Ex Parte ZHU et al

Patent Trial and Appeal Board

May 9, 2018

13585432 (P.T.A.B. May. 9, 2018)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR
13/585,432 08/14/2012
145835 7590 05/09/2018
McDonnell Boehnen Hulbert & Bergh off LLP/W aymo
300 South Wacker Drive
Chicago, IL 60606
UNITED ST A TES OF AMERICA
JiajunZHU
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www .uspto.gov
ATTORNEY DOCKET NO. CONFIRMATION NO.
12-156 4822
EXAMINER
KING, RODNEY P
ART UNIT PAPER NUMBER
3665
MAILDATE DELIVERY MODE
05/09/2018 PAPER
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte JIAJUN ZHU and DAVID I. FERGUSON
Appeal2017-006440
Application 13/585,432
Technology Center 3600
Before BENJAMIN D. M. WOOD, LISA M. GUIJT, and
PAUL J. KORNICZKY, Administrative Patent Judges.
GUIJT, Administrative Patent Judge.
DECISION ON APPEAL
Appellants 1 appeal under 35 U.S.C. § 134(a) from the Examiner's
rejection2 of claims 1, 3-5, 7, 8, 10-12, 14, 15, and 17-19 under 35 U.S.C.
§ 103(a) as unpatentable over Bageshwar (US 2010/0280699 Al; published
Nov. 4, 2010) and Kulik (US 2011/0275408 Al; published Nov. 10, 2011).
We have jurisdiction under 35 U.S.C. § 6(b).
We affirm.
1 Appellants identify the real party in interest as Google Inc. Appeal Br. 1.
2 Appeal is taken from the Final Office Action dated June 6, 2016.
Appeal2017-006440
Application 13/585,432
STATEMENT OF THE CASE
Claims 1, 8, and 15 are the independent claims on appeal. Claim 1,
reproduced below, is exemplary of the subject matter on appeal, with
disputed limitations italicized for emphasis.
1. A method comprising:
receiving using a computer system in a vehicle, ground
truth data that relates to a current state of the vehicle in an
environment, wherein the vehicle is configured to operate in an
autonomous mode in which the computer system controls the
vehicle in the autonomous mode based on data obtained by a
sensor coupled to the vehicle, and wherein the sensor is a
RADAR unit, a LIDAR unit, or a camera;
obtaining, using the computer system in the vehicle,
perceived environment data that relates to the current state of the
vehicle in the environment as perceived by the sensor,
detecting, using the computer system in the vehicle, a
difference between the perceived environment data and the
ground truth data; and
adjusting, using the computer system in the vehicle, one or
more parameters of the sensor to reduce the difference between
the perceived environment data and the ground truth data,
wherein the one or more parameters include a latency of the
sensor, a position of the sensor, or an orientation of the sensor
relative to the vehicle.
ANALYSIS
Appellants argue independent claims 1, 8, and 15 as a group. Appeal
Br. 3-8; Reply Br. 1-6. We select claim 1 as representative, with claims 8
and 15 standing or falling with claim 1. See 37 C.F.R. § 41.37(c)(l)(iv).
Regarding independent claim 1, the Examiner finds, inter alia, that
Bageshwar discloses receiving, using a computer in a vehicle (i.e.,
processing functionality 102), ground truth data that relates to a current state
of the vehicle in an environment, as claimed. Final Act. 3 (citing Bageshwar
2
Appeal2017-006440
Application 13/585,432
i-fi-f 14, 16, 17, 20). Appellants do not dispute the Examiner's finding.
However, we note that the Specification provides a very broad description of
the claim term "ground truth data." In particular, the Specification discloses
that "[ s ]ensor parameters may be optimized by collecting sensor produced
data and comparing the collected data to known data or ground truth data."
Spec. i-f 16 (emphasis added). The Specification also discloses that
[i]n one example, the vehicle may receive ground truth data that
relates to a current state of the vehicle in an environment. For
example, the vehicle may be traveling down a road, with other
vehicles driving in front of it. The ground truth data may define,
for example, external driving conditions, a current state of the
vehicle, and a current status of the other vehicles traveling in front
of the vehicle. The external driving conditions may include a
weather indication, a position of an obstacle in the environment,
a position of a landmark in the environment, and/ or a terrain map
of the environment, for example. Other driving conditions could
. . . be included. The current status of the other vehicles may
include information such as . . . velocity or speed of the other
vehicles, and the heading of the other vehicles. Other types of
status information could also be received. For example, the
ground truth data may indicate that another vehicle is in front of
the vehicle heading straight on a two-lane, 10 mile road at a speed
of 20 miles-per-hour, and there is a left-tum slant at mile 5.
Id. i-f 19; see also id. i174; i-fi-159, 69-71 (disclosing that ground truth data is
data indicating the speed and/or actual location of a vehicle other than the
claimed vehicle). The Specification further discloses that "ground truth data
may include data regarding the sensors," for example, "indicating a location
and operating altitude of a camera ... [system]" (id. i174) and also that
"ground truth data may be obtained directly from the automobile," for
example, "one of the plurality of sensors may be a confirmed reliable data
source used to obtain the ground truth data" (id. i175). In fact, the
3
Appeal2017-006440
Application 13/585,432
Specification discloses that "ground truth data may take the form of any data
set ... received by the [automobile's] computer system ... to compare
and/or validate the integrity of any data that is obtained or collected by one
of the plurality of sensors of the automobile." Id. i-f 7 4 (emphasis added).
Thus, we construe the claim term "ground truth data" to mean any data that
is from a reliable data source and representative of an actual state of an
environment (i.e., ground), consistent with the Specification. Cf claim 1,
supra (further limiting "ground truth data" to data "that relates to a current
state of the vehicle in an environment").
In addition, the Examiner finds that Bageshwar discloses a vehicle
configured to operate in an autonomous mode based on data obtained by a
sensor (including a RADAR or LIDAR unit, or a camera), and also
obtaining, using the vehicle's computer system, perceived environment data
perceived from the sensor, as claimed. Final Act. 3 (citing, e.g., Bageshwar
i-f 14 (disclosing that autonomous vehicle 100 includes images sensors 104,
for example, "cameras," which "provide data regarding landmarks in the
environment around the vehicle 100 to processing functionality 102"), Fig.
IA). The Examiner further finds that Bageshwar discloses detecting, using
the vehicle's computer system, a difference between perceived environment
data and ground truth data, because Bageshwar discloses comparing an
estimated position to a measured position. Id. at 4 (citing Bageshwar i-fi-1 18,
20).
The Examiner determines that Bageshwar does not specifically
disclose adjusting, using the vehicle's computer system, a parameter of the
sensor (including latency, position, or orientation relative to the vehicle) to
reduce a difference between perceived environment and ground truth data, as
4
Appeal2017-006440
Application 13/585,432
claimed, and relies on Kulik for disclosing that, similar to Bageshwar,
"orientation processor 250 can determine the difference between measured
and calculated parameters." Final Act. 4 (citing Kulik i-fi-160, 61); see also
id. i160 ("orientation processor 250 ... can save the differences as
correction or calibration factors to the orientation sensor measurements").
The Examiner further relies on Kulik for teaching the desirability of
calibrating sensors, as a "well known ... valuable engineering practice."
Ans. 3 (citing Kulik i13). The Examiner reasons that it would have been
obvious "to modify the teachings of Bageshwar ... to include the features of
Kulik ... to determine sensor correction" (Final Act. 4), and more
specifically, "to modify the sensor ... of Bageshwar ... to include the
calibration/correction features of Kuli[k] ... to provide the user with the
most accurate sensor readings of vehicle conditions deemed as valuable
information for the user to make appropriate counteractions in response to
the sensed information" (Ans. 4).
Appellants argue that the Examiner erred because "Bageshwar in view
of Kulik does not teach 'adjusting ... one or more parameters of the sensor'
wherein 'the sensor is a RADAR unit, a LIDAR unit, or a camera." Appeal
Br. 4. In support, Appellants submit that Bageshwar' s "image sensors 104
are not the sensors the need correction in Bageshwar," but rather, "what
needs correction in Bageshwar is the 'kinematic state' of the vehicle." Id. at
5 (citing Bageshwar i-f 17); see also Reply Br. 1-2. Appellants further
submit that, in Bageshwar, "image sensors 104 provide the 'ground truth
data' (measurements of the positions of landmarks in the vehicle's frame)
that is used to correct errors" and thus, "image sensors 104 do not need
correction." Id. at 6. Appellants conclude that "because the image sensors
5
Appeal2017-006440
Application 13/585,432
104 provide the 'ground truth data' that is used to correct errors, Bageshwar
teaches away from adjusting one or more parameters of the image sensors
104." Id.
We are not persuaded by Appellants' argument, which fails to address
the rejection as articulated by the Examiner. The Examiner proposes
implementing a calibration step within Bageshwar' s method, whereby
sensors 104 (i.e., a camera) are calibrated, as taught by Kulik, regardless of
whether the measurements from Bageshwar' s sensors 104 are further used to
perform error corrections. Appellants have not provided, nor can we find,
support in Bageshwar that discourages calibration of sensors 104 prior to
implementing the processing functionality to further calculate errors. Prior
art does not teach away from claimed subject matter merely by disclosing a
different solution to a similar problem unless the prior art also criticizes,
discredits, or otherwise discourages the solution claimed. See In re Fulton,
391 F.3d 1195, 1201 (Fed. Cir. 2004).
We also disagree with Appellants' characterization of sensors 104
(i.e., a camera) as "ground truth sensors" during a calibration 3 step, which
by definition, compares the perceived environment data, as perceived by
sensors 104, to a standard (or data that is from a reliable data source) to
ascertain a deviation-although the same sensor (i.e., a camera) may provide
ground truth data after calibration (i.e., after the sensor is confirmed to
provide reliable data of an actual state of the vehicle in the environment). In
3 An ordinary meaning of the term "calibrate" is "to standardize (as a
measuring instrument) by determining the deviation from standard esp. so as
to ascertain the proper correction factors." WEBSTER'S THIRD NEW INT'L
DICTIONARY 316 (1993).
6
Appeal2017-006440
Application 13/585,432
other words, during Kulik's calibration step, Bageshwar's sensors 104
provide perceived data that is compared against a standard (i.e., ground truth
data).
Appellants also argue that "Kulik does not provide any motivation to
correct the image sensors 104 of Bageshwar." Appeal Br. 6. In support,
Appellants contend that Kulik's teachings regarding calibration or correction
of orientation sensors do not correspond to the claimed sensor (i.e., a
camera), but rather, to accelerometers, gyroscopes, and compasses. Id. at 6-
7; see also Reply Br. 2-3. Again, however, Appellants' argument does not
address the rejection as articulated by the Examiner, which relies on Kulik's
teaching regarding the desirability of calibrating and correcting sensors, as
applied to Bageshwar' s sensors 104, which expressly includes a camera as a
sensor, as claimed.
Appellants, for the first time in the Reply Brief, more specifically
argue that "Bageshwar in view of Kulik does not teach 'adjusting ... one or
more parameters of the sensor ... wherein the one or more parameters
include a latency of the sensor, a position of the sensor, or an orientation of
the sensor relative to the vehicle." Reply Br. 1 (emphasis added). In
accordance with our rules, we do not consider such arguments unless good
cause is shown. See 37 C.F.R. § 41.41(b)(2). Notwithstanding, Appellants
do not provide any additional argument or evidence with respect to this
statement of the claim limitation, and as such, Appellants' argument does
not constitute separate argument for patentability of the claims pursuant to
37 C.F.R. § 41.37(c)(l)(iv). See In re Lovin, 652 F.3d 1349, 1357 (Fed. Cir.
2011) (holding that the Board had reasonably interpreted 37 C.F.R.
7
Appeal2017-006440
Application 13/585,432
§ 41. 3 7 ( c) (1) (vii) as requiring "more substantive arguments in an appeal
brief than a mere recitation of the claim elements and a naked assertion that
the corresponding elements were not found in the prior art"). Moreover, as
evidenced by Sharma, 4 it is well known that position and orientation
(including tilt) are adjustable parameters ofBageshwar's sensors 104 (i.e.,
camera). See Sharma Abstract (referencing "vision sensor[] ... tilt
parameters"), i-f 3 (disclosing that "[ t ]raditional camera calibration
parameters consist of intrinsic and extrinsic parameters," wherein
"[ e ]xtrinsic parameters relate the camera to other world coordinate
systems"), i-f 4 (referencing "the camera's orientation and mounting
parameters").
Finally, Appellants argue that the Examiner's determination that
"'adding the feature of calibration/correction to the sensor is well known to
be a valuable engineering practice' is completely unsubstantiated" and
"cannot be used to support an obviousness rejection." Reply Br. 3--4 (citing
In re Zurko, 258 F.3d 1379, 1385 (Fed. Cir. 2001) (holding that as
assessment of basic knowledge and common sense that is not based on any
evidence in the record lacks substantial evidentiary support).
Unlike in Zurko, however, the Examiner provides support for the
determination that the calibration and correction of sensors is a well-known
engineering practice by quoting Kulik: "'even if the device has been
calibrated in the factory, additional calibration, user or automatic, is useful
and sometimes even unavoidable."' Ans. 3 (quoting Kulik i-f 3). Moreover,
4 US 2010/0235129 Al; published Sept. 16, 2010. Sharma is cited in the
record. See Appellants' Information Disclosure Statement filed Aug. 14,
2012.
8
Appeal2017-006440
Application 13/585,432
in further support of the Examiner's finding, as Official Notice of a fact
capable of instant and unquestionable demonstration as being well-known,
Sharma discloses that "[a]utonomous vehicles need accurate and reliable
obstacle detection solutions to aid in navigation and obstacle avoidance" and
"[t]ypically sensors such as vision sensors5 ... are utilized to detect
obstacles in an autonomous vehicle's path." Sharma ,-r 1. Sharma continues
that "[the combining of sensory data] can give inaccurate outputs if the
sensors are not accurately registered (calibrated)" and therefore, "[a]s a
common practice, registration is done." Id. ,-r 2 (emphasis added). Put
another way, Appellants' claim 1 generally recites the well-known method
of calibrating (by definition) the sensor of an autonomous vehicle by
adjusting a parameter of the sensor. The specific type of sensor, and its
measured (or perceived) data and parameter(s), as claimed, are also known,
as discussed supra. Thus, we are not apprised of error in the Examiner's
findings or reasoning.
Accordingly, we sustain the Examiner's rejection of independent
claim 1, and claims 8 and 15 fall with claim 1. Appellants chose not to
present separate arguments for the patentability of the dependent claims, and
therefore, for the same reasons supra, we also sustain the Examiner's
rejection of dependent claims 3-5, 7, 10-12, 14, and 17-19. Appeal Br. 8;
Reply Br. 7.
5 Sharma discloses that a camera is a vision sensor. See, e.g., Sharma ,-r 18
("Sensors 110 are mounted on the vehicle 100 . ... [S]ensors 110 consist of
visions sensors (for example, a monocular camera")).
9
Appeal2017-006440
Application 13/585,432
DECISION
The Examiner's decision rejecting claims 1, 3-5, 7, 8, 10-12, 14, 15,
and 17-19 is AFFIRMED.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(l )(iv).
AFFIRMED
10