`
`1
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` UNITED STATES PATENT AND TRADEMARK OFFICE
` BEFORE THE PATENT TRIAL AND APPEAL BOARD
`C.A. No. IPR2018-00289
`U.S. Patent 8,872,646
`* * * * * * * * * * * * * *
`APPLE INC.,
`*
`Petitioner,
`*
`v. *
`*
`UNILOC LUXEMBOURG S.A., *
`Patent Owner. *
`* * * * * * * * * * * * * *
` VIDEO-STREAMED TELECONFERENCED
`DEPOSITION OF JOSEPH A. PARADISO, PH.D., a
`witness called on behalf of the Patent
`Owner, pursuant to the United States Patent
`and Trademark Office Patent Trial and Appeal
`Board Rules of Procedure, before Jessica L.
`Williamson, Registered Merit Reporter,
`Certified Realtime Reporter, and Notary
`Public in and for the Commonwealth of
`Massachusetts, at the Offices of Hogan
`Lovells US LLP, 100 High Street, Boston,
`Massachusetts, on Thursday, July 19, 2018,
`commencing at 9:11 a.m.
`
`Complete Legal
`
`214-746-5400
`Apple v. Uniloc, IPR2017-00289
`Uniloc's Exhibit 2002
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`2
`
`A P P E A R A N C E S
`
`BY: JAMIE H. MCDOLE, ESQ.
` Haynes and Boone, LLP
` 2323 Victory Avenue
` Suite 700
` Dallas, Texas 75219
` (214) 651-5121
` jamie.mcdole@haynesboone.com
` -and-
`BY: THOMAS KELTON, ESQ.
` Haynes and Boone, LLP
` 2505 N. Plano Road
` Suite 4000
` Richardson, Texas 75082-4101
` (972) 739-6923
` thomas.kelton@hanesboone.com
` Counsel for the Petitioner and the
` Deponent
`
`1
`
`23
`
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`3
`
`A P P E A R A N C E S, Continued
`
`BY: JEFFREY HUANG, ESQ.
` Etheridge Law Group
` 2600 East Southlake Boulevard
` Suite 120-324
` Southlake, Texas 76092
` (408) 797-9059
` jeff@etheridgelaw.com
` Counsel for the Patent Owner
` (Present via teleconference.)
`
`ALSO PRESENT:
`
` Anthony Piccirilli, Videographer
`
`1
`
`23
`
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`4
`
` I N D E X
`DEPONENT PAGE
`JOSEPH A. PARADISO, PH.D.
`Examination by Mr. Huang 6
`
` E X H I B I T S
`
` (None.)
`
`1
`2
`3
`4
`
`56
`
`78
`
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`5
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` P R O C E E D I N G S
` THE VIDEOGRAPHER: Good morning.
` We're now on the record. Please note that
` the microphones are sensitive and may pick
` up whispering and private conversations.
` Please turn off all cell phones or place
` them away from the microphones as they can
` interfere with the deposition audio.
` Recording will continue until all parties
` agree to go off the record.
` My name is Anthony Piccirilli
` representing Complete Legal. The date today
` is July 19th, 2018, and the time is
` approximately 9:11 a.m. This deposition is
` being held at Hogan & Lovells located at 100
` High Street in Boston, Massachusetts. The
` caption of this case is Apple Incorporated
` versus Uniloc Luxembourg S.A. The name of
` the witness is Joseph A. Paradiso. At this
` time the attorneys present in the room and
` everyone attending remotely will identify
` themselves and the parties they represent,
` after which our court reporter, Jessica
` Williamson, representing Complete Legal will
` swear in the witness, and we can proceed.
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`6
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` MR. MCDOLE: Jeffrey, do you want
` to go first?
` MR. HUANG: Sure. This is Jeffrey
` Huang with Ethridge Law Group on behalf of
` the patent owner.
` MR. MCDOLE: Jamie McDole from
` Hanes and Boone on behalf of Apple and the
` witness.
` MR. KELTON: Thomas Kelton, Hanes
` and Boone, on behalf of Apple and the
` witness.
`
` JOSEPH A. PARADISO, PH.D.,
` a witness called on behalf of the Patent
` Owner, having satisfactorily been identified
` by the production of a driver's license, and
` being first duly sworn, was deposed and
` testified as follows:
`
` DIRECT EXAMINATION
`
` BY MR. HUANG:
`Q. Okay. Hi. Good morning. Let's start with
` this: Is there any reason you can't give us
` your full and truthful testimony this
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`7
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` morning?
`A. No.
`Q. And if at any time point during this
` deposition you don't understand a question,
` would you please let me know?
`A. Definitely.
`Q. And can we agree that if you do provide an
` answer, that you have understood the
` question?
`A. Yes.
`Q. Okay. Great. Did you submit a declaration
` in this IPR, which is IPR2018-00289?
`A. Let me check.
` (Witness reviews documents.)
`A. Can you repeat the number?
`Q. Sure. It's IPR2018-00289.
`A. That number does not appear in my
` declaration, I don't believe, but I did
` submit a declaration, yes.
`Q. Okay. So you have your declaration in front
` of you --
`A. Yes, sir, I do.
`Q. Does that appear to be your declaration you
` submitted in the IPR we're going to be
` talking about today?
`
`Complete Legal
`
`214-746-5400
`
`
`
`8
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`A. Yes, it does.
`Q. And can you please turn to the last page of
` that declaration, page 86?
`A. Yes.
`Q. And on the last page you see a signature?
` Is that your signature?
`A. I'm getting there in a second.
`Q. Okay. Sorry.
`A. This is my signature, yes.
`Q. Great. Did you review your declaration
` before signing it?
`A. Yes, I did.
`Q. And since you signed your declaration, were
` there any errors or omissions that you have
` identified in this declaration?
`A. There are no errors that I found on
` reviewing it. This appeared to be fine.
`Q. Great. And does your declaration contain
` all of your opinions for this IPR?
` MR. MCDOLE: Objection, form.
`A. This declaration reflects the opinions that
` I make with a prior art that I had found and
` put together on patent -- the '646 patent.
`Q. Okay. And so you do not have any other --
` or any other opinions not specifically in
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`9
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` your declaration regarding this IPR; is that
` correct?
`A. I may develop opinions in the future. These
` are the opinions I had at the time I wrote
` the declaration.
`Q. And as you sit here today, do you have any
` further opinions in this IPR other than
` that -- what's in your deposition (sic)?
`A. The declaration reflects my opinions.
` Again, the more I read, the more opinions I
` get.
`Q. Okay. Let's start with page 40 of your
` declaration, paragraph 81.
`A. Sure.
`Q. Please let me know when you're there and you
` have read the paragraph.
`A. I am there.
`Q. Okay. Do you see the paragraph, part of the
` first sentence, it says that "McMahan is
` directed to 'a method of enhancing the
` accuracy of a sensor,'" and "As taught by
` McMahan, this is done by 'determining
` whether the measure [of the sensor] falls
` outside of an acceptable range for the
` output of the sensor, and, when the measure
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`10
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` falls outside the acceptable range,
` modifying the measure of the output such
` that the measure falls within the acceptable
` range for the sensor.'"
` Do you see that?
`A. Yes, I see that.
`Q. What value specifically falls within the,
` quote-unquote, acceptable range?
`A. Can you repeat the question?
`Q. Yes. What value specifically falls within
` the, quote-unquote, acceptable range.
`A. Let me review what I wrote.
`Q. Sure.
` (Witness reviews document.)
`A. The accessible range is a signal within the
` normal range of the output of the sensor,
` quoting from McMahan, and I believe that.
`Q. And what would the specific values be for
` the normal range of that sensor?
`A. I cannot give you specific values because it
` depends very much on the sensor. You see
` how the signals are produced from the way
` the sensor's used. You establish a normal
` range or you look at the specifications of
` the sensor, and/or, and then you establish a
`
`Complete Legal
`
`214-746-5400
`
`
`
`11
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` normal range. There are many ways --
`Q. How --
`A. -- to do this.
`Q. Sorry. Go ahead.
`A. There are many ways to do that.
`Q. Okay. What -- for example, what are some of
` the many ways to establish the acceptable
` range for a sensor?
`A. You run tests. You see what their output
` is, and you set your range based on what you
` determine to be a normal output.
`Q. Is there any other way?
`A. That's the main way.
`Q. And when you say you run tests, what sort of
` tests would you run?
`A. You would use it in the context of how the
` device is used.
`Q. And in that case, the values -- the values
` derived in the context that the device is
` meant to be used, those would be all values
` within the acceptable range?
`A. You would determine by the normal operation
` of the device and the response of the sensor
` what the acceptable range is. Anything
` outside -- you would define your bounds of
`
`Complete Legal
`
`214-746-5400
`
`
`
`12
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` what is not acceptable based on that, on the
` response of the sensor, what the sensor
` would normally do in the context also of how
` the sensor would be used.
`Q. And so in this case, I believe in your
` declaration you're applying this -- the
` teachings of McMahan to a proposed
` Pasolini/Goldman combination; is that
` correct?
`A. Yes.
`Q. And so in the proposed Pasolini/Goldman
` combination, what specifically would be the
` acceptable range for that proposed device?
`A. Oh, I cannot tell you because it depends on
` what the sensor is, what the values are,
` what the units are, how it actually is being
` used. There many factors there. To give a
` specific number, I don't think that's
` relevant here.
`Q. Why don't you think it's relevant?
`A. Because it's ambiguous.
`Q. What is ambiguous?
`A. Asking me to give you a specific range.
`Q. I'm sorry. Do you think the question is
` ambiguous, or is something else ambiguous?
`
`Complete Legal
`
`214-746-5400
`
`
`
`13
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`A. What you're asking me does not have a clear
` answer.
`Q. Would it ever have a clear answer?
`A. It would have a clear answer if I had a
` device on the table, I was looking at the
` output, and I knew the units that this
` criteria would be applied to and I saw how
` it was used and I've reviewed the specs of
` the particular sensor. That's how we do --
` we do this all the time, by the way. It's a
` very common operation with the sensors and
` accelerometers. We were doing that in the
` early 2000s.
`Q. Okay. So for -- first of all, for the
` proposed Pasolini/Goldman combination, what
` type of sensor would be present in that
` combination?
`A. Oh, for the proposed combination it could be
` any accelerometer because Goldman can
` accommodate in principle. At least the
` ideas of Goldman can accommodate any
` accelerometer.
`Q. And so generally, then, what would be a
` typical acceptable range for an
` accelerometer such as the one in the
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`14
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` proposed Pasolini/Goldman combination?
`A. Oh, this depends very much on what you do
` with it.
`Q. Okay. What are the typical things you would
` do with such a device?
`A. Well, if you look at the patent
` specification, I guess it's if you're
` carrying a device or moving a device or
` jostling a device or you have it on the
` table or it's not moving. There's a range
` of things.
`Q. Okay. So I think let's start with you said
` it was -- you started with moving a device,
` jostling a device. Let's just start with
` moving a device. What would be a typical
` range of -- what would be generally an
` acceptable range when you're moving a
` device --
`A. If you're moving -- if you look at it in
` terms of G forces, if you're moving a
` device, you're in the realm of a G, maybe a
` couple of Gs. It depends. It depends on
` how you move it. If you have it on your
` shoe, you get impact forces that are quite a
` bit higher at heel strike. If you have it
`
`Complete Legal
`
`214-746-5400
`
`
`
`15
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` in your pocket, it will be a little bit
` less, but you will see as you're walking,
` you know, some variation of acceleration on
` the order of some Gs.
`Q. Okay. And so does that range generally from
` the lowest to the highest --
`A. Oh, it depends on --
` MR. MCDOLE: Objection, form.
` (Reporter interruption.)
`A. Can you repeat your question? That would
` probably be the best way.
`Q. Yeah. So the acceptable range is a range of
` values, I would imagine. So for the
` activity we're talking about, which is
` moving, for that activity, what would be
` generally the high and low values in the
` acceptable range?
`A. It depends on whether you look at absolute
` value or whether you look at a bipolar
` value. If -- you would basically see one G
` if it's at rest depending on the
` orientation. If you look at the dynamic
` acceleration, you can see changes that go up
` to a G or maybe a few Gs, on that order. If
` it's mounted on the shoe, you can see peaks
`
`Complete Legal
`
`214-746-5400
`
`
`
`16
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` that are much higher with the heel strike.
` It depends on where this device is located
` on the body and again how it's being used.
`Q. Okay. Great. So when you say given a
` value, which is a G or a few Gs or
` something, is that the upper value in the
` range?
`A. I would want to look at the specific data
` for this device in the context of use. If
` it was on something like a shoe, it would be
` higher. If it was on -- in the pocket,
` maybe it would be lower, but I would want to
` see this thing actually used and look at the
` data stream before I set this threshold.
`Q. I understand. But in your examples, when
` you did give an example of a G or several
` Gs, was that value what you intend to be the
` upper range of the acceptable range?
`A. It varies. It could be that. It could be.
`Q. Okay. And then what would be the lower
` range of the acceptable range?
`A. If it's a bipolar signal, it would be
` negative that, because it can go both ways.
` If it's an absolute value, it's only one
` value. If it's at rest, and if you subtract
`
`Complete Legal
`
`214-746-5400
`
`
`
`17
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` out gravity, it's zero. There are many
` answers to that question.
`Q. Okay. Great. And so going back to what you
` said before about testing to determine the
` acceptable range, I think you had mentioned
` you look at the actual output of the sensor
` data and the activity you are doing to
` determine what the acceptable range was; is
` that right? Or can you correct me if I'm
` wrong about that?
`A. You would do that. You would have to do it,
` or you would have an algorithm do it as
` well, but you would have to have a clear
` idea of what the acceptable range is.
`Q. And how do you get the clear idea of what
` the acceptable range is?
`A. By seeing. As I've said already -- I've
` answered the question -- by seeing the
` response of the device under normal use.
`Q. And so, I'm sorry, just for, again, my
` clarity, so the response of the device in
` normal use, that is -- those are values
` within the acceptable range; is that
` correct?
`A. Yes. Yes, those are values.
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`18
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`Q. Okay.
`A. If you see the device in normal use, those
` are within the acceptable range, except in
` the case -- even within normal use you can
` have a circuit malfunctioning. We've seen
` it. We've built these for years. And we
` can have these glitches that appear. So
` what you do is that you look at the signal.
` You know, you can have an algorithm look at
` it, typically we look at it, and we see
` where there may be spikes appearing in the
` signal. This is an example, an example of
` something out of the range. And we can
` identify the magnitude of the spike, the
` unphysical spike, which doesn't correspond
` to normal activity, versus the typical
` motion. And we set that threshold high
` enough so we don't exclude normal motion.
` We will exclude the spike. And this is
` common practice -- was common practice well
` before the date of that patent for people
` working with things like accelerometers.
`Q. Okay. Thank you. Now, can you please turn
` to page 41 and paragraph 82 of your
` declaration.
`
`Complete Legal
`
`214-746-5400
`
`
`
`19
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`A. Yes.
`Q. If you want to read the whole paragraph,
` please do so, and let me know when you're
` ready.
`A. Okay.
` (Witness reviews document.)
`A. Yes, I've finished.
`Q. Great. In the first sentence you said that
` a person of ordinary skill in the art "would
` have been motivated to" --
` (Reporter interruption.)
`Q. Okay. I'll start over. In the first
` sentence of the paragraph 82 it states a
` person of ordinary skill in the art "would
` have been motivated to combine the teachings
` of McMahan with the Pasolini and Goldman
` combination in order to enhance the accuracy
` of the accelerometer of the Pasolini and
` Goldman combination."
` Do you see that?
`A. I see it.
`Q. Can you tell me where in Pasolini or Goldman
` it states that the accuracy of the
` accelerometer is insufficient?
`A. If you look, in general, using
`
`Complete Legal
`
`214-746-5400
`
`
`
`20
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` accelerometers, any kind of distortion of
` the data is going to affect the accuracy.
` This is just a fact independent of Pasolini
` and Goldman. When you use a device, if
` there's distortion in the device, you have
` difficulty with the accuracy. And a POSITA
` would be aware of glitches because we saw
` them, you know, we encountered them. The
` early accelerometers were even worse. And
` we would always put in a glitch detector
` because these glitches would affect the
` accuracy adversely.
` So a POSITA would be aware of the
` glitches just by looking at the data and
` would be aware that there's a reason to get
` rid of them. And I think McMahan gives a
` great example of doing this. So a POSITA
` would want to definitely combine the essence
` of what McMahan says in glitch removal to
` the devices of the sort that Pasolini and
` Goldman discuss.
`Q. So is there any discussion of this in either
` Pasolini or Goldman?
`A. I would have to review Pasolini and Goldman
` to be sure. I do not believe that Pasolini
`
`Complete Legal
`
`214-746-5400
`
`
`
`21
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` or Goldman talk specifically about glitches
` as we call them, which is why we include
` McMahan. But anybody who worked with these,
` a POSITA working with systems like this at
` the time would be very well aware of
` glitches because we were plagued by them.
`Q. And so you're saying specifically -- and by
` the way, I belive you have Pasolini and
` Goldman there --
`A. Yes, I do.
`Q. -- possibly in front of you? So if you want
` to go ahead and review Pasolini and Goldman,
` please do so and take your time. But my
` question would be, so specifically for
` either Pasolini or Goldman are you saying
` that those devices were also plagued by
` glitches such that its accuracy would be
` insufficient?
`A. Let me review these --
`Q. Sure.
`A. -- quickly and get back to you.
`Q. Take your time.
`A. Sure.
` (Witness reviews documents.)
`A. Okay. I have reviewed these pieces of art.
`
`Complete Legal
`
`214-746-5400
`
`
`
`22
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` Can you repeat the question?
`Q. Yes. So specifically in Pasolini or Goldman
` can you show me where it speaks about these
` -- I'm sorry, let's start over.
` Specifically in either Pasolini or
` Goldman can you show me where it speaks
` about being plagued by glitches, so to
` speak, as you referred to it?
`A. Okay. If you look at Pasolini, he doesn't,
` at least on this quick reread that I made
` here, explicitly talk about glitches, but he
` does talk about effects on accuracy of the
` accelerometer. So he's already concerned
` with these devices not being accurate. So
` that's raised in Pasolini. He talks about
` offsets and how his baseline restore
` basically is removing them dynamically as
` well as compensating for gravity, but he
` talks about offsets and offset drift which
` are issues with accelerometers.
` If you look at Goldman, he doesn't
` talk so much about that. He does talking
` about calibrating it. He tells you how to
` remove the offsets and calibrate by flipping
` it. So you calibrate the range and the
`
`Complete Legal
`
`214-746-5400
`
`
`
`23
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` offset that way. But he also does talk
` about issues in reading these out. So if
` you look at the very end of it, when he
` talks about sampling rates, he actually says
` that -- at the very end "If radio packets
` are sent by reading the accelerometer, it
` might" --
` (Reporter interruption.)
` THE WITNESS: Sorry.
`A. "If radio packets are sent while reading the
` accelerometer, it might cause a sample
` period to be missed as it takes about 4
` milliseconds to send a packet. Sample
` readings may also be lost due to garbage
` collection, though since a generational
` collection scheme is used by the Squawk Java
` VM, the expected delays... are only about
` 20-30 milliseconds."
` So he's talking about sample readings
` being affected, the actually data coming
` from the accelerometer being essentially
` distorted, which can create a glitch.
` That's one way glitches can be produced.
` There are several ways you can get them, but
` that's one way. So both of them discuss or
`
`Complete Legal
`
`214-746-5400
`
`
`
`24
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` point to accuracy issues with
` accelerometers.
` And in Goldman as a POSITA reading
` that, I would be worrying about glitches.
` And as POSITA at the time, I saw them
` constantly. You hook up an accelerometer on
` a scope, you move it around, depending on
` how you build your circuit, what
` accelerometer you use, you see these
` glitches. You know, many sensors can suffer
` from these. Depends oftentimes on how
` they're applied, but it's very common. So
` we knew about them, and I can see the
` pointers in the literature.
`Q. Okay. So let's stay with Goldman because
` we're there. In the passage you read it
` talks about it might cause a sample period
` to be missed, and also sample readings can
` be lost due to garbage collection --
`A. Yes.
`Q. -- is that correct?
`A. Yes.
`Q. And so -- I'm sorry. And so, in your
` opinion, you're saying that a sample period
` being missed or lost is equivalent to a
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`25
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` glitch?
`A. Depends on how it's -- how it's actually
` playing out. If you get part of a sample
` coming in or, you know, there's something
` wrong with your conversion, if there's some
` process that interferes with storing the
` data, you can glitch. It's one of the ways
` it happens.
`Q. What do you mean by "part of a sample coming
` in"?
`A. Let's say bits get mixed up. You do an
` A-to-D conversion, there's some other task
` that happens, depending on how you write
` your code you could be making a -- not doing
` it the optimal way, it could be a property
` of the device, many ways this can happen,
` you can, you know, get a bit thrown,
` shouldn't be there. So you suddenly see a
` large spike is a good example.
`Q. So a bit that is thrown or something that is
` there that shouldn't be there --
`A. It's one example.
`Q. -- that is one example --
`A. One example.
`Q. -- that is an example of sample period being
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`26
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` missed or lost?
`A. A sample period being missed or lost you
` would see a discontinuity, but there could
` be related things happening in the software
` that could perhaps change the actual data
` byte. Add a bit, mess up the conversion,
` there's a problem with timing so you read a
` register too soon, this is not unusual
` sometimes, and you can get a glitch. There
` are many ways --
`Q. (Inaudible.)
`A. -- glitches happen. This is one way. And
` I've seen it. I've seen it coding up
` accelerometers. It's one way. Another way
` is a problem with the circuit. Another way
` is a problem with the accelerometer.
`Q. Okay. What --
`A. Could be a bad data link. You'd see it
` there as well.
`Q. Okay. So all those various other ways, are
` they talked about in Goldman?
`A. No. But you can see Goldman is concerned
` with the accuracy of the accelerometer
` values, and glitches are in that family. So
` as a POSITA I want to get the best, you
`
`Complete Legal
`
`214-746-5400
`
`
`
`27
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` know, sensor stream that I can. I'm going
` to be concerned that all of these factors --
` these guys are pointing to issues with
` accuracy. Goldman is pointing one way, the
` beginning of one way you can start to get
` glitches. So I don't -- I don't want them.
` Let's say I drop the accelerometer.
` That's another way I can get a glitch. I
` get a huge value. It's not relevant to
` this. It could skew all my data. I get a
` spike.
`Q. But that isn't talked about in Goldman now,
` correct?
`A. No.
`Q. And so the issue that you identified that we
` talked about in Goldman, which is the sample
` period being missed or lost, does Goldman
` state that because of that issue its
` accuracy would therefore be insufficient?
`A. I believe he implies that. That's something
` you want to avoid, yes.
`Q. Okay. Let's go back to Pasolini. Could you
` leading with the column and line numbers
` refer me to the sections in Pasolini that
` you were referring to?
`
`Complete Legal
`
`214-746-5400
`
`
`
`28
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`A. Let me see. If you look at column 5,
` paragraph -- lines 25 to 30, he talks about
` some of the contributions to accura --
` inaccuracy in accelerometers.
`Q. Okay. And so in that -- well, actually,
` let's start with that. Is there any other
` section that you've identified --
`A. There may be. I'd have to read it again,
` but this is what I saw when I read here.
`Q. Okay. And so in the section that you
` referred us to, it's referring to
` "fabrication offsets and deviations" --
`A. Uh-huh.
`Q. -- "that can be caused by the fatiguing of
` the materials" --
`A. Yes.
`Q. -- correct?
`A. Yes.
`Q. And what does that mean?
`A. That means that accelerometers of these
` sorts usually work with a cantilever beam
` done in a MEMS architecture, and that has a
` certain spring constant, certain mechanical
` properties, and he is saying that those can
` change over time, which is causing drift and
`
`Complete Legal
`
`214-746-5400
`
`
`
`29
`
`JOSEPH A. PARADISO, PH.D.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` the scale factor or, if not, offset of the
` accelerometer, both.
`Q. Okay. In the very next sentence after that
` it states "Subtraction of the static-
` acceleration signals" -- I'm going to skip
` the symbols -- "from the acceleration
` signals," again, some more symbols I'll skip
` -- "advantageously enables compensation of
` said offsets."
` Do you see that?
`A. Yes, I see that.
`Q. So doesn't Pasolini provide the solution to
` its fabrication offsets and mediation issues
` right there in the same paragraph?
`A. Yes. He suggests the solution there, but
` what I called attention to that paragraph
` for isn't to describe glitches per se but to
` talk about the general concern that POSITAs
` had at that time with the accuracy of these
` sensors.
`Q. Okay.
`A. So Pasolini is already flagging accuracy
` issues. He's not saying they're perfect.
`Q. And given the issue and proposed solution by
` Pasolini, does Pasolini suggest that its
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`30
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
` accuracy is still nonetheless insufficient?
`A. Insufficient for what?
`Q. For the purpose of the Pasolini device.
`A. That's not what he's suggesting there. He's
` basically saying that there are artifacts
` that come from the special MEMS
` accelerometers. This is one of them. And
` his subtraction, his technique that he is
` putting forth in this patent can compensate
` for those. It's not a common filter or any
` kind of advanced system for this, but it is
` a simple way to compensate for bias drift.
` But he is raising a flag of accuracy as
` being a potential issue in these devices.
` He's suggesting a solution for it.
`Q. Okay. Can you turn to page 42, paragraphs
` 84 and 85 --
`A. Sure.
`Q. -- of your declaration?
`A. Yes.
`Q. So here where you're talking about applying
` Mizell to, again, Paso -- the proposed
` Pasolini and Goldman combination; is that
` correct?
`A. That's correct.
`
`Complete Legal
`
`214-746-5400
`
`
`
`JOSEPH A. PARADISO, PH.D.
`
`31
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`Q. And here your talking about applying Mizell
` to the proposed Pasolini and Goldman
` combination "in order to smooth and reduce
` noise in order to obtain a more accurate
` representation of the gravity signal."
` Do you see that?
`A. I see it.
`Q. Do you see that?
`A. Yeah, I do see it.
`Q. So if -- actually, yeah, let me start from
` the top. So where is it in -- let's start
` with Pasolini where it talks about the need
` to sooth and reduce noise for an a