Sylvania Electric Products, Inc. v. Brainerd, 369 F. Supp. 468 (D. Mass. 1974)

*469 Morris Relson, Darby & Darby, David R. Francescani, New York City, for plaintiff.

Robert B. Russell, Russell & Nields, Henry C. Nields, Boston, Mass., for defendant.

W. ARTHUR GARRITY, Jr., District Judge.

This case arises under the patent laws of the United States and especially those statutes, 35 U.S.C. §§ 101-103, which define the standard of patentable invention. Plaintiff, a large electronics corporation which has developed and marketed a system for automatically identifying railroad cars, seeks a declaration that its system does not infringe Patent No. 3,145,921 (hereinafter "Brainerd patent"), of which defendant is patentee, and that said patent is invalid. Defendant, an engineer at the Massachusetts Institute of Technology and long-time active railroad enthusiast, has counterclaimed for infringement,^[1] and by amendment allowed late in the pretrial period plaintiff has raised the defense of patent misuse to the infringement counterclaim. At trial, the misuse issue was severed pending adjudication of the questions of validity and infringement.

The lawsuit and the contested patent involve methods for automatically identifying objects. Such objects theoretically may be stationary or in motion, and may include railroad cars, buses, packages, and supermarket items. The result sought is the receipt of information which distinguishes one object from another by a device capable of discriminating on the basis of such information.

The Brainerd patent and the alleged infringing device, Sylvania's Kar Trak System, are both directed toward the automatic identification of railroad cars, generally termed "automatic car identification" or "A.C.I." For decades the railroads have been confronted with the difficult problem of tracking and coordinating the use of unmanned cars owned by different lines and hauled from point to point in trains often hundreds of cars long. Freight cars, for example, have no power source of their own, nor are they usually manned. Rather, until they become part of a train, they are inert objects with no capacity to change location or identify themselves. Long freight trains must be broken up and reassembled at various points, with the risk that an incorrectly identified car may be lost or misdirected. Although the problem has long been recognized, it remained unsolved until 1967 when the Association of American Railroads (AAR) formally recommended adoption of the Kar Trak system. Since 1967, *470 Kar Trak has enjoyed widespread acceptance and success in the industry.

Defendant's patent application was filed on July 2, 1959. Despite exhibits submitted by defendant in support of an earlier conception date and accompanying memoranda and argument, the court finds that defendant's claims were not fully disclosed prior to the date of filing and that no date of conception earlier than July 2, 1959 is warranted.

Reduced to simple terms, the Brainerd patent operates in the following manner. A device including both an "interrogating" unit and a "sensing" unit is positioned alongside a track. The "interrogating" unit includes a light source which emits a fan-shaped beam of light. Each railroad car to be identified is equipped with a label comprised of vertical strips of red, green and white retroreflective material in a picket fence arrangement. Pairs of red and green strips (or other sufficiently differently colored strips) are arranged to conform with a binary code and are separated by white strips which serve to advance a reading of a label but do not contain identification information.

As a moving train passes by the trackside device, the light beam, which is stationary, falls on each strip of the label sequentially. Reflected light from the label passes back to the "sensing" unit which is positioned "in close association" with the interrogating unit so that the returned light travels as closely as possible to the axis of the emitted light. It is characteristic of a retroreflective surface that the light from a specific source placed as close as possible to the receiving axis of the sensor will be reflected much more brightly than light from other sources and that unwanted light signals or "noise" are therefore less likely to be sensed. Scotchlite, a product of the 3M Company and a highly efficient retroreflector, is used in the Kar Trak system and, although not specifically mentioned in the Brainerd patent claims, it is identified in Brainerd's accompanying specifications.

Returned light bearing information from the color coded strips is "viewed" by the sensing unit through a light-limiting slit, one strip at a time. As the light passes through this slit, it falls on a dichroic mirror, which mirror functions to separate red light from green light. Red light passes through the mirror to a photocell which produces a "red signal"; green light is reflected from the mirror to a second photocell to produce a "green signal." These photocells, in turn, deliver electric currents to a data processing means which converts the information into intelligible form and transmits it.^[2] While this description omits certain peripheral features of the patent, including light-blocking shields and the possible use of a pulsed light source, such features are not essential to the operation of the patented system and, in the court's view, they would not suffice, standing alone, to make the device patentable.

The Brainerd patent combines various elements into one system. As previously noted, these elements include a light source, a retroreflective label with colored strips arranged in a coded pattern, a light-limiting slit, a dichroic mirror, photocells, and a data processing means. Whether such a combination is patentable depends upon analysis of the prior art and a determination of the relevant legal standard.

The most comprehensive piece of prior art in the field of automatic object identification *471 is the Bus Electronic Scanning Indicator (BESI) System designed for use on the London Transport bus routes by Alan Readman and T. E. Pick. Although modified at various times prior to defendant's conception date, BESI was fully operational at least as early as 1957. Like Brainerd, BESI utilized (1) a wayside device comprising a light source and a sensing means; (2) an "identification plate" made up of retroreflective material arranged in a coded pattern; (3) photocells with color-discriminating characteristics; and (4) a data processing means. Thus, in broadest outline, BESI clearly anticipated the Brainerd system.

Defendant argues, however, that Brainerd is distinguishable from BESI, and therefore patentable, in four critical aspects: (1) unlike Brainerd, BESI did not use Scotchlite; (2) although BESI used colors, it only employed two (red and white) whereas Brainerd employs three (red, green and white) while also utilizing the siding of a car as a fourth color (black) (3) BESI did not use a light-limiting slit; and (4) BESI did not use a dichroic mirror. In our opinion, none of these distinctions rises to a level of significance comparable to that of the underlying system concept which BESI teaches. In addition, we find that each of these elements was anticipated in the prior art. Scotchlite, like cats-eyes, is a retroreflective substance, the properties of which were widely known at all relevant times. Indeed, Pick and Readman had experimented with Scotchlite in the BESI system as early as 1956 but rejected its use in favor of cats-eyes, believing the latter to have higher reflecting qualities at the necessary distances. The 3M Company, Scotchlite's manufacturer, had promoted its use as a retroreflector as early as 1949 in a publication entitled "Reflective Characteristics of Scotchlite," and it had been used successfully by the Toni Company to sense and direct packages moving along a conveyor past a light source. The use of Scotchlite therefore was perfectly obvious to one skilled in the art and developing such a system in 1958 and 1959.^[3]

As for the use of four color states, the court finds that the Ritzerfeld Swiss Patent No. 291908, issued in 1953, provides for such use and clearly anticipates Brainerd. In addition, a paper discussing BESI and published by Pick and Readman in October 1958 clearly indicates the possible use of a third color in that system, thereby anticipating Brainerd's use of red, green and white. Finally, use of the light-limiting slit and dichroic mirror, the former being merely a sound engineering technique and the latter conceded by defendant to be an old element for separating colored light, are anticipated by the prior art Orthuber (U.S.Pat.No.2,899,132) and Millspaugh (U.S.Pat.No.2,797,256) patents respectively. Even without Orthuber and Millspaugh, their use would have been obvious to a skilled artisan in the field.

Thus, none of the elements combined in the Brainerd patent was unrevealed in the prior art, nor do any of these elements operate in the context of the Brainerd system in a novel or unexpected way. Defendant contends, however, that this fact is not dispositive of the patentability of the system and argues that, although the elements functioning alone were obvious, to combine them in this manner to produce this specific result was nonobvious. We do not agree. At the outset, it is clear that a patent monopoly is to be granted only where an applicant's invention matches "the general strictness with which the overall test [of novelty, utility and nonobviousness] is to be applied." Graham v. John Deere Co., 1966, 383 U.S. 1, 19, 86 S. Ct. 684, 694, 15 L. Ed. 2d 545. In *472 addition, where a combination patent is in dispute, a court must scrutinize it "with a care proportioned to the difficulty and improbability of finding invention in an assembly of old elements." Great Atlantic & Pacific Tea Co. v. Supermarket Equipment Corp., 1950, 340 U.S. 147, 152, 71 S. Ct. 127, 130, 95 L. Ed. 162; see Bartelt Engineering Co. v. Pneumatic Scale Corp., D.Mass., 1969, 298 F. Supp. 649, 659.

The test of the patentability of a combination has been variously stated. In Great Atlantic & Pacific Tea Co. v. Supermarket Equipment Corp., supra, the Supreme Court stated that such a device is patentable "only when the whole in some way exceeds the sum of its parts" and added that a patent should not be held valid absent a "finding that old elements which made up this device perform any additional or different function in the combination than they perform out of it." Id. 340 U.S. at 152, 71 S. Ct. at 130. Applying this test in Griffith Rubber Mills v. Hoffar, 9 Cir., 1963, 313 F.2d 1, the court held that "the old elements, including the known material in the new use, must perform additional and different functions in the combination than out of it; the results must be unusual and surprising more must be derived from the combination than that which might be reasonably expected as the sum of the old ideas drawn from the public domain." Id. at 4. Many other cases adopt the same or similar language in weighing the validity of a combination patent. See, e. g., Bobertz v. General Motors Corp., 6 Cir., 1955, 228 F.2d 94, 98; Knickerbocker Plastic Co. v. Allied Molding Corp., 2 Cir., 1950, 184 F.2d 652, 654; Soundscriber Corporation v. United States, 1966, 360 F.2d 954, 960, 175 Ct.Cl. 644; Rosenberg v. Standard Food Products Corp., E.D.N.Y., 1971, 331 F. Supp. 1065.

Even assuming that Brainerd's device, in the absence of the Kar Trak system, would have satisfied the railroad industry's need for A.C.I. and would have enjoyed commercial success, we cannot find it other than an obvious assemblage of elements arranged in logical sequence, each functioning in its time-proven manner and the whole merely equalling the sum of the individual functions. Here, unlike certain chemical combinations or perhaps a particular machine, the elements do not coact in a novel or unusual manner to produce an additional unexpected function. One element functions to emit light, one to reflect light, one to sense light, and one to convert and transmit the sensed light in intelligible form; together, however, they constitute no new function but merely produce a desired result. Thus, in applying the restrictive standard announced in Great Atlantic & Pacific Tea Co., supra, we find that the Brainerd patent is obvious not only in light of the specific prior art related to object identification but also in light of the known functions of its component elements and the absence of any unusual or additional result derived from their combination, and hold therefore that it is invalid.

Because a decision on infringement as well as validity is in the interest of judicial economy, we now turn to that issue. Among various arguments advanced by plaintiff in support of a finding of noninfringement, that which identifies those components shared by the Brainerd patent and Kar Trak as having been drawn from the prior art is rejected, despite our finding that the Brainerd system is in fact anticipated in the prior art. To do otherwise would seem to assume that which must be ascertained. Moreover, consideration of questions of infringement must be premised upon an assumption of validity, viz., that the Brainerd patent constitutes a non-obvious and patentable combination and ordering of prior art elements.

Advancing on the assumption that Brainerd's system is a valid, ingenious and patentable arrangement of old elements in a non-obvious, novel and functional sequence, does Kar Trak infringe? *473 In roughest outline, both systems operate in an identical manner. Both reflect an emitted beam of light against a retroreflective coded surface. Both utilize a light-sensitive means and data-processing means to "read" the returned light and convert it to intelligible form. Both employ retroreflective labels comprised of colored stripes arranged in meaningful coded sequence, which stripes must be read sequentially if an accurate result is to be obtained.^[4]

There is, however, one fundamental difference between the two systems which when implemented in Kar Trak completely alters the structure and function of the whole system. Whereas Brainerd's system employs a stationary light source and is dependent upon the motion of a train to draw a label into and across its line of "vision," Kar Trak employs a timed rotating scanner by means of a mirrored wheel which draws the light beam vertically across the label, from top to bottom, without regard to the speed or direction of the car or whether the car is even moving. A number of results, all of which distinguish Kar Trak from Brainerd, flow from the use of this rotating scanner: (1) unlike Brainerd's horizontal "picket fence" label, the stripes in the Kar Trak label are parallel to the ground and arranged vertically as in a ladder; (2) whereas Brainerd is capable only of one scan per label, Kar Trak's rotating scanner permits multiple scans even of labels affixed to the fastest-moving trains, such extra scans necessarily enhancing the probability that a label will be correctly read; (3) Kar Trak reads labels on cars which are stationary before the scanner, or are moving in reverse; (4) the timed rate of scan allows Kar Trak to rely on a "time-base" in the operation of its data processor. This last fact has two further consequences. First, it eliminates the need for "control" stripes (in Brainerd's system, white stripes) to step the reading along from one identification stripe to the next. Instead, identification signals are received from red, blue, white or black stripes, and the decoder recognizes a pattern of signals within specific and predetermined time intervals. Second, by providing time limits within which a label must be read to be valid, said limits being based upon the speed of scan and the size of a standardized label, Kar Trak rejects as spurious signals which do not comply with these limits and which must, therefore, originate from a source other than a valid label. Brainerd is unable to distinguish between spurious and valid signals by means of the criterion of time.

Thus, by the addition of a rotating scan, Kar Trak is capable of using a ladder-like label with four, rather than two, identification colors. By successfully eliminating the need to use certain colored stripes solely for "control" purposes, Kar Trak is capable of packing more identification stripes, and hence more information, into less space. As a result of the special needs imposed by a time-base system, Kar Trak's data processor functions differently than Brainerd's.

On the basis of these findings, we rule that Kar Trak does not infringe upon Brainerd. Even assuming an identity of result between the two, there is clearly no identity of means and operation. See, e. g., Bartelt Engineering Co., Inc. v. Pneumatic Scale Corp., D. Mass., 1969, 298 F. Supp. 649, 660; Nickerson v. Bearfoot Sole Co., 6 Cir., 1962, 311 F.2d 858, 879. Although both systems reach a similar result through a similar sequence of events, and although the Brainerd claims, read broadly, do describe the outlines of the Kar Trak system, these similarities end far short of the essential mechanics of, and essential *474 differences between, the two systems. The Kar Trak system is "so far changed in principle" from Brainerd "that it performs the same or similar function in a substantially different way . . . ." Decca Limited v. United States, 1970, 420 F.2d 1010, 1014, 190 Ct.Cl. 454, citing Westinghouse v. Boyden Power Brake Co., 1898, 170 U.S. 537, 18 S. Ct. 707, 42 L. Ed. 1136.

Finally, the fact that the Brainerd system has never been constructed or tested has relevance to the finding of non-infringement, even beyond the fact that such "paper" patents must be strictly construed both as to validity and infringement. See Lockwood v. Langendorf United Bakeries, Inc., 9 Cir., 1963, 324 F.2d 82, 88. By providing for repeated scans and using a timing criterion to discriminate against interfering light, Kar Trak would appear less likely to misread, or fail to real, a label than Brainerd. In practice, Kar Trak has proven sufficiently accurate to satisfy the high demands of the railroad industry. Whether Brainerd is capable of similar accuracy is necessarily conjecture. Kar Trak underwent testing for seven years before final implementation. Brainerd has never been tested; indeed, a flaw in its data processing system was only detected by the plaintiff during the course of this litigation, fully twelve years after the patent application was first filed. Thus there is added force to plaintiff's contention that its distinctive time-based operation, absent in Brainerd, is a major and essential element in the development of any accurate identification system.

Accordingly, it is ordered that judgment be entered for the plaintiff Sylvania Electric Products, Inc. declaring that (a) the defendant Brainerd's patent No. 3,145,921 is invalid and (b) on the contrary assumption that defendant's patent No. 3,145,921 is valid, it is not infringed by the plaintiff's Kar Trak system. In light of these rulings, there is no occasion to give further consideration to the misuse issue.

[1] Certain counterclaims by defendant based on alleged confidential disclosure of the patent claims to plaintiff, and improper use by plaintiff of information derived thereby, were stricken from the case on motion for partial summary judgment. Other counterclaims alleging patent interference and antitrust violations were dropped by pretrial stipulation.

[2] At trial, plaintiff demonstrated, and defendant conceded, that the data processing means as drawn in the specification drawings is inoperable. There was testimony, however, that this aspect of the invention could have been rectified by a reasonably skilled artisan in the field. Furthermore, the data processing means is not advanced by defendant as an essential element in his patent. The court, therefore, has not heavily weighed the inoperability of this aspect of the patent in reaching its determination on validity.

[3] Defendant has also argued that the alignment of the sensor as close as possible to the axis of the light source is an essential aspect of his patent. However, such an alignment was widely known to be necessary to the effective use of Scotchlite or similar retroreflectors. Indeed, BESI reveals the combination of interrogating and sensing units "closely associated" in one wayside device.

[4] It is true, as plaintiff, has argued, that Kar Trak, unlike Brainerd, is not dependent on the placement of white, "control" stripes to advance a reading of a coded label. Nevertheless, plaintiff cannot dispute the fact that a correct sequential reading of "identification" stripes is prerequisite to correct car identification in either system.