Waymo LLC

Patent Trials and Appeals Board

Feb 18, 2021

2020005727 (P.T.A.B. Feb. 18, 2021)

UNITED STATES PATENT AND TRADEMARK OFFICE
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
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
15/847,390 12/19/2017 Jiajun Zhu XSDV 3.0-460
CONCON RE II
5238
146033 7590 02/18/2021
WAYMO LLC
BOTOS CHURCHILL IP LAW
430 Mountain Ave
Suite 401
New Providence, NJ 07974
EXAMINER
LILLIS, EILEEN DUNN
ART UNIT PAPER NUMBER
3993
NOTIFICATION DATE DELIVERY MODE
02/18/2021 ELECTRONIC
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.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address(es):
pto@bciplaw.com
waymo@bciplaw.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

Ex parte JIAJUN ZHU, DAVID I. FERGUSON, and DMITRI A. DOLGOV
____________

Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1
Technology Center 3900
____________

Before MICHELLE R. OSINSKI, JILL D. HILL, and
CYNTHIA L. MURPHY, Administrative Patent Judges.

OSINSKI, Administrative Patent Judge.
DECISION ON REQUEST FOR REHEARING

STATEMENT OF THE CASE
Appellant1 filed a request for rehearing under 37 C.F.R. § 41.52
(hereinafter “Request” or “Req. Reh’g”), dated December 30, 2020, of our
Decision on Appeal (hereinafter “Decision” or “Dec.”) dated November 4,
2020. In the Decision, we affirmed the Examiner’s decision to reject:
(i) claims 21–37 under 35 U.S.C. § 112, first paragraph, as failing to comply
with the written description requirement on a first asserted basis;

1 We use the term “Appellant” to refer to “applicant” as defined in 37 C.F.R.
§ 1.42. Appellant identifies the real party in interest as Waymo LLC.
Appeal Br. 2.
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

2
(ii) claims 21–26, 30, and 31 under 35 U.S.C. § 102(b) as anticipated by
Trepagnier; (iii) claims 21–26 and 30–34 under 35 U.S.C. § 103(a) as
obvious over Trepagnier and Dieterle; and (iv) claims 32–37 under
35 U.S.C. § 103(a) as obvious over Trepagnier and Svendsen. Dec. 38.2
Rehearing is limited to matters overlooked or misapprehended by the Board
in rendering the initial decision. See 37 C.F.R. § 41.52(a).

OPINION
Written Description
Appellant argues that “[b]oth the rejection and the Decision have
ignored the underlying presumption of an adequate written description” and
“both fail to rebut the presumption that the specification as filed is not, at a
minimum, ‘adequate’ with respect to the terms at issue.” Req. Reh’g 3. In
particular, Appellant points to the use of a “string-cit[ation]” in the Final
Office Action. Id. at 4. We are not persuaded of reversible error where the
Examiner does not solely rely on such a string citation of the claimed
features of accessing behavior data and filtering received sensor data in
making the rejection. See Final Act. 20–25.

2 In the Decision, we also reversed the Examiner’s decision to reject:
(i) claims 23, 24, 27–29, and 33–37 under 35 U.S.C. § 112, first paragraph,
as failing to comply with the written description requirement on an
additional asserted basis; (ii) claims 22, 29, and 32–37 under 35 U.S.C.
§ 112, second paragraph, as being indefinite; and (iii) claims 27–29 under
35 U.S.C. § 103(a) as obvious over Trepagnier and Foessel. Dec. 12–18, 38.
In addition, we did not reach the additional rejection of claims 32–37 under
35 U.S.C. § 102(b) as anticipated by Svendsen. Id.
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

3
Appellant also argues that “[i]t is unclear from the Decision whether
the Board is reversing or affirming the Examiner for the other three features”
of (1) autonomously navigating, (2) determining that another vehicle has
performed an action, and (3) altering the control strategy. Req. Reh’g 4.
Because we determined there to be a lack of algorithm or flow of logic
described in Appellant’s Specification sufficient to guide one of ordinary
skill in the art to program a computer to perform the steps of (1) accessing,
by one or more computing devices, behavior data, and (2) filtering received
sensor data to identify actions of interest of another vehicle, we did not need
to rely on any other basis for rejecting claims 21–37 under 35 U.S.C. § 112,
first paragraph, as failing to comply with the written description
requirement.
Appellant argues that “[b]ecause the Board did not apply the
(erroneous) written description standard under which the Examiner rejected
all claims, the basis set forth in the Decision is an undesignated new ground
of rejection.” Req. Reh’g 5. Although the Final Office Action and Answer
did not explicitly use the terminology “2019 § 112 Guidance,” in our view,
the Examiner’s analysis shows that the Examiner is relying on such
Guidance. The 2019 § 112 Guidance states that “the specification must
provide a sufficient description of an invention, not an indication of a result
that one might achieve” and that “[t]he level of detail required to satisfy the
written description requirement varies depending on the nature and scope of
the claims and on the complexity and predictability of the relevant
technology.” Computer-Implemented Functional Claim Limitations for
Compliance with 35 U.S.C. 112, 84 Fed. Reg. 57, 61 (Jan. 7, 2019). The
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

4
2019 § 112 Guidance also indicates that “[i]f the specification does not
provide a disclosure of the computer and algorithm(s) in sufficient detail to
demonstrate to one of ordinary skill in the art that the inventor possessed the
invention that achieves the claimed result, a rejection under 35 U.S.C. 112(a)
for lack of written description must be made.” Id. at 62.
Similarly, the Examiner’s basis of rejection is that the Specification
does not describe the algorithms or flow of logic to perform the steps of
(1) accessing behavior data (2) filtering received sensor data, such that “[t]he
claims recite computer-implemented methods/systems, but the
[Specification] does not provide a description as to how the associated
manipulative steps would be effected.” Final Act. 3; see also id. (“[The
Specification] does not describe any algorithms or flow of logic to perform
many of the actions that are listed in [the flow chart of Figure 7 and that also
appear in the claims]” and instead “the [S]pecification indicates that the
listed actions are simply results that may be attributed to various components
associated with the computer system 110.”). To the extent that the Examiner
refers to the claimed steps “not appear[ing] to be universally accepted in the
autonomous vehicle control art such that the performance of those actions
using a computer would be self-explanatory” (Final Act. 20), we do not
understand such language to be an indication that the Examiner is employing
his or her own standard as asserted by Appellant. Rather, we understand
such a reference to be an indication of the Examiner’s consideration of the
complexity and predictability of the relevant technology, and consequently,
the level of detail that might be required to satisfy the written description
requirement in accordance with the 2019 § 112 Guidance.
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

5
Accessing Behavior Data
Appellant asserts that “[t]he Decision appears to misapprehend the
nature of the rejection” in that “[t]he rejection itself did not assert that it was
unclear what behavior data is.” Req. Reh’g 5. Appellant asserts that rather
“the [F]inal Office Action . . . admits . . . that ‘As explained in [the
Specification], the “vehicle 101” can access and compare the stored data to
determine that a detected object vehicle has specifically changed heading, or
more broadly, performed a behavior/action of interest (cols. 7–8, ll. 51–30;
Figs. 5, 6).’” Id. The Examiner’s rejection explains that:
the suggestion in [the Specification] that “instructions 132 may
allow for computer system 110 to identify when a detected
vehicle has performed one or more of the behaviors of interest”
is insufficient to describe or teach how the computer achieves the
result[] of accessing behavior data . . . . Given both the database
136 that includes a road graph, and the database 137 that includes
the detected position and movement associated with a vehicle,
simply “combining both sets of data” would not appear to
inherently result in the access of “behavior data” associated with
that detected vehicle. In this regard, the detected vehicle data is
local (i.e. relative to vehicle 101: cols. 7–8, ll. 51–2), but the map
data would appear to be global. [The Specification] does not
describe or teach how the local/global data would be reconciled
with “a set of actions or behaviors of interest” that may be
contained in database 138.
Final Act. 23–24 (citations omitted). In other words, the rejection
questions how relevant behavior data for a vehicle is ascertained
simply by the use of map data.
Appellant further argues that “the arguments presented on appeal
appear to have been overlooked and/or misapprehended, because the
Decision evaluates the term ‘behavior data’ in the context of what a human
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

6
driver might do.” Req. Reh’g 5 (citing Dec. 8–9). Appellant further argues
that “[a]t no point has Appellant ever asserted the ‘idea of the computer 110
storing a human driver’s knowledge about potential behaviors of nearby
objects for later access’” and “[f]or this reason alone, the rejection should be
reversed.” Id. Appellant continues that “[b]y interpreting a portion of the
claimed feature (the ‘behavior data’) to be ‘a human driver’s knowledge’,
the Board effectively, and improperly, has misconstrued the actual claim
features” and “the entire analysis and basis for the Decision is flawed.” Id.
The basis for the Decision is that the Specification fails to describe an
algorithm or flow of logic causing computer 110 to access behavior data
specifically for another vehicle, as opposed to behavior data for other types
of objects that may be encountered by an autonomous vehicle, as well as
how computer 110 would determine that the behavior of another vehicle is
relevant to the autonomous vehicle simply by the use of map data. Dec. 10;
see also Final Act. 23 (The “[Specification does indicate that ‘database 138
may include a set of actions or behaviors of interest,’” but “[m]issing here is
an explanation or teaching as to how the computer achieves the desired
result[] of accessing behavior data.”) and Ans. 23 (“Appellant does not . . .
point to any description in [the Specification] and explain how the computer
might actually achieve the ‘potential actions’ subject matter, or the ‘actions
of interest’ subject matter, relative to the ‘vehicle’ being claimed in
reissue.”) In short, the Specification does not describe adequately how
relevant behavior data for another vehicle is ascertained simply by the use of
map data.
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

7
Appellant argues that “despite the fact that the Decision expressly
admits the disclosure of ‘behaviors of . . . [a] vehicle changing lanes or
routes’, the Decision affirms the rejection because ‘[t]he Specification does
not describe an algorithm or flow of logic causing the computer 110 to
access behavior data for a vehicle, as opposed to behavior data for a
pedestrian or bicycle.’ (Decision at 10).” Req. Reh’g 6. “Appellant notes
that this determination appears to conflict with the conclusion reached in
counterpart application No. 15/847,064.” Id. According to Appellant, “the
new ground of rejection by the Board misapprehends the issue on appeal”
“[b]ecause the Board acknowledges that the [S]pecification discloses
‘behaviors of . . . [a] vehicle changing lanes or routes’ and the term at issue
is ‘accessing, by the one or more computing devices, behavior data
identifying one or more potential actions of the other vehicle that are to
result in a change in the control strategy.’” Id. at 7 (emphasis omitted).
Although we acknowledge the description in the Specification of another
vehicle changing lanes as a behavior data that may result in a change in the
control strategy of an autonomous vehicle, the problem identified in the
written description rejection is the lack of an explanation as to how the
computing device of the autonomous vehicle ascertains that (i) certain
actions relate to a vehicle, and not another type of object that may be
encountered by an autonomous vehicle, and (ii) another vehicle and its
behaviors are relevant to the autonomous vehicle simply by the use of map
data. See Final Act. 23; Ans. 23.
Appellant further argues that “[t]he assertion presented in the instant
Decision regarding vehicle versus pedestrian or bicycle behavior data is not
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

8
consistent with the 2019 § 112 guidance, which explains that” the
Specification need not “contain examples explicitly covering the full scope
of the claim language.” Req. Reh’g 6 (quoting 2019 § 112 Guidance, 84 Fed
Reg. 61). In our view, rather than requiring examples for all possible types
of other objects that may be encountered by an autonomous vehicle, we
understand the Examiner’s basis for the affirmance of the rejection to be a
lack of an algorithm or flow of logic as to how, the computing device would
ascertain that behavior is from another vehicle and would be relevant to the
autonomous vehicle so as to warrant causing a change in control strategy
simply by the use of map data.
We have considered the exemplary passages from the Specification
identified by Appellant as “explain[ing] in detail how the memory stores
behavior data, such as in a database” (Req. Reh’g 7), including, for example
that “the sensors may provide the location and shape information of objects
surrounding the vehicle to computer 110, which in turn may identify the
object as another vehicle” and “database 138 may include a set of actions or
behaviors of interest, such as the vehicle changing lanes or routes, and
instructions 132 may allow for computer system 110 to identify when a
detected vehicle has performed one or more of the behaviors of interest” and
“[i]n particular, computer system 110 may access the recorded position and
movement stored in database 137, as well as a road graph of the
environment stored in database 136” and “[b]y combining both sets of data,
computer system 110, may then determine when one or more of the key
behaviors have occurred.” Id. at 8 (quoting Spec. 7:1–4, 41–50). We have
also considered Appellant’s contention that “it should be readily apparent
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

9
that the term in dispute merely requires accessing (e.g., retrieving) behavior
data from storage (e.g., memory including a database), which can be done by
‘read[ing] with the aid of an electronic device, such as a hard-drive, memory
card, ROM, RAM, DVD or other optical disks, as well as other write-
capable and read-only memories.” Id. at 11. We, however, are not
persuaded that the basis of the Examiner’s rejection rests on a lack of the
understanding of storage of data in a memory, but rather on the lack of an
algorithm or flow of logic as to how relevant behavior data for an object is
ascertained. The exemplary identified passages to not explain adequately, in
our view, an algorithm or flow of logic as to how raw sensor data such as
location and shape information can identify an object as a vehicle in order to
assess its behavior and how the computer is programmed to determine that
data from another vehicle is relevant to the autonomous vehicle so as to
warrant causing a change in control strategy (e.g., how the computer is
programmed to combine position and movement data for other vehicles with
road graph environment data in order to know that the position and
movement data for other vehicles is relevant to the autonomous vehicle).
Appellant additionally argues that “the Examiner and the Board
appear to concede, in the context of the anticipation rejection of claim 21
based on [Trepagnier] that the prior art has sufficient written description for
accessing behavior data merely because “Trepagnier has a ‘VSO’ (Variable
Structure Observer) that ‘satisfies the “behavior data” requirement.’” Req.
Reh’g 12 (citing Final Act. 32–33; Dec. 22–23). In contrast to merely
considering the VSO, the Decision refers to Trepagnier’s disclosure of
creating an identifier and mathematical model of motion for an obstacle that
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

10
is used to predict the obstacle’s position and velocity in order to enable the
vehicle to safely stop or avoid obstacles even if the obstacle’s sensor data is
interrupted. Dec. 24. We are not persuaded that reliance on Trepagnier in
connection with the prior art rejections mandates reversal of the Examiner’s
decision that the Specification fails to comply with the written description
requirement.
Filtering
Appellant argues that “the Decision misapprehends the discussion of
the filtering operation as addressed on appeal” in that “the Decision (and the
Answer) ignores express discussions of filtering in the [S]pecification,
which were raised on appeal.” Req. Reh’g 12. Appellant continues that “by
not addressing the express teachings throughout Appellant’s disclosure, the
Board has impermissibly sustained the rejection of claim 32 for lack of
written description.” Id. at 13. We have considered the exemplary passages
from the Specification identified by Appellant, including, for example, that
“the data may be filtered to only include instances where a nearby vehicle
has performed an action of interest” and that “this determination is made by
comparing data for the nearby vehicle with map data accessed by the
autonomous vehicle” and that although “vehicle 101 will collect data
regarding vehicle 510’s change in heading, computer 110 will also
determine that the change in heading does not correspond to an action of
interest, as vehicle 510 merely travels along the same road graph element.”
Id. (quoting Spec. 2:9–15; 8:18–30) (emphasis omitted). The exemplary
identified passages to not explain adequately, in our view, an algorithm or
flow of logic as to how the computer is programmed to compare data for
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

11
other vehicles with map data in order to know that the data for other vehicles
is indicative of behavior data that does or does not correspond to an action of
interest. See, e.g., Ans. 27 (stating that the Specification “does not describe
or identify how a sensor data filter would be used for comparing two sets of
data that are based on different coordinate systems”).
Appellant also argues that the statement in the Decision that there is a
question raised “as to ‘how a sensor data filter would be used for comparing
two sets of data that are based on different coordinate systems’” “repeats a
misstatement of fact raised for the first time in the Answer.” Req. Reh’g 13.
Appellant also argues that the Examiner’s assertions “involve at least one
assertion irrelevant to the claim term and which does not properly relate to
the cited portion of the Specification.” Id. at 14.
Appellant states that “[f]irst, nothing in claim 32 refers in any way to
comparing two sets of data based on different coordinate systems.” Req.
Reh’g 13. Although claim 32 does not explicitly recite a comparison of two
sets of data based on different coordinate systems, such a comparison is part
of the filtering that is explicitly recited. For example, the Specification
states, as identified by Appellant, that “[v]ehicle 101 may also filter the data
collected for vehicle 510 so that it only contains instances where vehicle 510
has performed an action of interest” and “[w]hile vehicle 101 will collect
data regarding vehicle 510’s change in heading, computer 110 will also
determine that the change in heading does not correspond to an action of
interest, as vehicle 510 merely travels along the same road graph element.”
Spec. 8:18–28. In order to be able to make such an assessment (i.e., whether
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

12
a change in heading corresponds with a change in road element), there must
be some kind of comparison of collected data with map data.
Appellant also states that “[s]econd, the cited portion of the
[S]pecification at column 8, lines 9–17 occurs before the portion starting at
line 18 that states ‘Vehicle 101 may also filter the data collected for vehicle
510 so that it only contains instances where vehicle 510 has performed an
action of interest.’” Req. Reh’g 13–14 (emphasis omitted). Appellant
continues that “[r]eference to the preceding paragraph of the [S]pecification
in the Answer is a red herring that is irrelevant to the claimed feature” and
“nothing in the paragraph cited by the Examiner in the Answer even
mentions different coordinate systems or a comparison thereof.” Id. at 14.
As described above, the portion of the Specification beginning at column 8,
line 18 also relates to determining whether a change in heading involves a
change in road graph element as part of determining whether a change in
heading is an action of interest. This portion of the Specification also must
contemplate a comparison of collected data with map data.
Anticipation by Trepagnier
Appellant argues that “the [F]inal Office Action failed to set forth a
proper prima facie case for anticipation because the rejection fails to actually
address the features as claimed” and “[t]he Answer was likewise deficient.”
Req. Reh’g 15. We disagree that the Final Office Action and Answer failed
to address the language of the claim. See Final Act. 32–33 and Ans. 40–41
(summarizing the basis for the Examiner’s rejection of the claims and
including reference to the various claim limitations).
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

13
Appellant also argues that there “is a misapprehension in the Decision
about what Appellant explained on appeal regarding Trepagnier” when
considering the Decision’s statement that “The Examiner takes the position
that ‘the VSO satisfies the ‘behavior data’ requirement.” Req. Reh’g 15
(citing Dec. 22). Appellant continues that “one term at issue is ‘behavior
data identifying one or more potential actions of the other vehicle that are to
result in a change in the control strategy.” Id. Appellant indicates that
“[t]he rejection improperly stated that the ‘identifier’ for an obstacle is
‘behavior data’ as claimed.” Id. at 15–16 (citing Final Act. 32). Appellant
continues that “the identifier is ‘detected’” and “the identifier of Trepagnier
is derived from the detected information of the object by a sensor system,
which necessarily involves past information about the object.” Id. at 16.
Even if the identifier is based on identifying an obstacle’s position,
geometry, and velocity vector so as to produce a “state vector” that is
“constantly updated based on the incoming stream of sensor data”
(Trepagnier 15:53–62), the VSO algorithm also creates a mathematical
model of motion that produces predicted positions of the obstacle. The
mathematical model identifies potential actions of the obstacle (e.g., another
vehicle) in that it is identifying potential positions and velocities of the
obstacle over time. See Final Act. 32 (“With respect to path-planning, and
as explained in Trepagnier at column 16, lines 60–62: ‘After an initial path
is created that follows the center of the corridor, the path is checked against
the obstacle repository to determine if it is a safe path.’ Here, the ‘obstacle
repository’ is populated with the detected vehicle information (i.e. the
detected vehicle ‘identifier’ and the mathematical model of motion). Thus,
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

14
the broad claim language reciting ‘accessing’ and ‘determining’, is clearly
satisfied.”).
Appellant argues that the prediction of an obstacle’s position and
velocity in order to enable the vehicle to stop safely or avoid obstacles if
sensor data for the obstacle is interrupted “is not what is claimed” because
“[t]he behavior data itself is ‘behavior data identifying one or more potential
actions of the other vehicle that are to result in a change in the control
strategy.’” Req. Reh’g 16 (boldface omitted). Appellant continues that
“[e]either the identifier actually includes potential actions of another vehicle
or the VSO predicts future positions of the other vehicle based on the past
information about the object” and “[i]t would not make sense for both to be
true.” Id. We agree with Appellant that the VSO predicts future positions of
the other vehicle based on past information about the object; we, however,
do not agree that this results in a reversible error by the Examiner. The
Examiner’s rejection points to the mathematical model, populated in the
obstacle repository and constituting behavior data identifying potential
actions of the other vehicle.
Appellant maintains that “[a]s best understood, the identifier with a
‘mathematical model of motion’ are used by the VSO to derive a continued
future path based on the past information of the object” and “[a] benefit to
this approach, per Trepagnier, is that it ‘filters out fluctuations that can occur
in [the detected] sensor data,’” but “it is not what is claimed.” Id. at 17
(quoting Trepagnier 16:34–35). We are not persuaded that Trepagnier’s
determination of a continued future path of an object based on the past
information of the object, and the subsequent checking of a path for an
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

15
autonomous vehicle against the obstacle repository populated with the
detected vehicle identifier and mathematical model of motion of objects, as
well as the adjustment of the path so as to avoid all known obstacles, fails to
constitute the steps of: accessing behavior data identifying potential actions
that are to result in a change in the control strategy, determining that the
other vehicle has performed an action identified in the behavior data, and
altering the control strategy.
Obviousness over Trepagnier and Dieterle
Claims 21–26, 30, and 31
Appellant “submits that the same misapprehensions in the Decision
also apply to the obviousness rejection based on Trepagnier and Dieterle as
to claims 21–26 and 30–31” “[b]ecause all of the deficiencies of Trepagnier
. . . regarding the anticipation rejection apply equally to the obviousness
rejections.” Req. Reh’g 17. We have addressed such an argument in the
previous section relating to Trepagnier.
Claims 32–34
Appellant argues that the Decision’s upholding of the rejection of
claim 32 based on Trepagnier’s disclosure of filtering at column 16, lines
32–37 and column 5, lines 37–48 “is a misapprehension of the reference and
the claimed feature” because “[f]iltering of erroneous received sensor data
fluctuations is not what is claimed.” Req. Reh’g 17–18. Appellant argues
that “[n]either the discussion at column 16 nor the Kalman filter discussion
at column 5 of Trepagnier disclose filtering in the manner claimed,” which
relates to “filtering collected data . . . [so as] to identify ‘instances’ where a
‘vehicle has performed an action of interest’ in contrast to other actions, not
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

16
of interest.” Id. at 18. Appellant continues that “[b]y not considering the
disclosure of filtering in Appellant’s [S]pecification and as set forth on
Appeal, the Decision erroneously affirms the rejection in view of
Trepagnier.” Id.
Appellant, however, has not pointed out where any such argument
was made previously in the briefing, and we have not found any such
previous argument in our review of the record. That is, we do not see where
Appellant argued in the Appeal Brief or Reply Brief that the Examiner’s
reliance on Trepagnier for the claimed filtering step was not supported by a
preponderance of the evidence and/or that the Examiner did not properly
consider the disclosure of filtering in Appellant’s Specification when
determining that Trepagnier rendered obvious the claimed filtering step. See
Appeal Br. 42–43; Reply Br. 22–24. The Board could not have overlooked
or misapprehended arguments that were not before it.
Appellant also asserts that the Decision’s determination “‘that a
rationale having rational technical underpinning has been articulated’ for
combining Dieterle with Trepagnier” is “based on a misapprehension.” Req.
Reh’g 18 (quoting Dec. 30). More particularly, Appellant asserts that the
feature that is “the basis for one of ordinary skill in the art to modify
Trepagnier with Dieterle”—the absence of map information—“is not present
in any of the claims on appeal in this case.” Id.
Although in the claimed invention the step of determining whether the
vehicle cannot currently operate in a fully autonomous mode based on the
identified actions of interest of another vehicle relates to “whether there are
any obstacles or objects in the immediate vicinity which may prohibit or
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

17
reduce the ability of the vehicle to avoid a collision” (Spec. 9:57–60), as
opposed to the step in Dieterle that relates to whether there is a lack of
“infrastructure data” (e.g., lane data) (Dieterle 6:28–32), we are not
persuaded that the Examiner’s articulated basis for combining the teachings
of Trepagnier and Dieterle lacks rational underpinnings. One of ordinary
skill in the art may be prompted to do what the inventor has done, but for a
different purpose or to solve a different problem. It is not necessary that the
prior art suggest the modification to achieve the same advantage or result
discovered by an applicant. See, e.g., In re Kemps, 97 F.3d 1427, 1430 (Fed.
Cir. 1996) (citing In re Dillon, 919 F.2d 688, 693 (Fed. Cir. 1990) (en banc)
(“Although the motivation to combine here differs from that of the applicant,
the motivation in the prior art to combine the references does not have to be
identical to that of the applicant to establish obviousness.”)).
Here, Trepagnier teaches an autonomous vehicle control system and
recognizes that “[o]ne of the challenges for an autonomous vehicle . . . lies
in building a map of the world . . . and locating itself within that map” and
“[t]he data collected from the obstacle and lane detection sensors are
referenced to some absolute location in the world or some location relative
to the vehicle.” Trepagnier 10:1–8. Trepagnier continues that “a map of the
area surrounding the autonomous vehicle can be created, and from this map
the path of the autonomous vehicle can be planned.” Id. at 10:14–16.
Dieterle, also in the context of an autonomous vehicle control system,
recognizes that if “infrastructure data” (e.g., lane width data) is not available,
then “the lateral guidance system is switched off.” Dieterle 6:28–31.
Considering that Trepagnier recognizes the need for collecting data so as to
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

18
be able to locate the autonomously controlled vehicle, the Examiner’s
articulated reason for combining Dieterle’s teaching of switching off
guidance systems if there is insufficient data to build the map has rational
underpinnings. That is, if there is insufficient information (e.g., map
information and/or lane information) so as to be able to build a map and
thereby assess the actions of an obstacle or object in the immediate vicinity
of an autonomously controlled vehicle, there is a rational underpinning to
undertaking the step of determining whether the vehicle cannot currently
operate in a fully autonomous mode based on identified actions of interest of
another vehicle (e.g., movement of another vehicle).
Appellant argues that “[t]he Examiner’s assertion in the Answer that
‘one skilled in the art would recognize that the logical map could obviously
be used to also predict the movement of the detected obstacles in the
environment of the autonomous vehicle’ (Answer pp. 45–46) is simply
speculation that is not supported by the cited references or any other
information in the record.” Req. Reh’g 18. We are not persuaded by such
argument in that the passage in the Answer identified by Appellant relates to
an alternative rejection of independent claim 21 in which the Examiner relies
on Trepagnier’s “Mission Manager module,” rather than having anything to
do with the rejection of independent claim 32. Ans. 45–46.
Appellant also argues that relevant case law recognizes that the
“motivation question arises in the context of the general problem
confronting the inventor rather than the specific problem solved by the
invention.” Req. Reh’g 18 (boldface omitted) (quoting MPEP § 2144(IV)).
Appellant asserts that “[t]he Decision misapprehends this by finding that a
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

19
‘rational technical underpinning[] has been articulated.’” Id. (quoting
Dec. 30). Appellant continues that “[t]here is no ‘absence of map
information’ concern in the general problem confronting the inventors of the
claimed technology.” Id. Appellant then points to a portion of the
Specification, which states in part, that a “device [implementing the
disclosed system] may . . . determine when the surrounding vehicles have
performed one of several predefined actions by comparing the sensor data
with stored road graph data.” Id. at 18–19 (quoting Spec. 2:38–49).
Appellant “submits that because the Decision did not consider the teachings
of the [S]pecification when addressing whether the motivation was proper is
reversible error” and “[w]hen the full record on appeal is considered,
Appellant submits that the asserted ‘underpinning’ is totally irrelevant to the
‘general problem confronting the inventor’, and thus there is no proper
motivation to combine.” Id. at 19.
We disagree that the articulated reasoning or its underpinning is
“totally irrelevant” to the general problem confronting the inventor, as urged
by Appellant. Req. Reh’g 19. When considering a portion of the
Specification referencing the comparison of sensor data with stored road
graph data in order to determine when surrounding vehicles have performed
a predefined action (id. at 18–19 (citing Spec. 2:38–49)), the absence of road
graph data (or alternatively termed “infrastructure data” in Dieterle) appears
to be relevant to the problem of determining whether a surrounding vehicle
has performed a predefined action that necessitates modification to the
control strategy of the autonomous vehicle, which is a problem identified by
Appellant.
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

20
Obviousness over Trepagnier and Svendsen
Appellant submits that “[t]he deficiencies discussed above regarding
Trepagnier and claim 32 apply to this rejection as well” and “[t]hus, for
similar reasons, the Decision misapprehends the rejection of independent
claim 32 based on a combination of Trepagnier and Svendsen.” Req.
Reh’g 19. We have addressed such deficiencies in the previous sections
relating to Trepagnier.

CONCLUSION
We have carefully considered each and every argument raised by
Appellant. For the above reasons, we are unpersuaded that the Board
misapprehended or overlooked Appellant’s arguments or otherwise erred in
sustaining the rejections of claims 21–37 under 35 U.S.C. §§ 112, first
paragraph, 102(b), and 103(a). Accordingly, Appellant’s Request has been
granted to the extent that we have reconsidered our Decision in light of the
arguments in Appellant’s Request, but is denied with respect to our making
any modification to the Decision.
In summary:
Outcome of Decision on Rehearing
Claims
Rejected
35 U.S.C. § Reference(s)/Basis Denied Granted
21–37 112, first
paragraph
Written Description 21–37
21–26, 30,
31
102(b) Trepagnier 21–26, 30, 31
21–26,
30–34
103(a) Trepagnier, Dieterle 21–26, 30–34
32–37 103(a) Trepagnier, 32–37
Appeal 2020-005727
Application 15/847,390
Patent US 9,216,737 B1

21
Claims
Rejected
35 U.S.C. § Reference(s)/Basis Denied Granted
Svendsen
Overall
Outcome
21–37

Final Outcome of Appeal after Rehearing:
Claims
Rejected
35 U.S.C. § Reference(s)/Basis Affirmed Reversed
21–37 112, first
paragraph
Written Description 21–37
22, 29,
32–37
112, second
paragraph
Indefiniteness 22, 29,
32–37
21–26, 30,
31
102(b) Trepagnier 21–26, 30,
31

21–26,
30–34
103(a) Trepagnier, Dieterle 21–26, 30–
34

27–29 103(a) Trepagnier, Foessel 27–29
32–37 103(a) Trepagnier,
Svendsen
32–37
32–37 102(b) Svendsen3
Overall
Outcome
21–37

No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a). See 37 C.F.R.
§ 1.136(a)(1)(iv).
DENIED

3 As explained in the Decision (Dec. 38), we do not reach this rejection.