Application No.: l l/987,44O
`Examiner: Bharadwaj, Kalpana
`Art Unit: 2129
`
`REMARKS
`
`In view of the following particulars, reconsideration of the pending application
`
`is respectfully requested.
`
`1.
`
`Rejection of claims 1—8 under 35 U.S.C. § 1031a) as being unpatentable over
`
`US.
`
`atent 6 301 572 Harrison
`
`in View of U.S.
`
`atent a
`
`lication
`
`2002/0054694 g Vachtsevanos!
`
`Reconsideration of this rejection is respectfully requested in view of the
`
`amendments to independent claims 1 and 5, from which the remaining claims in the
`
`rejection depend, and the following remarks which demonstrate that the proposed
`
`combination of Harrison and Vachtsevanos fails to render the pending claims prima
`
`facie obvious.
`
`In observing pending claim 1, the claim relates to a device for overall machine
`
`toll monitoring including the feature that the first neural network classifies into normal
`
`and abnormal categories, an amount of characteristics extracted from a target signal
`
`generated when the machine tool is racing prior to machining a workpiece.
`
`According to this arrangement, the present invention achieves the technical
`
`effects that the first neural network determines whether an attachment state of a tool is
`
`normal or not, and an unbalance in the attachment state of the tool.
`
`Regarding this feature, the Office action alleges that
`
`
`ITFigs. 2 to 4 do not disclose any preprocessing blocks of Harrison. The
`
`preprocessing is only explicit in Fig. 1. Further, contrary to what the applicant is
`
`arguing, both the preprocessing block(50) and the processing block(22) are input to the
`
`Fuzzy art Neural Network (34), via an input matrix. Therefore, the processing by the
`
`neural network of the signals before the machine toll is racing, does get input to the
`
`neural networkfor classification. (see page 8 in the Oflice A ction).l|
`
`However, in observing Harrison at column 4 lines 9—10, column 4 lines 59—60,
`
`column 5 lines 48—49 and column 6, lines 14—17, Harrison specifically states that Figs. 2
`
`-2-
`
`

`

`Application No.: l l/987,44O
`Examiner: Bharadwaj, Kalpana
`Art Unit: 2129
`
`to 4 are block diagrams illustrating details of the Input Preprocessing block of Fig.1.
`
`Therefore, it is respectfillly submitted that Figs. 2 to 4 are diagrams relating to the Input
`
`Preprocessing block.
`
`The Input Preprocessing, disclosed in detail in Figs. 2 to 4, relates to a process
`
`for preparing signals extracted from a Vibration sensor 12 and tachometer 46 for
`
`analysis by the Fuzzy ART Neural Network 34 before inputting the signals to the
`
`Fuzzy ART Neural Network and thereby provides the Fuzzy ART Neural Network
`
`with the accurate signals.
`
`In particular, such preparation processes are explained in
`
`detail at column 5, lines 38—54, col. 6, lines 14—22, and col. 7 line 66 to col. 8, line 7,
`
`as shown below.
`
`The ’I‘ttcl‘ttameters 46 measure the exact Speed that rotating
`filaments in the turbine are turning. Various lecttnoittgt‘es
`may 13::- used 111st Elites: tachotrtetmfs, in the CEKSC of an. ILAMESIIN}
`gas. turbine, the tachnmetcrs em: implenmnlcd using a mag—
`netic gmratunth sensor, and adjusted to yield digital pulses
`at a rate pr-t‘rpt‘n‘iirjmal it: the turbine rotational frequency.
`These 131113130 meter output Signals are converted to a digital
`number ill a.
`'Iktchmmelct‘ Prue: ~38ng Block 43. The digital.
`tacitt‘nneter values am: updated each time the input. digital
`signals are processed by the Fourier Tm nsform fit) of FIG. 2-.
`The Frequency Umnain Input Preprocessor Block .511 is
`illustrated in [46.231136inputvtbraiiun Signals and Minam-
`e-tet' signals are. prepared for analysis by the Fuzzy ARI.‘
`Neural Network 34 in the Frequency Dmnain tnpui Pfiipl‘fi—
`CESSIJE’. "the outputs of the Frequency Domain Input Pmpmr-
`ccasor are Cttmlmnems of the Input Matrix ~52 to the. neural
`network .
`
`(001.5, lines 38-54)
`
`

`

`Application No.: l l/987,44O
`Examiner: Bharadwaj, Kalpana
`Art Unit: 2129
`
`Refine thing, infiput is the Fast REYES-3T ’I‘ransthrm, FEW", fill,
`[he time. domain Sign].
`is windowed using ii Flat—’11)}:
`Winrhiw lllfl, see PEG. 3, in prim-Edi: acmra‘ie signal. :uupfl
`[Miles and minimize scallopihg Law.» in the spertrum. "I'lir:
`swirl-um Es enlivened from the samples number domain
`into real values by the: Absolute Value- black His-l. “liken the
`frequencymurmur: signs] is converted ism a.
`lsgariihmic
`representation 1136 and m arle unipolar positive in block .103.
`
`(col. 6, lines 14-22)
`
`A dirtieiasimaal shift is 2113:: created m, a stair-step manner,
`136 of FIG. 4, at periodic-
`l’requencies to help ensure.
`
`those frequencies in the
`separation of the ‘iiii‘rirmahoh at
`neural memory. The frequeney period Chosen for increment?
`ing the dimensional shifi was smalier than the smallest
`sideliaud sepa ration 1h :11 was errgjsecttisd in tha- Sp-ecu‘mh. ’lihis
`Siflif-Sittp 'l‘ra‘ietimmldimension shift allows the.
`,l‘“:.rz.z:v ART
`neural network. to 1mm 1hr:- inpu‘i iiifmriiiai‘imi very rapidly
`bLLJllSL. vigilance trials in the. melt-arm}: are reduced.
`
`(col. 7 line 66 to col. 8, line 7)
`
`That is, the Input Preprocessing disclosed in Harrison is clearly distinguished
`
`from pending claim 1, which requires “the first neural network classifying into normal
`
`and abnormal categories, an amount of characteristics extracted from a target signal
`
`generated when the machine tool is racing prior to machining a workpiece.”
`
`Vachtsevanos does not make up for the shortcomings of Harrison.
`
`Accordingly, the proposed combination of Harrison and Vachtsevanos fails to
`
`disclose, teach or suggest the limitation “an amount of characteristics extracted from the
`
`target signal generated when the machine tool
`
`is racing prior to machining the
`
`

`

`Application No.: l l/987,44O
`Examiner: Bharadwaj, Kalpana
`Art Unit: 2129
`
`workpiece is classified into normal and abnormal categories by the first neural
`
`network,” as recited in pending claim 1.
`
`Claims 2-4 are also considered to be patentable as containing all of the
`
`elements of claim 1, as well as for their respective individually recited features.
`
`In observing pending claim 5, the claim requires 1) a first deviation which is a
`
`normalized value of a magnitude of the difference vector between weight coefficients of
`
`neurons in an output layer included in the first neural network and the amounts of
`
`characteristics extracted by the first characteristics extracting unit, and a second
`
`deviation which is a normalized value of a magnitude of the difference vector between
`
`weight coefficients of neurons in an output layer included in the second neural network
`
`and the amounts of characteristics extracted by the second characteristics extracting
`
`unit, and 2) a determination unit determines that there exists a fault in the machine tool
`
`if one of the first and the second deviation is greater than a preset threshold.
`
`The Office action relies on Harrison (col. 10 lines 60-65) for the teaching of a
`
`determination unit determining that there exists a fault in the machine tool if one of the
`
`first and the second deviation is greater than a preset threshold. That is, the rejection
`
`alleges that the first and the second deviations correspond to the vibration disclosed
`
`in Harrison (col. 10, lines 60-65).
`
`However, as set forth above, claim 5 expressly defines that the first deviation
`
`is a normalized value of a magnitude of the difference vector between weight
`
`coefficients of neurons in an output layer included in the first neural network and the
`
`amounts of characteristics extracted by the first characteristics extracting unit, and the
`
`second deviation is a normalized value of a magnitude of the difference vector between
`
`weight coefficients of neurons in an output layer included in the second neural network
`
`and the amounts of characteristics extracted by the second characteristics extracting
`
`unit.
`
`It is noted that the vibration mentioned in Harrison is merely related to a
`
`mechanical Vibration, which is Clearly different from the Claimed normalized value.
`
`-5-
`
`

`

`Application No.: l l/987,44O
`Examiner: Bharadwaj, Kalpana
`Art Unit: 2129
`
`Vachtsevanos does not make up for the shortcomings of Harrison.
`
`Accordingly, the proposed combination of Harrison and Vachtsevanos fails to
`
`disclose, teach or suggest a first deviation which is a normalized value of a magnitude of
`
`the difference vector between weight coefficients of neurons in an output layer included
`
`in the first neural network and the amounts of characteristics extracted by the first
`
`characteristics extracting unit, and a second deviation which is a normalized value of a
`
`magnitude of the difference vector between weight coefficients of neurons in an output
`
`layer included in the second neural network and the amounts of characteristics extracted
`
`by the second characteristics extracting unit, as recited in pending claim 5.
`
`Claims 6-8 are also considered to be patentable as containing all of the
`
`elements of claim 5, as well as for their respective individually recited features.
`
`In View of these observations, it is respectfully submitted that the proposed
`
`combination of Harrison and Vachtsevanos fails to render the pending claims of this
`
`rejection prima facie obvious. Accordingly, withdrawal of this rejection is kindly
`
`requested.
`
`

`

`Application No.: l l/987,440
`Examiner: Bharadwaj, Kalpana
`Art Unit: 2129
`
`2.
`
`Conclusion
`
`As a result of the amendment to the claims, and further in view of the
`
`foregoing remarks, it is respectfully submitted that the application is in condition for
`
`allowance. Accordingly, it is respectfully requested that every pending claim in the
`
`present application be allowed and the application be passed to issue.
`
`If any issues remain that may be resolved by a telephone or facsimile
`
`communication with the applicants’ attorney, the examiner is invited to contact the
`
`undersigned at the numbers shown below.
`
`BACON & THOMAS, PLLC
`625 Slaters Lane, Fourth Floor
`
`Respectfully submitted,
`
`Alexandria, Virginia 22314—1176
`Phone: (703) 683-0500
`
`Facsimile: (703) 683-1080
`
`Date: April 19, 2011
`
`KEVIN D. WILLIAMS
`
`Attorney for Applicants
`Registration No. 63,716
`
`