Wavetronix v. EIS Electronic Integrated Sys.

Citations: 573 F.3d 1343; 91 U.S.P.Q. 2d (BNA) 1468; 2009 U.S. App. LEXIS 16732; 2009 WL 2245213Docket: 2008-1129

Court: Court of Appeals for the Federal Circuit; July 29, 2009; Federal Appellate Court

Original Court Document: View Document

SummaryOriginalCytator Case TreatmentBackward CitatorAnalysis

EnglishEspañolSimplified EnglishEspañol Fácil

Wavetronix LLC filed a patent infringement lawsuit against EIS Electronic Integrated Systems regarding United States Patent No. 6,556,916, titled 'System and Method for Identification of Traffic Lane Positions.' Wavetronix claims that EIS’s Remote Traffic Microwave Sensor (RTMS X3) infringes on one independent claim and several dependent claims of the patent, either literally or under the doctrine of equivalents. The district court granted summary judgment of non-infringement to EIS, which Wavetronix is appealing. EIS has cross-appealed the dismissal of its counterclaims related to the patent's invalidity and unenforceability. The Federal Circuit, comprised of Circuit Judges Newman and Schall, and District Judge Patel, affirmed the district court's summary judgment of non-infringement. The patent pertains to technology designed for monitoring vehicular traffic, particularly the flow of vehicles in specific lanes to aid urban planning. Accurate traffic monitoring requires devices to identify lane boundaries, which is achieved by 'teaching' the device the locations of lanes based on vehicle movement. The ruling emphasizes that any alleged errors in the district court's opinion do not automatically necessitate reversal of the judgment.

Drivers are generally assumed to operate near the centers of lanes, which can help identify lane boundaries based on where vehicles tend not to drive. The patent in question relates to a method for teaching a traffic monitoring device the locations of traffic lanes through the detection of vehicle movements. Wavetronix developed a radar-based traffic monitoring system in 2000 to enhance automated lane definition, as previous systems required manual adjustments. The ’916 patent, filed on September 27, 2001, and issued on April 29, 2003, includes only claim 1, which details a method for defining traffic lanes using sensor-detected vehicle positions and a generated probability density function (PDF).

The specification includes preferred embodiments illustrated with histograms that represent vehicle traffic density across spatial distances from a sensor, indicating lane centers at the peaks and lane boundaries at the valleys. Statistical techniques are employed to determine lane boundaries for vehicle detection purposes.

The accused device, EIS’s RTMS X3, also detects and counts vehicles across multiple lanes. While lane locations can be determined manually or with the auto set-up feature, Wavetronix claims this feature infringes on their patent. The RTMS X3 transceiver has a field of view divided into thirty-two range slices, with each slice approximately the width of a typical traffic lane, but it requires matching these slices to actual lanes to accurately assign vehicles to specific lanes.

The human installer can manually define traffic lanes on a computer screen by aligning lane boundaries with displayed blips that represent passing vehicles, or utilize an auto set-up feature via the Setup Wizard. When the auto set-up is selected, the installer specifies the expected number of lanes, prompting the RTMS X3 to monitor traffic for one minute. This process requires a roadway with significant vehicle presence. During the minute, the transceiver emits radar signals that reflect off vehicles, generating data processed in several steps.

Every ten milliseconds, the system produces thirty-two values reflecting the strength of these signals across thirty-two range slices. Each value is compared to a detection threshold and stored in a Q-vector array, which records the difference between reflected strength and the threshold. A positive value in the Q-vector indicates vehicle presence in that slice within the last ten milliseconds. The Q-vector is updated continuously, and its contents are sent to the laptop, creating a REQ array that processes approximately 20% of the Q-vector data due to transmission limitations.

The Setup Wizard identifies 'local maxima' in the REQ array, focusing on the first local maximum closest to the sensor for each ten-millisecond interval. The NAMP array then compiles these first local maxima across the one-minute setup to generate a probability density function estimation, as referenced in the ’916 patent. This array records the first local maxima rather than individual vehicles, as a single vehicle typically produces multiple maxima as it passes through the sensor's field of vision. Illustrations in the document demonstrate this process with hypothetical vehicles traversing the RTMS X3's detection area.

The excerpt outlines the technical details regarding the NAMP Array Histogram used by Wavetronix. The court does not formally adopt the term 'histogram' to describe the NAMP array. The RTMX X3 sensor analyzes thirty-two range slices, resulting in thirty-two corresponding values after a one-minute setup period. A lane is defined when the count in a range slice exceeds approximately one percent of the total Q-vectors processed, requiring a minimum of about ten counts for identification. The system will repeat the detection phase if fewer lanes are detected than expected, and will eliminate the least significant lanes if more are detected than expected. Following this setup, the system is prepared to monitor vehicle traffic in the identified lanes.

In terms of procedural history, Wavetronix initiated the action on January 27, 2005. The district court granted summary judgment in favor of Wavetronix on EIS’s defenses of best mode and inequitable conduct in 2006. The court conducted extensive pre-trial hearings and ultimately granted summary judgment of non-infringement on September 21, 2007, without issuing a claim construction order. Both Wavetronix and EIS filed appeals, which are under the jurisdiction of 28 U.S.C. 1295(a)(1). The summary judgment was reviewed de novo, with emphasis on the criteria that no genuine issues of material fact exist and that the moving party is entitled to judgment as a matter of law, as outlined under Fed. R. Civ. P. 56.

The district court ruled that the EIS system does not infringe the ’916 patent, following a two-step infringement analysis: determining the meaning of disputed claim terms and comparing the accused device to those claims. The only contested term requiring construction is 'probability density function estimation.' Both parties requested the appellate court to construe this term rather than remand it for further construction. The appellate court accepted this invitation for three reasons: the district court had addressed claim construction, both parties agreed to the construction, and the record was sufficiently developed to allow for a decision without prejudice. 

The term is construed based on its ordinary and customary meaning as understood by someone skilled in the art at the time of the invention. The court will reference public sources, including the claims, specification, prosecution history, and relevant scientific evidence. There is consensus that 'probability density function estimation' involves estimating a 'probability density function,' which is a well-defined statistical term representing the probabilities of possible outcomes from an experiment.

EIS acknowledges that "PDF" is a recognized mathematical term, defined as a function whose integral over a region indicates the probability that a random variable falls within that region. Both parties agree on the qualifications of their experts regarding "ordinary skill in the art." The patent's language does not clearly define "PDFE," and neither party utilizes the patent’s prosecution history for clarification. Although the patent specification provides some guidance, it lacks a consistent, explicit definition of "PDFE." Terms such as "PDF as estimated," "PDF estimator," and "estimated PDFs" appear inconsistently throughout the specification. The relationship between "PDF" and "PDFE" is not fully clarified, but the specification outlines "PDF" more clearly, describing it graphically with peaks indicating lane centers and valleys representing lane boundaries. This graphical representation aligns with the mathematical definition of PDF, which models probabilities rather than actual data. The visual representation of a PDF on a histogram yields a smooth curve, in contrast to the disjointed appearance of actual data plots.

Collecting and plotting more data points results in a smoother line, but achieving perfect smoothness equivalent to a mathematical function requires infinite data, which is impractical in real-world scenarios like traffic monitoring. Thus, a significant number of data points can approximate the mathematical curve representing vehicle distribution. The primary distinction between a Probability Density Function (PDF) and a Probability Density Function Estimator (PDFE) is that a PDF is a perfectly smooth mathematical construct, while a PDFE is an approximation based on actual data, lacking perfect smoothness but reflecting real-world conditions. 

The specification illustrates various embodiments that utilize data points to estimate a hypothetical PDF. For instance, one embodiment utilizes normalized data presented as a percentage of the total, while others may use raw counts of vehicles. Figures 6, 7, and 8 of the specification illustrate different data representations, with Figure 7 indicating normalized data because its y-axis ranges from zero to one, whereas Figure 8 presents raw counts that require a calculation to compare against totals. 

The ability to compare data across a range is crucial for understanding both the presence and absence of vehicles, thereby aiding in defining traffic lanes based on the PDFE. There is no evidence indicating that the key distinguishing feature of a PDF is its ability to function over a range, as both PDF and PDFE share this characteristic. Consequently, the court construes a PDFE as "a finite data set large enough to approximate a function of a continuous variable whose integral over a region gives the probability that a random variable falls within the region." This definition applies to both generating a PDFE and defining traffic lanes from it as per the claims of the patent. 

The patent suggests that certain embodiments may involve only one range bin per lane, although using multiple bins would optimize lane definitions. However, if only one bin is used, the method does not appear to employ any technique other than comparing values across lanes to define them. The specification warns that range bins resulting in histograms without distinct peaks or valleys fail to provide useful insights into the beginnings and endings of vehicle lanes.

The teaching related to the patented system is crucial for assessing both literal infringement and infringement under the doctrine of equivalents. Literal infringement is determined by evaluating whether the accused device falls within the claims as construed, which is a factual question. The court will affirm a summary judgment of non-infringement only if there is no genuine dispute regarding the accused device's alignment with the claims when viewed favorably towards the non-movant.

Wavetronix claims that the RTMS X3 Setup Wizard utilizes NAMP array data to perform the step of 'defining traffic lanes' as outlined in claim 1. Wavetronix describes the Setup Wizard's operation as inferring the likelihood of vehicle travel in certain ranges based on first local maxima. EIS counters that the accused device does not engage in any probability analysis, asserting that the NAMP array data does not represent probabilities due to its coarse nature, which cannot reveal lane centers or boundaries. EIS claims its system uses a simpler method that relies on a common threshold for the NAMP data and requires human intervention to determine lane boundaries.

EIS argues that the NAMP array lacks the essential properties of the claimed PDFEs from the ’916 patent, failing to fulfill the requirement of defining traffic lanes. The proper construction of the PDFE claim demands detailed representations of vehicle frequencies across positions, which is necessary for accurately defining lane boundaries. The NAMP array's data processing does not facilitate comparisons across lanes, as lane definitions are made independently for each range slice without reference to others. Consequently, the characteristics of the NAMP data do not align with the construction of PDFE, rendering the accused system incapable of meeting the claimed limitations.

The highest value (x+2) and lowest value (x-1) in the NAMP array do not affect the definition of traffic lanes, as each lane corresponds to a single value rather than lane centers or boundaries. This indicates that the NAMP array data is insufficient to qualify as a PDFE under the ’916 patent. The RTMS X3's lane selection process is based on this data, leading to the conclusion that it does not infringe the 'defining traffic lanes' step outlined in claim 1 of the patent.

Regarding the doctrine of equivalents, Wavetronix argues that the district court overlooked potential infringement by equivalents, though EIS claims this argument was waived. Assuming it was not, the evaluation focuses on whether the accused product includes elements that are identical or equivalent to the patented claims. Infringement by equivalents is a factual matter, requiring that any differences between the accused product and the patent claims be deemed insubstantial by someone skilled in the art.

To establish equivalence, a plaintiff must demonstrate that the accused product performs the same function in a similar manner and yields similar results for each claim limitation. Courts cannot apply the doctrine of equivalents if it would negate a claim element entirely. The Setup Wizard does not align with the patented method's definition, as it merely verifies whether operator-defined range slices match actual lanes based on a specific threshold, rather than utilizing a comparative analysis of range slices.

The example of NAMP values illustrates that threshold values can categorize all values as lanes regardless of their peaks or valleys, which are crucial to the ’916 patent's method. The patent's claim requires lane boundaries to be defined through vehicle traffic analysis, revealing lane centers based on probability, contrary to the Setup Wizard's approach, which lacks meaningful analysis of where lanes start and end.

The system in question utilizes data solely to determine if a pre-selected range slice, representing the full width of a traffic lane, is within the trafficked roadway, without comparing data across different bins to identify lane boundaries. Even if the EIS system is argued to “define” traffic lanes, it does so distinctly from the method described in the ’916 patent, indicating that differences are not trivial. Wavetronix claims that the NAMP array generates a PDFE for lane definition; however, multiple arrays process data. Under patent law, two physical components can collectively fulfill an element of a claimed invention without negating any claim limitations. If data from a different array can be identified as a PDFE, it might suggest that the combination of that array with the NAMP array infringes the lane-defining step by equivalents. The REQ array does make comparisons that could relate to PDFEs, but even if these comparisons were deemed PDFEs, the method of combining the REQ and NAMP arrays to define lanes differs significantly from the patented process, which does not accommodate multiple PDFEs. The RTMS X3 only confirms whether a pre-selected range slice falls within a traffic lane and does not help identify lane boundaries. Thus, even with the combination of arrays, it cannot be said to define lanes as the patent describes, resulting in no infringement by equivalents. 

Additionally, the district court ruled that EIS could not succeed on its counterclaims of inequitable conduct and best mode, a decision affirmed upon cross-appeal. The overall conclusion reached is that the RTMS X3 does not infringe claim 1 of the ’916 patent, and the counterclaims are dismissed. Each party will bear its own costs on appeal.