/24 20130101; G01B 21/042 20130101;
G01S 13/931 20130101; G01B 11/26 20130101 |
International
Class: |
G01S 7/497 20060101
G01S007/497; G01S 17/89 20060101 G01S017/89; G01B 11/26 20060101
G01B011/26; G01S 13/86 20060101 G01S013/86; G01S 13/93 20060101
G01S013/93; G01S 17/42 20060101 G01S017/42 |
Claims
1. A computer-implemented method for calibrating a first laser unit
and a second laser unit both affixed to a common autonomous car,
comprising: transforming point cloud information obtained from the
first laser unit into a first point cloud in a coordinate system
associated with the autonomous car based at least in part on a
first transformation matrix, the first transformation matrix
defined at least in part by a position and orientation of the first
laser unit relative to the autonomous car; transforming point cloud
information obtained from the second laser unit into a second point
cloud in the coordinate system associated with the autonomous car
based at least in part on a second transformation matrix, the
second transformation matrix defined at least in part by a position
and orientation of the second laser unit relative to the autonomous
car; determining an overlapping region between the first and second
point clouds based at least in part on paired points between the
first and second point clouds, wherein the paired points include at
least a nearest neighbor in the second point cloud for at least one
point in the first point cloud; deriving a plurality of first
features that at least partially characterize a surface of the
first point cloud in the overlapping region; deriving a plurality
of second features that at least partially characterize a surface
of the second point cloud in the overlapping region; generating at
least one calibration rule for calibration between the first and
second laser units based at least in part on evaluating a function
that includes at least the first and second features, wherein the
at least one calibration rule includes at least one of a
translational transformation or a rotational transformation; and
causing calibration between the first and second laser units based
on the at least one calibration rule.
2. A computer-implemented method for calibrating at least a first
emitter/detector unit and a second emitter/detector unit both
carried by a common mobile platform, comprising: transforming point
cloud information obtained from the first emitter/detector unit
into a first point cloud in a reference system associated with the
mobile platform; transforming point cloud information obtained from
the second emitter/detector unit into a second point cloud in the
reference system associated with the mobile platform; determining
an overlapping region between the first and second point clouds
based at least in part on identifying at least a nearest neighbor
in the second point cloud for at least one point in the first point
cloud; comparing surface attributes of the first and second point
clouds in the overlapping region; and generating at least one
calibration rule for calibration between the first and second
emitter/detector units based at least in part on comparing the
surface attributes, wherein the at least one calibration rule
includes at least one of a translational transformation or a
rotational transformation.
3. The method of claim 2, wherein transforming point cloud
information obtained from the first emitter/detector unit is based
at least in part on a first set of transformation rules.
4. The method of claim 3, wherein the first set of transformation
rules is at least partially defined in accordance with a position
and orientation of the first emitter/detector unit relative to the
mobile platform.
5. The method of claim 3, wherein transforming point cloud
information obtained from the second emitter/detector unit is based
at least in part on a second set of transformation rules and
wherein the second set of transformation rules differs from the
first set of transformation rules.
6. The method of claim 3, wherein the reference system associated
with the mobile platform comprises a coordinate system.
7. The method of claim 3, wherein the first set of transformation
rules comprises a transform matrix.
8-11. (canceled)
12. The method of claim 2, wherein determining an overlapping
region comprises creating a tree-structured data structure for at
least one of the first or second point clouds.
13. The method of claim 12, wherein the tree-structured data
structure comprises a K-Dimensional (KD) tree data structure.
14-23. (canceled)
24. The method of claim 2, further comprising causing calibration
of the first and second emitter/detector units in accordance with
the at least one calibration rule.
25. A non-transitory computer-readable medium storing
computer-executable instructions that, when executed, cause one or
more processors associated with a mobile platform to perform
actions comprising: transforming point cloud information obtained
from a first emitter/detector unit into a first point cloud in a
reference system associated with the mobile platform; transforming
point cloud information obtained from a second emitter/detector
unit into a second point cloud in the reference system associated
with the mobile platform; determining an overlapping region between
the first and second point clouds based at least in part on
identifying at least a nearest neighbor in the second point cloud
for at least one point in the first point cloud; comparing surface
attributes of the first and second point clouds in the overlapping
region; and generating at least one calibration rule for
calibration between the first and second emitter/detector units
based at least in part on comparing the surface attributes, wherein
the at least one calibration rule includes at least one of a
translational transformation or a rotational transformation.
26-30. (canceled)
31. The computer-readable medium of claim 25, wherein the first
emitter/detector unit includes at least one laser sensor.
32. The computer-readable medium of claim 31, wherein the at least
one laser sensor has a field of view (FOV) that is smaller than at
least one of 360 degrees, 180 degrees, 90 degrees, or 60
degrees.
33. The computer-readable medium of claim 31, wherein the first
emitter/detector unit includes a plurality of laser sensors rigidly
fixed relative to each other.
34-36. (canceled)
37. The computer-readable medium of claim 25, wherein comparing
surface attributes comprises matching a surface associated with the
first point cloud with a surface associated with the second
point.
38. The computer-readable medium of claim 37, wherein matching a
surface associated with the first point cloud with a surface
associated with the second point comprises determining normal
vector information with respect to at least a portion of the first
point cloud.
39. The computer-readable medium of claim 25, wherein comparing
surface attributes further comprises evaluating a target function
defined at least in part by a plurality of points of the first and
second point clouds that are within the overlapping region.
40. The computer-readable medium of claim 39, wherein the target
function comprises a rotational component and a translational
component.
41. The computer-readable medium of claim 40, wherein comparing
surface attributes further comprises keeping the translational
component fixed.
42. The computer-readable medium of claim 39, wherein generating at
least one calibration rule comprises optimizing the target
function.
43. The computer-readable medium of claim 42, wherein optimizing
the target function is based at least in part on a least squares
method.
44-47. (canceled)
48. A vehicle including a programmed controller that at least
partially controls one or more motions of the vehicle, wherein the
programmed controller includes one or more processors configured
to: transform point cloud information obtained from a first
emitter/detector unit into a first point cloud in a reference
system associated with the vehicle; transform point cloud
information obtained from a second emitter/detector unit into a
second point cloud in the reference system associated with the
vehicle; determine an overlapping region between the first and
second point clouds based at least in part on identifying at least
a nearest neighbor in the second point cloud for at least one point
in the first point cloud; compare surface attributes of the first
and second point clouds in the overlapping region; and generate at
least one calibration rule for calibration between the first and
second emitter/detector units based at least in part on the
comparison of surface attributes, wherein the at least one
calibration rule includes at least one of a translational
transformation or a rotational transformation.
49-57. (canceled)
58. The vehicle of claim 48, wherein determining an overlapping
region comprises creating a tree-structured data structure for at
least one of the first or second point clouds.
59. (canceled)
60. The vehicle of claim 48, wherein comparing surface attributes
comprises matching a surface associated with the first point cloud
with a surface associated with the second point.
61. (canceled)
62. The vehicle of claim 48, wherein comparing surface attributes
further comprises evaluating a target function defined at least in
part by a plurality of points of the first and second point clouds
that are within the overlapping region.
63-66. (canceled)
67. The vehicle of claim 48, wherein the at least one calibration
rule comprises a rule for transformation between coordinate systems
of the first emitter/detector unit and the second emitter/detector
unit.
68. The vehicle of claim 48, wherein the one or more processors are
further configured to detect a difference between the generated at
least one calibration rule against one or more previously generated
calibration rules.
69. The vehicle of claim 48, wherein the vehicle includes at least
one of an unmanned aerial vehicle (UAV), a manned aircraft, an
autonomous car, a self-balancing vehicle, or a robot.
70-92. (canceled)
Description
TECHNICAL FIELD
[0001] The presently disclosed technology is generally directed to
calibration of emitter/detector sensors, such as laser sensors,
that are carried by mobile platforms.
BACKGROUND
[0002] Laser sensors, such as LiDAR sensors, typically transmit a
pulsed laser signal outwards, detect the pulsed signal reflections,
and measure three-dimensional information (e.g., laser scanning
points) in the environment to facilitate environment mapping. To
accomplish precise mapping of an environment that surrounds a
mobile platform, an omni-directional laser sensor with 360-degree
horizontal field of view (FOV) is typically mounted on the mobile
platform to constantly scan its surroundings. Omni-directional
laser sensors are typically expensive, non-customizable, and have a
limited vertical FOV. Accordingly, there remains a need for
improved sensing techniques and devices for mobile platforms.
SUMMARY
[0003] The following summary is provided for the convenience of the
reader and identifies several representative embodiments of the
presently disclosed technology.
[0004] In some embodiments, a computer-implemented method for
automatically calibrating at least a first emitter/detector unit
and a second emitter/detector unit both carried by a common mobile
platform includes transforming point cloud information obtained
from the first emitter/detector unit into a first point cloud in a
reference system associated with the mobile platform; transforming
point cloud information obtained from the second emitter/detector
unit into a second point cloud in the reference system associated
with the mobile platform; determining an overlapping region between
the first and second point clouds; comparing surface attributes of
the first and second point clouds in the overlapping region; and
generating at least one calibration rule for calibration between
the first and second emitter/detector units based at least in part
on comparing the surface attributes. In some embodiments,
transforming point cloud information obtained from the first
emitter/detector unit is based at least in part on a first set of
transformation rules, which is at least partially defined in
accordance with a position and orientation of the first
emitter/detector unit relative to the mobile platform. In some
embodiments, the reference system associated with the mobile
platform comprises a coordinate system. In some embodiments, the
first set of transformation rules comprises a transform matrix. In
some embodiments, the first emitter/detector unit includes at least
one laser sensor that has a field of view (FOV) smaller than at
least one of 360 degrees, 180 degrees, 90 degrees, or 60 degrees.
In some embodiments, the first emitter/detector unit includes a
plurality of laser sensors rigidly fixed relative to each other. In
some embodiments, determining an overlapping region comprises
determining at least a nearest neighbor in the second point cloud
for at least one point in the first point cloud. In some
embodiments, determining an overlapping region comprises creating a
tree-structured data structure for at least one of the first or
second point clouds. In some embodiments, comparing surface
attributes comprises matching a surface associated with the first
point cloud with a surface associated with the second point. In
some embodiments, comparing surface attributes further comprises
evaluating a target function defined at least in part by a
plurality of points of the first and second point clouds that are
within the overlapping region. In some embodiments, generating at
least one calibration rule comprises optimizing the target
function. In some embodiments, the at least one calibration rule
comprises a rule for transformation between coordinate systems of
the first emitter/detector unit and the second emitter/detector
unit. In some embodiments, the computer-implemented method further
includes detecting a difference between the generated at least one
calibration rule against one or more previously generated
calibration rules. In some embodiments, the computer-implemented
method further includes causing calibration of the first and second
emitter/detector units in accordance with the at least one
calibration rule.
[0005] In other embodiments, a non-transitory computer-readable
medium stores computer-executable instructions. The instructions,
when executed, cause one or more processors associated with a
mobile platform to perform actions including: transforming point
cloud information obtained from a first emitter/detector unit into
a first point cloud in a reference system associated with the
mobile platform; transforming point cloud information obtained from
a second emitter/detector unit into a second point cloud in the
reference system associated with the mobile platform; determining
an overlapping region between the first and second point clouds;
comparing surface attributes of the first and second point clouds
in the overlapping region; and generating at least one calibration
rule for calibration between the first and second emitter/detector
units based at least in part on comparing the surface attributes.
In some embodiments, transforming point cloud information obtained
from the first emitter/detector unit is based at least in part on a
first set of transformation rules, which is at least partially
defined in accordance with a position and orientation of the first
emitter/detector unit relative to the mobile platform. In some
embodiments, the reference system associated with the mobile
platform comprises a coordinate system. In some embodiments, the
first set of transformation rules comprises a transform matrix. In
some embodiments, the first emitter/detector unit includes at least
one laser sensor that has a field of view (FOV) smaller than at
least one of 360 degrees, 180 degrees, 90 degrees, or 60 degrees.
In some embodiments, the first emitter/detector unit includes a
plurality of laser sensors rigidly fixed relative to each other. In
some embodiments, determining an overlapping region comprises
determining at least a nearest neighbor in the second point cloud
for at least one point in the first point cloud. In some
embodiments, determining an overlapping region comprises creating a
tree-structured data structure for at least one of the first or
second point clouds. In some embodiments, comparing surface
attributes comprises matching a surface associated with the first
point cloud with a surface associated with the second point. In
some embodiments, comparing surface attributes further comprises
evaluating a target function defined at least in part by a
plurality of points of the first and second point clouds that are
within the overlapping region. In some embodiments, generating at
least one calibration rule comprises optimizing the target
function. In some embodiments, the at least one calibration rule
comprises a rule for transformation between coordinate systems of
the first emitter/detector unit and the second emitter/detector
unit. In some embodiments, the actions further include detecting a
difference between the generated at least one calibration rule
against one or more previously generated calibration rules. In some
embodiments, the actions further include causing calibration of the
first and second emitter/detector units in accordance with the at
least one calibration rule.
[0006] In still further embodiments, a vehicle includes a
programmed controller that at least partially controls one or more
motions of the vehicle. The programmed controller includes one or
more processors that are configured to: transform point cloud
information obtained from a first emitter/detector unit into a
first point cloud in a reference system associated with the
vehicle; transform point cloud information obtained from a second
emitter/detector unit into a second point cloud in the reference
system associated with the vehicle; determine an overlapping region
between the first and second point clouds; compare surface
attributes of the first and second point clouds in the overlapping
region; and generate at least one calibration rule for calibration
between the first and second emitter/detector units based at least
in part on the comparison of surface attributes. In some
embodiments, transforming point cloud information obtained from the
first emitter/detector unit is based at least in part on a first
set of transformation rules, which is at least partially defined in
accordance with a position and orientation of the first
emitter/detector unit relative to the vehicle. In some embodiments,
transforming point cloud information obtained from the second
emitter/detector unit is based at least in part on a second set of
transformation rules and wherein the second set of transformation
rules differs from the first set of transformation rules. In some
embodiments, the first set of transformation rules comprises a
transform matrix. In some embodiments, the first emitter/detector
unit includes a plurality of laser sensors rigidly fixed relative
to each other. In some embodiments, determining an overlapping
region comprises determining at least a nearest neighbor in the
second point cloud for at least one point in the first point cloud.
In some embodiments, comparing surface attributes comprises
matching a surface associated with the first point cloud with a
surface associated with the second point, and matching a surface
associated with the first point cloud with a surface associated
with the second point comprises determining normal vector
information with respect to at least a portion of the first point
cloud. In some embodiments, comparing surface attributes further
comprises evaluating a target function defined at least in part by
a plurality of points of the first and second point clouds that are
within the overlapping region, wherein the target function
comprises a rotational component and a translational component. In
some embodiments, the at least one calibration rule comprises a
rule for transformation between coordinate systems of the first
emitter/detector unit and the second emitter/detector unit. In some
embodiments, the one or more processors are further configured to
detect a difference between the generated at least one calibration
rule against one or more previously generated calibration rules. In
some embodiments, the vehicle includes at least one of an unmanned
aerial vehicle (UAV), a manned aircraft, an autonomous car, a
self-balancing vehicle, or a robot. In some embodiments, the one or
more processors are further configured to cause calibration of the
first and second emitter/detector units in accordance with the at
least one calibration rule.
[0007] In yet still further embodiments, a computer-implemented
method for automatically detecting disturbance to an
emitter/detector unit carried by a mobile platform includes
transforming first point cloud information into a first point cloud
in a reference system associated with the mobile platform, the
first point cloud information obtained from the emitter/detector
unit at a first point in time; transforming second point cloud
information into a second point cloud in the reference system
associated with the mobile platform, the second point cloud
information obtained from the emitter/detector unit at a second
point in time; determining an overlapping region between the first
and second point clouds; comparing surface attributes of the first
and second point clouds in the overlapping region; and detecting a
disturbance to the emitter/detector unit based at least in part on
comparing the surface attributes. In some embodiments, transforming
first point cloud information obtained from the emitter/detector
unit is based at least in part on a set of transformation rules,
which is at least partially defined in accordance with a position
and orientation of the emitter/detector unit relative to the mobile
platform. In some embodiments, the reference system associated with
the mobile platform comprises a coordinate system. In some
embodiments, the reference system associated with the mobile
platform corresponds to a reference system at the first point in
time. In some embodiments, the set of transformation rules
comprises a transform matrix. In some embodiments, the
emitter/detector unit includes at least one laser sensor that has a
field of view (FOV) smaller than at least one of 360 degrees, 180
degrees, 90 degrees, or 60 degrees. In some embodiments, the
emitter/detector unit includes a plurality of laser sensors rigidly
fixed relative to each other. In some embodiments, determining an
overlapping region comprises determining at least a nearest
neighbor in the second point cloud for at least one point in the
first point cloud. In some embodiments, determining an overlapping
region comprises creating a tree-structured data structure for at
least one of the first or second point clouds, wherein the
tree-structured data structure comprises a K-Dimensional (KD) tree
data structure. In some embodiments, comparing surface attributes
comprises matching a surface associated with the first point cloud
with a surface associated with the second point, wherein matching a
surface associated with the first point cloud with a surface
associated with the second point comprises determining normal
vector information with respect to at least a portion of the first
point cloud. In some embodiments, comparing surface attributes
further comprises evaluating a target function defined at least in
part by a plurality of points of the first and second point clouds
that are within the overlapping region, wherein the target function
comprises a rotational component and a translational component. In
some embodiments, detecting a disturbance to the emitter/detector
unit comprises optimizing a target function based at least in part
on a least squares method. In some embodiments, detecting a
disturbance to the emitter/detector unit further comprises
generating at least one rule for transformation between the first
point cloud and the second point cloud. In some embodiments, the
mobile platform includes at least one of an unmanned aerial vehicle
(UAV), a manned aircraft, an autonomous car, a self-balancing
vehicle, a robot, a smart wearable device, a virtual reality (VR)
head-mounted display, or an augmented reality (AR) head-mounted
display. In some embodiments, the computer-implemented method
further includes issuing a warning in response to detecting the
disturbance to the emitter/detector unit.
BRIEF DESCRIPTION OF THE DRAWINGS
[0008] FIG. 1A illustrates a scanning pattern of a laser sensor
that can be utilized in accordance with some embodiments of the
presently disclosed technology.
[0009] FIG. 1B illustrates a three-dimensional point cloud
generated by a laser sensor in accordance with some embodiments of
the presently disclosed technology.
[0010] FIG. 2 illustrates a layout of multiple laser sensors
carried by a mobile platform to achieve wide angle (e.g.,
omni-directional) horizontal field of view (FOV) coverage in
accordance with some embodiments of the presently disclosed
technology.
[0011] FIG. 3 illustrates a laser unit including multiple laser
sensors that can be utilized in accordance with some embodiments of
the presently disclosed technology.
[0012] FIG. 4 illustrates a layout of multiple laser units carried
by a mobile platform to achieve wide angle (e.g., omni-directional)
horizontal FOV coverage in accordance with some embodiments of the
presently disclosed technology.
[0013] FIG. 5 illustrates a mobile platform with two laser units
(or laser sensors) with overlapping FOVs in accordance with some
embodiments of the presently disclosed technology.
[0014] FIG. 6 illustrates a calibration process for two laser units
(or laser sensors) in accordance with some embodiments of the
presently disclosed technology.
[0015] FIG. 7A illustrates movement of a mobile platform from a
first point in time to a second point in time, in accordance with
some embodiments of the presently disclosed technology.
[0016] FIG. 7B illustrates a process for detecting disturbance to a
laser unit in accordance with some embodiments of the presently
disclosed technology.
[0017] FIG. 8 illustrates one frame of laser points of a laser
sensor in accordance with some embodiments of the presently
disclosed technology.
[0018] FIG. 9 illustrates a sequence of frames of point data
generated by a laser unit (or laser sensor) carried by a mobile
platform that moves during a period of time, in accordance with
some embodiments of the presently disclosed technology.
[0019] FIG. 10 illustrates a combined point cloud generated in
accordance with some embodiments of the presently disclosed
technology.
[0020] FIG. 11 illustrates a mobile platform that carries multiple
sensors in addition to a laser unit, in accordance with some
embodiments of the presently disclosed technology.
[0021] FIG. 12 illustrates information that can be provided by the
multiple sensors of FIG. 11, in accordance with some embodiments of
the presently disclosed technology.
[0022] FIG. 13 illustrates data collection frequency differences of
the multiple sensors and the laser unit of FIG. 11, in accordance
with some embodiments of the presently disclosed technology.
[0023] FIG. 14 illustrates a process for combining time sequenced
point information generated by a laser unit to form a point cloud
in accordance with some embodiments of the presently disclosed
technology.
[0024] FIG. 15 illustrates examples of mobile platforms configured
in accordance with some embodiments of the presently disclosed
technology.
[0025] FIG. 16 is a block diagram illustrating a representative
architecture for a computer system or other control device that can
be utilized to implement various portions of the presently
disclosed technology.
DETAILED DESCRIPTION
1. Overview
[0026] To achieve accurate and comprehensive environment mapping
while overcoming the deficiencies associated with omni-directional
laser sensors, multiple laser sensors can be used (e.g.,
strategically affixed to a mobile platform) to enable wider
horizontal field of view (FOV) coverage, omni-directional
horizontal coverage, partial or complete spherical coverage, or any
other suitable customized coverage of the surrounding environment.
Laser sensors with limited FOV can be significantly cheaper than
omni-directional laser sensors and as used herein typically refer
to laser sensors with a horizontal FOV smaller than 360 degrees,
180 degrees, 90 degrees, or 60 degrees. Due to the multiplicity of
sensors that are typically used to achieve the desired angular
coverage, proper calibration between sensors is used to accurately
align of point cloud data generated by different sensors, thereby
providing a meaningful and reliable mapping of the surrounding
environment. Incorrect calibration between the sensors can distort
the alignment of point clouds, cause errors in environment mapping,
and thus bring about undesirable changes to the navigation,
movement, and/or other functions of the mobile platform.
Furthermore, when the mobile platform is actively deployed,
external vibration or other disturbances may cause changes in the
position or orientation of originally fixed sensors, therefore
causing calibration errors. Accordingly, detecting such changes
accurately in real time can further contribute to the reliability
and safety of the mobile platform.
[0027] The technology disclosed herein is generally directed to
calibrating and/or detecting errors in laser sensors carried by a
mobile platform. As will be discussed in further detail below, some
embodiments of the presently disclosed technology include a
multi-laser calibration method that accounts for at least: 1)
limitations to the FOV of the laser sensors; and 2) the desire or
requirement for omni-directional, spherical, and/or other
customizable coverage via a multi-laser system. In some
embodiments, the presently disclosed technology uses individual
laser units each including multiple fixed laser sensors to reduce
the number of calibrations within the system. In some embodiments,
the presently disclosed technology utilizes a method for detecting
a common overlapping region between the point clouds generated by
at least two laser sensors or units, as well as surface matching or
comparison of the point clouds within the overlapping region, to
generate rules for high precision calibration rules for laser
sensors or units.
[0028] Several details describing structures or processes that are
well-known and often associated with mobile platforms (e.g., UAVs
or other types of movable objects) and corresponding systems and
subsystems, but that may unnecessarily obscure some significant
aspects of the presently disclosed technology, are not set forth in
the following description for purposes of clarity. Moreover,
although the following disclosure sets forth several embodiments of
different aspects of the presently disclosed technology, several
other embodiments can have different configurations or different
components than those described herein. Accordingly, the presently
disclosed technology may have other embodiments with additional
elements and/or without several of the elements described below
with reference to FIGS. 1-16.
[0029] FIGS. 1-16 are provided to illustrate representative
embodiments of the presently disclosed technology. Unless provided
for otherwise, the drawings are not intended to limit the scope of
the claims in the present application.
[0030] Many embodiments of the technology described below may take
the form of computer- or controller-executable instructions,
including routines executed by a programmable computer or
controller. The programmable computer or controller may or may not
reside on a corresponding mobile platform. For example, the
programmable computer or controller can be an onboard computer of
the mobile platform, or a separate but dedicated computer
associated with the mobile platform, or part of a network or cloud
based computing service. Those skilled in the relevant art will
appreciate that the technology can be practiced on computer or
controller systems other than those shown and described below. The
technology can be embodied in a special-purpose computer or data
processor that is specifically programmed, configured or
constructed to perform one or more of the computer-executable
instructions described below. Accordingly, the terms "computer" and
"controller" as generally used herein refer to any data processor
and can include Internet appliances and handheld devices (including
palm-top computers, wearable computers, cellular or mobile phones,
multi-processor systems, processor-based or programmable consumer
electronics, network computers, mini computers and the like).
Information handled by these computers and controllers can be
presented at any suitable display medium, including an LCD (liquid
crystal display). Instructions for performing computer- or
controller-executable tasks can be stored in or on any suitable
computer-readable medium, including hardware, firmware or a
combination of hardware and firmware. Instructions can be contained
in any suitable memory device, including, for example, a flash
drive, USB (universal serial bus) device, and/or other suitable
medium.
2. Representative Embodiments
[0031] FIG. 1A illustrates a scanning pattern 102a of a laser
sensor that can be utilized in accordance with some embodiments of
the presently disclosed technology. As illustrated in FIG. 1A, the
FOV of the laser sensor is limited (e.g., no larger than 60 degrees
in both horizontal and vertical directions). FIG. 1B illustrates a
three-dimensional point cloud generated by a laser sensor (e.g.,
the laser sensor illustrated in FIG. 1A). Compared with a
conventional omni-directional laser sensor, a laser sensor with
limited FOV cannot provide a uniformly-distributed, 360-degree
three-dimensional point cloud, but may provide denser or sparser
scanning points in certain portions of its FOV (e.g., a higher
density of points in a forward area of the sensor FOV as
illustrated in FIG. 1B).
[0032] FIG. 2 illustrates a layout of multiple laser sensors 204
carried by a mobile platform 202 to achieve wide angle (e.g.,
omni-directional) horizontal FOV coverage in accordance with some
embodiments of the presently disclosed technology. The laser
sensors 204 can be distributed and/or oriented differently to
achieve wide angle FOV coverage, partial or complete spherical
coverage, as well as other customizable FOV coverage. In the
illustrative layout of FIG. 2, depending on the FOV of each laser
sensor 204, the overlapping region 206 between the FOVs of two
neighboring sensors 204 may be limited, which may provide
insufficient or inadequate geometrical information for
high-resolution calibrations. In addition, depending on the
quantity of laser sensors 204 needed to achieve a wide angle
coverage, the number of calibrations between sensors can be
significant. For example, even if only a pairwise calibration
between neighboring sensors 204 is to be performed, at least twelve
set of calibration rules need to be determined for the layout shown
in FIG. 2.
[0033] The presently disclosed technology can: 1) achieve an
overlapping region of an adequate size (e.g., exceeding a threshold
of overlapped FOVs) between sensors; 2) strategically distribute
and orient multiple sensors with respect to the mobile platform to
achieve wide angle (e.g., omni-directional), spherical, and/or
other customizable coverage; and/or 3) reduce the number of laser
sensors utilized. In this regard, FIG. 3 illustrates a laser unit
300 that includes multiple laser sensors 302, 304, and 306 that can
be used in accordance with some embodiments of the presently
disclosed technology.
[0034] With reference to FIG. 3, the laser unit 300 may include a
mechanical structure 308 (e.g., a metal frame) to which two or more
laser sensors with a limited FOV are rigidly or fixedly connected
(e.g., welded to a metal frame) at a certain position and
orientation with respect to one another. In some embodiments, the
mechanical structure 308 may correspond to a part of a mobile
platform. In the embodiment as illustrated in FIG. 3, three laser
sensors 302, 304, and 306 are positioned and oriented in a manner
that expands a 60-degree horizontal FOV coverage for each sensor to
approximately 160 degrees of horizontal FOV for the laser unit 300.
Illustratively, angles between neighboring laser units can be set
to 50 degrees to allow for suitable FOV overlap.
[0035] The calibration rules for calibrating between or among laser
sensors within a laser unit 300 can be known and fixed. For
example, laser sensors within one laser unit 300 may be
pre-calibrated with respect to each other manually, using the same
calibration technology as disclosed herein, or based on calibration
methods known to those of skill in the relevant art. As discussed
above, the relative positions and orientations of the multiple
laser sensors are unlikely to change within a laser unit 300.
Therefore, based on the pre-calibration, the laser unit 300 can
consistently generate or otherwise output point cloud data that
covers a wider FOV than each constituent laser sensor.
[0036] FIG. 4 illustrates a layout of multiple laser units 404
carried by a mobile platform 402 to achieve wide angle (e.g.,
omni-directional) horizontal FOV coverage in accordance with some
embodiments of the presently disclosed technology. Each laser unit
404 may have a configuration like that of the laser unit 300
illustrated in FIG. 3. The laser units 404 (possibly in combination
with laser sensors 204 described above with reference to FIG. 2)
can be distributed and/or oriented differently to achieve wide
angle FOV coverage, partial or complete spherical coverage, as well
as other customizable FOV coverage. With reference to FIG. 4, four
laser units 404 are distributed and oriented in four respective
directions, e.g., at 45 degrees, 135 degrees, 225 degrees, and 315
degrees in accordance with a coordinate system centered on the
mobile platform 402. As discussed earlier, laser sensors within
individual laser units 404 are fixed relative to each other.
Therefore the number of calibrations (e.g., between laser units
404) is reduced as compared to the configuration of FIG. 2. For
example, if only pairwise calibrations between neighboring laser
units 404 are to be performed, only four sets of calibration rules
need to be determined. At the same time, an overlapping region
between FOVs of neighboring laser units 404 can be large enough to
provide sufficient geometrical information to enhance the precision
of the calibration between the laser units 404.
[0037] FIG. 5 illustrates a mobile platform 512 with first and
second laser units (or laser sensors) 508 and 510 with
corresponding FOVs 504 and 506 that overlap, in accordance with
some embodiments of the presently disclosed technology. As
illustrated in FIG. 5, the two laser units 508 and 510 are arranged
in a forward-looking manner and a relative distance between them is
limited to provide an overlapping region 502 of adequate size. In
other embodiments, the laser units or sensors can be arranged in
different manners to provide adequately sized overlapping regions.
With reference to FIG. 5, illustratively, a set of indicators "x"
represents a point cloud 514 generated by the first laser unit 508
and a set of indicators "o" represents a point cloud 516 generated
by the second laser unit 510.
[0038] With continued reference to FIG. 5, the mobile platform 512
can be associated with a coordinate system F.sup.r, the first laser
unit 508 can be associated with a coordinate system F.sub.1.sup.l,
and the second laser unit 510 can be associated with a coordinate
system F.sub.2.sup.l. Initial transformation matrices for
transforming coordinates between the mobile platform coordinate
system F.sup.r and the laser unit coordinate systems F.sub.1.sup.l
and F.sub.2.sup.l can be expressed as .sub.rT.sub.l.sub.1 and
.sub.rT.sub.l.sub.2, respectively. In some embodiments, the initial
transformation matrices can be determined prior to deploying the
mobile platform, for example, in accordance with the position and
orientation of the respective laser units relative to the mobile
platform. Suitable manual or semi-automatic calibration methods
known to those of skill in the relevant art may be used to
determine the initial transformation matrices. In some embodiments,
the transformation matrices are all homogeneous matrices of a
4.times.4 size. In some embodiments, an initial transformation
matrix .sub.l.sub.1T.sub.l.sub.2 for transforming directly between
coordinate systems of the first and second laser units 508 and 510
can be determined prior to deploying the mobile platform as well,
for example, based on the values of .sub.rT.sub.l.sub.1 and
.sub.rT.sub.l.sub.2.
[0039] FIG. 6 illustrates a calibration process for two laser units
(or laser sensors) in accordance with some embodiments of the
presently disclosed technology. The calibration process of FIG. 6
can be implemented by a controller (e.g., an onboard computer of a
mobile platform, an associated computing device, and/or an
associated computing service). In step 605, the process includes
transforming point cloud information obtained by the two laser
units into point clouds in a reference system based on initial
transformation rules. For example, laser point clouds 514 and 516
are captured by the two laser units 508 and 510 in accordance with
their corresponding coordinate systems F.sub.1.sup.l and
F.sub.2.sup.l, respectively. The laser point clouds 514 and 516 can
be projected into the coordinate system F.sup.r of the mobile
platform 512 to form corresponding first and second transformed
point clouds C.sub.1.sup.r and C.sub.2.sup.r, based on initial
transformation matrices .sub.rT.sub.l.sub.1 and
.sub.rT.sub.l.sub.2, respectively.
[0040] In step 610, the process includes determining an overlapping
region between the transformed point clouds in the reference
system. Illustratively, assuming the initial transformation
matrices .sub.rT.sub.l.sub.1 and .sub.rT.sub.l.sub.2 are properly
determined (e.g., close to their truth values), the surface shapes
of transformed point clouds C.sub.1.sup.r and C.sub.2.sup.r within
the overlapping region should not deviate significantly from each
other. Illustratively, the controller sets up K-Dimensional (KD)
tree structures t.sub.1.sup.r and t.sub.2.sup.r respectively,
according to the two point clouds C.sub.1.sup.r and C.sub.2.sup.r
projected to the coordinate system of the mobile platform 512. By
using this type of tree structure, for any point of the first point
cloud C.sub.1.sup.r, the nearest neighbor in the second point cloud
C.sub.2.sup.r can be retrieved quickly, and vice versa. If a
distance between a query point in the first point cloud (e.g.,
C.sub.1.sup.r) and its nearest neighbor in the second point cloud
(e.g., C.sub.2.sup.r) is less than a specified threshold, the
process can include labeling or classifying the query point as
being located in the overlapping region. The process can identify
all query points in either or both point clouds C.sub.1.sup.r and
C.sub.2.sup.r that satisfy the threshold requirement, thus
determining the overlapping region.
[0041] In step 615, the process includes evaluating a match between
surfaces of the transformed point clouds in the overlapping region.
Illustratively, for each point of a specific transformed point
cloud (e.g., C.sub.1.sup.r) in the overlapping region, the nearest
neighbor thereof in the specific point cloud can also be quickly
retrieved by using the corresponding KD tree structure (e.g.,
t.sub.1.sup.r) that was set up in step 610. Accordingly, for each
point of the specific transformed point cloud, a specified number
of (nearest) neighbor points can be selected to form a plane, from
which a normal vector corresponding to the point can be readily
determined.
[0042] In some embodiments, to match surfaces represented or
indicated by the point clouds in the overlapping region, the
controller may implement a Point-to-Plane Iterative Closest Point
(Point-to-Plane ICP) method. Illustratively, after the overlapping
region is found and the normal vectors for at least one point cloud
within the overlapping region are determined, the controller can
use the Point-To-Plane ICP method to minimize the following target
function:
E = i ( ( Rp i + t - q i ) n i ) 2 ##EQU00001##
where p.sub.i denotes a point within the overlapping region of the
first point cloud C.sub.1.sup.r, q.sub.i denotes a nearest point
within the overlapping region of the second point cloud
C.sub.2.sup.r, R denotes a rotation matrix, t denotes a translation
vector, and n.sub.i denotes a normal vector with respect to
p.sub.i. When an adequate number of points (e.g., exceeding a
threshold number) are available in the overlapping region, the
minimization can be achieved based, for example, on a least squares
method.
[0043] In step 620, the process includes generating calibration
rules for calibration between the two laser units or sensors.
Illustratively, when the target function is minimized, an optimal
solution of transformation (e.g., optimal values for rotation
matrix R and translation vector t) is achieved. In some
embodiments, a translational change between the laser units or
sensors is less likely to occur because different laser units or
sensors are fixedly connected (e.g., via brackets). However, the
laser units or sensors may be more likely to rotate, for example,
if they are connected to the brackets by screws. In these
embodiments, the translation vector t can be fixed to a constant
(e.g., a value determined based on a prior minimization of the
target function), so that the controller may estimate the
rotational matrix R more efficiently. In some embodiments, after
the calibration rules are generated, the controller calibrates the
two laser units or sensors based thereon. Illustratively, the
controller may align point clouds generated from the two laser
units or sensors using the rotational matrix rotation matrix R and
translation vector t, until they are updated in a next round of
calibration.
[0044] In step 625, the process includes comparing newly generated
calibration rules against previously generated calibrations rules.
Illustratively, the onboard computer may compare newly determined
optimal values for the rotation matrix R and/or the translation
vector t against their optimal values determined in an initial
round of calibration, a most recent round of calibration, an
average or weighted average of several recent rounds, or the like.
In step 630, the process includes determining whether the
difference that results from the comparison in step 625 exceeds a
threshold. If not, the process proceeds to step 605 for a new round
of calibration. If the difference exceeds the threshold, the
process proceeds to step 635.
[0045] In step 635, the process includes taking one or more further
actions. The difference exceeding the threshold may indicate that
the two laser units or sensors cannot be reliably calibrated with
each other. For example, the physical position or orientation of at
least one of the laser units or sensors may have deviated
substantially from a preset configuration. In this case, the
controller may issue a warning to an operator of the mobile
platform. Alternatively, the controller may suspend the navigation
or other functions of the mobile platform in a safe manner.
[0046] FIG. 7A illustrates the movement of a mobile platform 702
between a first point in time t.sub.1 and a second point in time
t.sub.2, in accordance with some embodiments of the presently
disclosed technology. Illustratively, a transformation matrix
.sub.r.sub.1T.sub.r.sub.2 for transforming between the mobile
platform's coordinate system F.sub.1.sup.r at time t.sub.1 and the
mobile platform's coordinate system F.sub.2.sup.r at time t.sub.2
is known or can be determined, for example, based on measurements
collected by a global positioning system and/or an inertial
measurement unit carried by the mobile platform 702. Based on a
known transformation relationship (e.g., the transformation matrix
.sub.r.sub.1T.sub.r.sub.2) between the mobile platform's two
positions at two different points in time, representative systems
can detect disturbances to and/or calibrate the transformation
between the coordinate systems of a single laser unit (or laser
sensor) 704 and the mobile platform body can be achieved.
[0047] FIG. 7B illustrates a process for detecting disturbance to a
laser unit (e.g., the laser unit 704 carried by the mobile platform
702 as illustrated in FIG. 7A) in accordance with some embodiments
of the presently disclosed technology. The process can be
implemented by a controller (e.g., an onboard computer of the
mobile platform, an associated computing device, and/or an
associated computing service). The FOVs of the laser unit 704 at
times t.sub.1 and t.sub.2 may have a sufficiently large overlapping
region to facilitate a disturbance detection process similar to the
calibration process described above with reference to FIG. 6.
Similar to generating calibration rules for calibration between two
different laser units, the process of FIG. 6 can be modified to
generate calibration rules for calibration of a single laser unit
at two points in time (or two positions/orientations), and to
detect disturbances to the laser unit based thereon.
[0048] Step 705 of the disturbance detection process includes
transforming point cloud information obtained by the laser unit at
two points in time into respective point clouds in a reference
system. Illustratively, given (1) the transformation matrix
.sub.r.sub.1T.sub.r.sub.2 of the mobile platform 702 between two
times t.sub.1 and t.sub.2, (2) initial values for a transformation
matrix .sub.r.sub.1T.sub.l.sub.1 between the mobile platform
coordinate system F.sub.1.sup.r and the laser unit coordinate
system F.sub.1.sup.l at time t.sub.1, and (3) initial values for a
transformation matrix .sub.r.sub.2T.sub.l.sub.2 between the mobile
platform coordinate system F.sub.2.sup.r and the laser unit
coordinate system F.sub.2.sup.l at moment t.sub.2, the controller
can project point clouds collected by the laser unit 704 at the two
times, t.sub.1 and t.sub.2, to the mobile platform's coordinate
system F.sub.1.sup.r at time t.sub.1. As discussed earlier, the
initial values for a transformation matrix between the mobile
platform and the laser unit can be known or pre-determined in
accordance with the position and orientation of the laser unit
relative to the mobile platform. Suitable manual or semi-automatic
calibration methods known to those of skill in the relevant art may
be used to determine the initial transformation matrix, for
example, prior to deploying the mobile platform. Accordingly, in
some embodiments, the initial values for .sub.r.sub.1T.sub.l.sub.1
and .sub.r.sub.2T.sub.l.sub.2 may be the same.
[0049] Step 710 of the disturbance detection process includes
determining an overlapping region between the transformed point
clouds. Illustratively, the projected point clouds can be expressed
as .sub.r.sub.1P.sub.1 (for time t.sub.1) and .sub.r.sub.1P.sub.2
(for time t.sub.2), wherein theoretically
.sub.r.sub.1P.sub.2=.sub.r.sub.1T.sub.r.sub.2.sub.r.sub.2P.sub.2.
Similar to the calibration process of FIG. 6, the controller may
set up KD tree structures for both projected point clouds and
determine their respective subsets of points in an overlapping
region: .sub.r.sub.1{circumflex over (P)}.sub.1 and
.sub.r.sub.1{circumflex over (P)}.sub.2, respectively.
[0050] Step 715 of the disturbance detection process includes
evaluating a match between surfaces indicated or represented by the
transformed point clouds in the overlapping region. Illustratively,
similar to the calibration process of FIG. 6, the controller may
compare the surfaces of the two projected point clouds within the
overlapping region. After the controller estimates normal vectors
n.sub.1,i for each point within .sub.r.sub.1{circumflex over
(P)}.sub.1, the controller may evaluate the following target
function:
H = i ( ( Rp 1 , i + t - p 2 , i ) n 1 , i ) 2 ##EQU00002##
wherein p.sub.1,i.di-elect cons.{circumflex over (P)}.sub.1,
p.sub.2,i.di-elect cons..sub.r.sub.1{circumflex over
(P)}.sub.2.
[0051] Step 720 of the disturbance detection process includes
detecting whether there is disturbance to the laser unit.
Illustratively, the controller can minimize the target function H,
for example, based on a least squares method. If the minimized
value of function H exceeds a threshold, or if the rotational
matrix R and/or the translation vector t exceed or deviate from a
certain threshold (e.g., with respect to
.sub.r.sub.1T.sub.r.sub.2), in step 725 of the process, the
controller can determine that the laser unit 704 has been disturbed
between times t.sub.1 and t.sub.2 (e.g., due to a loosened screw or
impact by an external object). In this case, the process proceeds
to step 730 where the one or more further actions can be taken. For
example, the controller may issue a warning to an operator of the
mobile platform 702 and/or take other actions. If no disturbance is
detected in step 725, the process proceeds to step 705. Similar to
the calibration process of FIG. 6, this disturbance detection
process using the coordinate systems of a single laser unit and the
mobile platform body can be performed repeatedly and periodically
while the mobile platform is deployed.
[0052] In the use of certain laser units or sensors, the number
and/or distribution of laser scanning points in a single frame may
not provide a sufficiently dense point cloud to facilitate
calibration, mapping, object detection, and/or positioning. This
problem may be particularly apparent in the use of low-cost
small-angle LiDAR sensors. For example, for a typical low-cost
small-angle LiDAR, the number of laser points in a single frame can
be fewer than 4000 or even 2000, whereas a more expensive
omni-directional LiDAR may produce 288000 laser scanning points in
a single frame.
[0053] FIG. 8 illustrates one frame 800 of laser scanning points
produced by a laser sensor (e.g., a small-angle laser sensor that
implements a scanning pattern 102a of FIG. 1). As illustrated, a
sparse set of laser scanning points 810 is distributed in a
three-dimensional coordinate system 820 in a non-uniform manner. In
some embodiments, the sparse and non-uniform distribution of points
810 may not provide enough data in an overlapping region for the
calibration process disclosed herein to perform reliably.
[0054] FIG. 9 illustrates a sequence of frames of scanning point
data generated by a laser unit (or laser sensor) that is carried by
a mobile platform, which moves during a period of time. As
illustrated in FIG. 9, a laser unit (or laser sensor) 915 carried
by a mobile platform 910 generates multiple frames 920 of scanning
point data during a time period from t.sub.i to t.sub.i+k. For
example, frame 920a is generated at time t.sub.+1 with the mobile
platform 910 (and the laser unit or sensor) situated in a first
position/orientation, frame 920b is generated at a subsequent time
t.sub.i+1 with the mobile platform 910 (and the laser unit or
sensor) situated in a second position/orientation, and frame 920c
is generated at a subsequent moment t.sub.i+2 with the mobile
platform 910 (and the laser unit or sensor) situated in a third
position/orientation. Portions of the presently disclosed
technology can generate a combined point cloud based on temporally
sequenced sets of point data, such as the sequence of frames
920.
[0055] FIG. 10 illustrates a combined point cloud 1000 generated in
accordance with some embodiments of the presently disclosed
technology. As illustrated in FIG. 10, a dense set of laser
scanning points 1010 that combines multiple sets of laser scanning
points (e.g., similar to those of frame 800 in FIG. 8) is
distributed in a three-dimensional coordinate system 1020 in a
relatively uniform manner to provide comprehensive
three-dimensional environmental information. In some embodiments,
the calibration process disclosed herein utilizes such combined
point clouds rather than single frame of point data.
[0056] To combine multiple frames of point data in a manner that
reduces noise and error, the presently disclosed technology
includes estimating a relative transformation matrix between
successive frames by using multiple types of sensors carried by a
mobile platform.
[0057] FIG. 11 illustrates a mobile platform 1120 that carries
multiple sensors in addition to a laser unit (or sensor), in
accordance with some embodiments of the presently disclosed
technology. As illustrated, the mobile platform may carry a stereo
camera 1104, an inertial measurement unit 1106, a wheel encoder
1110, and/or a global positioning system (GPS) 1102, in addition to
a laser unit 1108. Those of skill in the relevant art will
appreciate that fewer, more, or alternative sensors may be used by
the presently disclosed technology. For example, instead of using
the stereo camera 804, a set, array, or system of multiple cameras
can be used.
[0058] FIG. 12 illustrates information that can be provided by the
multiple sensors of FIG. 11. The stereo camera 1104 can provide
three-dimensional coordinates of environmental features 1202 (e.g.,
one or more distinctive points in three dimensional space of
surrounding environment), which may establish a constraint
relationship between successive frames (e.g., corresponding to
observations from two different positions 1220a and 1220b).
Illustratively, the sampling frequency or data acquisition rate of
the stereo camera 1104 is between 20 Hz and 40 Hz. The inertial
measurement unit 1106 can provide high-frequency acceleration
information and angular velocity information. Illustratively, the
sampling frequency or data acquisition rate of the inertial
measurement unit is 200 Hz or higher. Via integration, a
transformation matrix of the mobile platform 1120 between two
successive frames can be calculated. The wheel encoder 1110 can
provide the rotation speed of the powered wheels (e.g., rear
wheels) and steering information of the front wheels, and can
provide, according to a known wheel size, constraints on forward
speeds and deflection angles between successive frames.
Illustratively, the sampling frequency or data acquisition rate of
the wheel encoder is about 20 Hz. Depending on outdoor signal
conditions, the GPS 1102 can provide the position of the mobile
platform 1120 and attitude information thereof in global system.
Illustratively, the sampling frequency or data acquisition rate of
the GPS is below 5 Hz. Illustratively, the laser unit 1108 (e.g.,
including one or more LiDAR sensors) has a sampling frequency or
data acquisition rate of 10 Hz.
[0059] The table below summarizes typical data acquisition
frequency information of the representative sensors illustrated in
FIGS. 11 and 12:
TABLE-US-00001 Sensor Frequency Laser 10 hz Stereo camera 20 hz to
40 hz Inertial measurement unit >200 hz Wheel encoder
approximately 20 hz Global positioning system 5 hz
[0060] FIG. 13 illustrates data collection frequency differences of
the multiple sensors and the laser unit of FIG. 11, in accordance
with some embodiments of the presently disclosed technology.
[0061] FIG. 14 illustrates a process for combining time sequenced
point information generated by a laser unit to form a point cloud
in accordance with some embodiments of the presently disclosed
technology. The process can be implemented by a controller (e.g.,
an onboard computer of a mobile platform, an associated computing
device, and/or an associated computing service). As part of the
presently disclosed technology, generating a combined point cloud
can include estimating relative states associated with the laser
unit over a period of time, instead of estimating all subsequent
states with respect to a global coordinate system. Illustratively,
embodiments of the presently disclosed technology estimate relative
position information of the laser unit with respect to two or more
different frames that it generates in the period of time, thereby
enabling accurate accumulation of laser point data from different
frames in this period of time. This approach can facilitate or
enhance subsequent calibration, object detection, mapping, and/or
positioning operations.
[0062] Step 1405 of the process includes obtaining observation
data, corresponding to a period of time, from multiple observation
sensors (e.g., the multiple sensors as illustrated in FIG. 8). In
some embodiments, methods in accordance with the presently
disclosed technology include may make an approximation that data
from different sensors is synchronized. For example, in a
representative case the data acquisition frequency of the target
laser unit is 10 Hz, the frequency of the stereo camera is 40 Hz,
the frequency of the wheel encoder is 20 Hz, the frequency of the
inertial measurement unit is 200 Hz, and the frequency of the GPS
is 5 Hz. As an approximation, observation data from different
sensors can be considered as accurately aligned according to
different frequency multiples. Accordingly, using a 1-second time
window as an example, the controller can obtain 200 accelerometer
and gyroscope readings (from the inertial measurement unit), 40
frames of stereo camera observation, 20 groups of speed and
deflection angle observations (from the wheel encoder), and 5
pieces of GPS positioning information. Based on these, embodiments
of the presently disclosed technology can estimate relative
positions between 10 laser unit data acquisition events or
positions thereof with respect to a particular local coordinate
system (such as a local coordinate system corresponding to the
first of the 10 data acquisition events).
[0063] In some embodiments, the presently disclosed technology
includes a further approximation that the position of the laser
unit coincides with that of the stereo camera, thereby further
simplifying the problem to be solved. As discussed with reference
to FIG. 12, the observation data from the different sensors can be
described mathematically as follows: [0064] 1) According to the
observation data from the stereo camera, illustratively
three-dimensional coordinates and/or descriptor(s) of one or more
environmental features (e.g., feature 1202) can be extracted from
frames produced by the camera at positions 1220a and 1220b,
respectively. These coordinates and/or descriptor(s) can be matched
with respect to the feature 1202. In an objective function for
optimization, this type of observation can be embodied by an error
item relating to the re-projection of feature(s) onto the camera
coordinate systems at different positions. For example, the cost
term based on an environment feature and two consecutive frames of
stereo camera observation includes 3 parts: (a) a re-projection
error between the left camera and right camera at a frame
corresponding to position 1220a; (b) a re-projection error between
the left camera and right camera at a frame corresponding to
position 1220b; and (c) a re-projection error between the left (or
right) camera at two positions 1220a and 1220b. [0065] 2) According
to the observation data from the inertial measurement unit with
known timestamp and initial values, a constraint relationship of a
rotation matrix, a translation vector, and a speed between two
consecutive camera frames can be calculated, for example, by using
suitable integration techniques known to those of skill in the
relevant art. This type of observation can be embodied by an error
item between the post-integration state and a real state in the
objective function. Illustratively, the variables to be estimated
at each frame, e.g., camera frames corresponding to positions 1220a
and 1220b, include the camera's orientation (e.g., an element in
Special Orthogonal group), and position and velocity (e.g.,
elements in R.sup.3 space group.) Integration using observations
captured from the inertial measurement unit provides the
constraints between the variables explained above. In some
embodiments, while a state is optimized iteratively, suitable
pre-integration technique is adopted to improve computational
efficiency. [0066] 3) A motion model including the speed and
deflection angle of the mobile platform can be derived based on
observation data from the wheel encoder. Similarly, via
integration, a state constraint between consecutive camera frames
can be obtained, and the expression of this type of observation can
be similar to that of the inertial measurement unit. In some
embodiments, in contrast to the situation of the inertial
measurement unit, only a sub-space of the state is constrained
(e.g., the position and the yaw angle of the mobile platform) based
on the wheel odometer observations. Due to possible noise of the
wheel encoder, the covariance of this error term can be set to be
relatively larger in some embodiments. [0067] 4) The observation
data from the GPS can directly provide a constraint on a state of
the mobile platform at a particular time. In the objective
function, this type of observation can be expressed as an error
between an estimated state provided by the GPS and a real state
value. Due to the low data acquisition frequency of the GPS in some
embodiments, the GPS observation may only be used when its noise
level lower than certain threshold and/or its accuracy guaranteed
within certain range.
[0068] In embodiments for which the position of the laser unit is
approximated as coinciding with that of the stereo camera, a
controller (e.g., an onboard computer of the mobile platform, an
associated computing device, and/or an associated computing
service) obtains observation data that can be provided by the
sensors for a period of time from time 1 until time k. The
observation data can be expressed as follows:
.sub.k={C.sub.1:k,.sub.1:k-1,,.sub.1:q}
where
[0069] 1) the first element C.sub.1:k denotes observation
information obtained by the stereo camera, and may be defined as
follows:
where z.sub.i,j denotes an observation of a j.sup.th feature in the
i.sup.th frame by the stereo camera;
[0070] 2) the second element .sub.1:k-1 denotes a set of data
acquired by the inertial measurement unit until the k.sup.th point
in time, where .sub.i={I.sub.i, I.sub.i+1, I.sub.i+2, . . . ,
I.sub.i+m} denotes a set of all observations by the inertial
measurement unit between the i.sup.th frame produced by the camera
and the i+1.sup.th frame produced by camera (e.g., a total of 20
readings from the inertial measurement unit between 2 successive
camera observations);
[0071] 3) the third element .sub.1:p denotes the observation by the
wheel encoder, which may be expressed as follows:
where v.sub.i,j.sup.W denotes speed information obtained by the
wheel encoder at the i.sup.th point in time and the j.sup.th point
in time and q.sub.i,q.sup.W denotes a rotation transformation
(e.g., quaternion expression), which can be derived or otherwise
obtained by a deflection angle calculation, between the i.sup.th
point in time and the j.sup.th point in time; and
[0072] 4) the last element .sub.1:q expresses the observation
obtained by the GPS:
where p.sub.i.sup.G denotes a global position of the i.sup.th point
in time, and q.sub.i.sup.G denotes rotation with respect to a
global coordinate system.
[0073] Step 1410 of the process includes evaluating states
associated with the laser unit at different points in time within
the time period based on the observation data. For example, using a
factor graph, the controller may establish a relationship between
an a priori probability and an a posteriori probability associated
with states of the laser unit (coincident with the stereo
camera):
where k=[1, 2, . . . , k] denotes a set of observation indexes of
the camera, m denotes a set of observation indices of the GPS, and
a state of the laser unit can be expressed as:
x.sub.k=[p.sub.k,v.sub.k,q.sub.k]
where p.sub.k, v.sub.k, and q.sub.k respectively denote a position,
a speed, and a quaternion (rotation) of the laser unit with respect
to a particular coordinate system at the k.sup.th point in time. In
the above formula, each p( ) is called a factor of the factor
graph.
[0074] In some embodiments, using a mathematical derivation based
on an assumption of zero-mean Gaussian white noise, the controller
may compute a maximum-a-posteriori of the above factor graph based
formula by solving for a minimum of the following formula:
where r.sub.* represents different residual types, and
.SIGMA..sub.k denotes covariance matrices corresponding to
different types of residuals, and is used to describe the
uncertainty of the observation. In this regard, those of skill in
the relevant art can determine residual models for different
sensors and determine Jacobian matrices between optimization
iterations. The controller can calculate optimal values for the
laser unit states based on the minimization, for example, based on
a gradient-based optimization method.
[0075] Step 1415 of the process includes determining transformation
rules for transforming between multiple reference systems (e.g., at
different points in time) and a target reference system.
Illustratively, according to the following approximations: (1) the
positions of the stereo camera and laser unit coincide with each
other; and (2) timestamps of data acquired by the laser unit and
data acquired by the camera are exactly the same, the controller
can compute relative transformation matrices for the laser unit at
different points in time with respect to a target point in time
(i.e., when the subject period of time starts, half-way through the
subject time period, or when the subject period of time ends) using
corresponding states as determined.
[0076] In some embodiments, the approximations that (1) the
positions of the stereo camera and laser unit coincide with each
other; and (2) timestamps of data acquired by the laser unit and
data acquired by the camera are exactly the same are not used. In
these embodiments, the presently disclosed technology can account
for two factors: (1) relative changes (e.g., the transformation
matrix .sub.cT.sub.l between the stereo camera and the laser unit;
and (2) a timestamp difference between different sensors. Regarding
the first factor (1), because the laser unit and the stereo camera
are not likely to move relative to each other during the subject
period of time, the controller may calculate a relative position of
the laser unit at any q.sup.th point in time with respect to any
p.sup.th point in time during the subject time period by simply
calculating a relative position of the camera at time q with time
p. As for the second factor (2) where timestamps between different
sensors cannot be perfectly synchronized, the controller may use
interpolation (e.g., based on a polynomial fitting) to compute
relative position information in a coordinate system (e.g., a
coordinate system of the mobile platform) at the time of any
specified timestamp.
[0077] Step 1420 of the process includes transforming data obtained
by the laser unit at different points in time based on the
transformation rules. Illustratively, using the relative
transformation matrices as determined in step 1415, the controller
can re-project data (e.g., laser scanning points) acquired at
different points in time (e.g., different frames) in the subject
time period, to the target point in time. In some embodiments, the
controller can exclude certain points in time from the
re-projection process due to excessive noise, data error, or other
factors. Step 1425 of the process includes generating a combined
point cloud using the transformed data. Illustratively, the
controller can add the re-projected data from multiple (selected)
frames to the frame of point data initially associated with the
target point in time, thereby accumulating temporally sequenced
frames of data to form a combined point cloud as if the data were
all acquired by the laser unit at the target point in time.
[0078] FIG. 15 illustrates examples of mobile platforms configured
in accordance with various embodiments of the presently disclosed
technology. As illustrated, a representative mobile platform as
disclosed herein may include at least one of an unmanned aerial
vehicle (UAV) 1502, a manned aircraft 1504, an autonomous car 1506,
a self-balancing vehicle 1508, a terrestrial robot 1510, a smart
wearable device 1512, a virtual reality (VR) head-mounted display
1514, or an augmented reality (AR) head-mounted display 1516.
[0079] FIG. 16 is a block diagram illustrating an example of the
architecture for a computer system or other control device 1500
that can be utilized to implement various portions of the presently
disclosed technology. In FIG. 16, the computer system 1600 includes
one or more processors 1605 and memory 1610 connected via an
interconnect 1625. The interconnect 1625 may represent any one or
more separate physical buses, point to point connections, or both
connected by appropriate bridges, adapters, or controllers. The
interconnect 1625, therefore, may include, for example, a system
bus, a Peripheral Component Interconnect (PCI) bus, a
HyperTransport or industry standard architecture (ISA) bus, a small
computer system interface (SCSI) bus, a universal serial bus (USB),
IIC (I2C) bus, or an Institute of Electrical and Electronics
Engineers (IEEE) standard 674 bus, sometimes referred to as
"Firewire".
[0080] The processor(s) 1605 may include central processing units
(CPUs) to control the overall operation of, for example, the host
computer. In certain embodiments, the processor(s) 1605 accomplish
this by executing software or firmware stored in memory 1610. The
processor(s) 1605 may be, or may include, one or more programmable
general-purpose or special-purpose microprocessors, digital signal
processors (DSPs), programmable controllers, application specific
integrated circuits (ASICs), programmable logic devices (PLDs), or
the like, or a combination of such devices.
[0081] The memory 1610 is or includes the main memory of the
computer system. The memory 1610 represents any form of random
access memory (RAM), read-only memory (ROM), flash memory, or the
like, or a combination of such devices. In use, the memory 1610 may
contain, among other things, a set of machine instructions which,
when executed by processor 1605, causes the processor 1605 to
perform operations to implement embodiments of the present
invention.
[0082] Also connected to the processor(s) 1605 through the
interconnect 1625 is a (optional) network adapter 1615. The network
adapter 1615 provides the computer system 1600 with the ability to
communicate with remote devices, such as the storage clients,
and/or other storage servers, and may be, for example, an Ethernet
adapter or Fiber Channel adapter.
[0083] The techniques introduced herein can be implemented by, for
example, programmable circuitry (e.g., one or more microprocessors)
programmed with software and/or firmware, or entirely in
special-purpose hardwired circuitry, or in a combination of such
forms. Special-purpose hardwired circuitry may be in the form of,
for example, one or more application-specific integrated circuits
(ASICs), programmable logic devices (PLDs), field-programmable gate
arrays (FPGAs), etc.
[0084] Software or firmware for use in implementing the techniques
introduced here may be stored on a machine-readable storage medium
and may be executed by one or more general-purpose or
special-purpose programmable microprocessors. A "machine-readable
storage medium," as the term is used herein, includes any mechanism
that can store information in a form accessible by a machine (a
machine may be, for example, a computer, network device, cellular
phone, personal digital assistant (PDA), manufacturing tool, any
device with one or more processors, etc.). For example, a
machine-accessible storage medium includes
recordable/non-recordable media (e.g., read-only memory (ROM);
random access memory (RAM); magnetic disk storage media; optical
storage media; flash memory devices; etc.), etc.
[0085] The term "logic," as used herein, can include, for example,
programmable circuitry programmed with specific software and/or
firmware, special-purpose hardwired circuitry, or a combination
thereof.
[0086] Some embodiments of the disclosure have other aspects,
elements, features, and steps in addition to or in place of what is
described above. These potential additions and replacements are
described throughout the rest of the specification. Reference in
this specification to "various embodiments," "certain embodiments,"
or "some embodiments" means that a particular feature, structure,
or characteristic described in connection with the embodiment is
included in at least one embodiment of the disclosure. These
embodiments, even alternative embodiments (e.g., referenced as
"other embodiments") are not mutually exclusive of other
embodiments. Moreover, various features are described which may be
exhibited by some embodiments and not by others. Similarly, various
requirements are described which may be requirements for some
embodiments but not other embodiments.
[0087] As discussed above, the disclosed technology can utilize
low-cost laser sensors to achieve wide angle FOV coverage, provide
high precision calibration between laser sensors, detect
disturbances to laser sensors, and generate combined point clouds
based on point data obtained at different times. While advantages
associated with certain embodiments of the technology have been
described in the context of those embodiments, other embodiments
may also exhibit such advantages, and not all embodiments need
necessarily exhibit such advantages to fall with within the scope
of the present technology. Accordingly, the present disclosure and
associated technology can encompass other embodiments not expressly
shown or described herein.
[0088] To the extent any materials incorporated herein conflict
with the present disclosure, the present disclosure controls.
* * * * *