This decision concerns predicting travel time for passengers or baggage using statistical data analysis. The applicant argued that the statistical processing provides accurate/reliable future predictions. However, the Board considered that even if a statistical prediction model might involve a great deal of ingenuity, it pertains to the field of statistics applied to logistics planning which is not a field of technology. Since the prediction of a travel time is not a technical activity, a more accurate or faster prediction due to the prediction algorithm cannot be considered technical either, and so are any results thereof (e.g., accurate travel schedule) as the prediction does not affect a technical system or the physical travel process.
Here are the practical takeaways from the decision T 1148/18 (Travel process prediction/THE AQUA ENTERPRISE COMPANY) of December 15, 2022, of the Technical Board of Appeal 3.5.01.
The invention is a system that predicts travel time for passengers or baggage using scheduled flights. It compares passage time predictions for different flights, like flight A and flight B. The comparison uses statistical analysis of past travelers’ data, including their passage times through airport checkpoints and related flight information.
The system calculates average passage times, like arrival airport exit times, and performs a statistical test. Based on the test results, the user gets advice, such as “Make travel plan with sufficient time to spare.”
Claim 1 of Main Request
A travel process prediction system predicting a travel process of a travel object traveling with transportation repeatedly operated at specific time, characterized by comprising:
a means (2) for specifying passage time at which a travel object actually passes through each of a plurality of passage points at a departure/arrival facility of transportation;
a means (2) for acquiring transportation specifying information which specifies transportation used by the travel object; and
a travel process prediction apparatus (1) predicting a travel process of a travel object under a specific condition,
wherein the travel process prediction apparatus (1) includes:
a storage means (14) for storing explanatory data comprising an explanatory text, where the explanatory text which explains a calculation result of a statistical calculation by a means (11) is associated in advance with a feature of information used in the statistical calculation and a result of the statistical calculation, and for storing passage time specified at each passage point and the acquired transportation specifying information in an associated manner for each of a plurality of travel objects;
a means (16) for accepting a plurality of pieces of transportation specifying information which specifies a plurality of pieces of transportation which is candidates to be used by an arbitrary travel object;
a means (16) for accepting a request for comparing passage time at which the travel object passes through a specific passage point when the travel object uses each of the plurality of pieces of transportation;
a means (11) for extracting, from the storage means (14), for each of the plurality of pieces of transportation specifying information, a plurality of pieces of passage time concerning the specific passage point, associated with transportation specifying information having a same content as the accepted transportation specifying information;
a means (11) for calculating, for each of the plurality of pieces of transportation specifying information, a mean or variance of the extracted passage time;
a means (11) for statistically testing a difference in the mean or variance of the passage time calculated for each of the plurality of pieces of transportation specifying information;
a means (11) for extracting the explanatory text from the explanatory data in accordance with the feature of information used in the statistical calculation and the result of the statistical calculation; and
a means (16) for outputting a test result and the explanatory text.
Is it patentable?
In the first instance, the Examination Division refused the European patent application on the grounds that the claims lacked an inventive step (Article 56 EPC).
The Board agreed with the Division and considered that due to the general wording of claim 1, a broad interpretation is used, and it was not inventive over general purpose computer system:
3. The Board judges that, regardless of the above, claim 1, in technical terms, defines a general purpose computer system which runs a software for implementing a non-technical concept, namely predicting a travel process or, more specifically, a travel time such as a duration of stay in an airport.
In light of the description the travel time depends on a number of (possibly interrelated) factors, for example airline schedules, passenger volume, weather conditions, entry formalities or customs/security handling. These factors are either of an administrative nature or a matter of logistics planning. They are, however, not based on technical considerations, for example regarding the operation of a technical system in the airport.
To identify the relevant factors and come up with a statistical prediction model might involve a great deal of ingenuity and be far from trivial. This, however, pertains to the field of statistics applied to logistics planning which is not a field of technology as required by Article 52(1) EPC.
In the Board’s view predicting a travel time based on historical data is conceptually similar to predicting arrival dates for delivery of mail based on previous delivery times (see T 0983/11 – Coordinated marketing/PITNEY BOWES, reasons, point 2.4) or predicting future purchases based on previous ones (see T 0977/17 – Storing electronic receipts/OTTO GROUP SOLUTION PROVIDER, reasons, point 2.2).
These are all activities excluded per se from patentability under Article 52(2)(a) and/or (c) and (3) EPC (cf. also T 154/04 – Estimating sales activity/DUNS LICENSING ASSOCIATES, reasons, points 19 and 20).
The Board considered the technical effect of the distinguishing features over a general-purpose computer system as follows:
4. Claim 1 differs from a general purpose computer system merely in the functions provided by the claimed means.
These functions essentially define the various aspects of statistics involved in predicting the travel time, i.e. data collection, organisation/storage, analysis, interpretation and presentation.
Specifically, they include:
– collecting (“specifying”, “acquiring”) and “storing” logistics (passage time, transportation specifying information) and statistical data (explanatory data comprising an explanatory text);
– receiving (“accepting”) user input (i.e. flight information and a specific passage point) to perform a comparative analysis;
– retrieving (“extracting”) historical data for the given user input and performing statistical calculations (“calculating”, “statistically testing”);
– providing (“extracting”, “outputting”) the results of the calculations and an explanation of the statistics.
First, these steps relate to mental or mathematical activities and are normally part of any statistical data analysis. It is a person skilled in the application of statistical mathematics to logistics planning, not a technically skilled person, that performs this kind of statistical analysis. Mathematical/statistical methods as such do not have technical character (Article 52(2)(a) and 52(3) EPC), and cannot contribute to inventive step.
Second, the output of the analysis is not used for a technical purpose. It merely supports the user in creating a (possibly more accurate) travel schedule – see paragraph  of the application.
5. Under the COMVIK approach (see T 641/00 – Two identities/COMVIK) the above functions are, as they are part of the non-technical requirements specification, given to the skilled person within the framework of the objective technical problem to be solved.
The Board, therefore, considers that the skilled person, a computer programmer, faces the objective technical problem of implementing them in a general purpose computer system.
The computer implementation is obvious because it merely claims “means” for performing the various functions without any technical details.
The applicant argued that the features provide a technical effect of error processing. But the Board disagreed:
6. In the written procedure the appellant relied heavily on arguments concerning features not in the claim, for example regarding the operation and (error) processing of check machines.
The Board notes that the claims define the matter for which protection is sought (Article 84 EPC). Hence, they must be examined for compliance with the provisions of the EPC. In particular, technical features not present in the claims cannot support an inventive step argument and, thus, are not relevant for discussing inventive step.
7. The Board, however, agrees with the appellant that Figure 1 indeed shows a hardware architecture going beyond a general purpose computer system, but this is not reflected in the claim.
As mentioned above, the claimed “means” neither refer to specific hardware means (e.g. readers, servers, etc.) nor to a specific hardware/software configuration that could provide a further technical effect (cf. T 1173/97 – Computer program product/IBM, Headnote), e.g. by allocating functions to different system components.
The claim merely provides a functional definition of “means” and, thus, is not limited to any particular technical features. In other words, the means may be realised by simply programming the corresponding functions on a computer.
8. The statistical calculations might, as argued by the appellant, provide accurate/reliable prediction results. This is, however, an inherent property of the calculations themselves and not a result of a particular implementation or the underlying technical system.
9. In the Board’s view the prediction of a travel time is not a technical activity – see point 3 above. Hence, a more accurate or faster prediction, in the sense of an enhanced speed of the prediction algorithm, can not be considered technical either (see T 1954/08 – Marketing simulation/SAP, reasons, point 6.2).
Also, any results thereof, e.g. a more accurate travel schedule, are not technical as the prediction does not affect a technical system involved in, and thus impacting on, the (physical) travel process.
For the same reason also the appellant’s argument that using a large amount of (measurement) data and applying complex statistical calculations were not usual at the priority date of the application fails. These are, as outlined above, non-technical activities and can be included in the problem specification. Thus, for assessing inventive step it is irrelevant whether or not they were known in the prior art.
The applicant argued that the features provide a technical effect of feedback or self-learning system enabling more accurate/reliable future predictions. Again, the Board disagreed:
10. The appellant argued that storing explanatory data/text would provide a feedback or self-learning system enabling more accurate/reliable future predictions.
The Board cannot see that the explanatory data (see Figure 18) is in any way related to a feedback/learning loop for controlling a technical system. The purpose of the explanatory text is to explain to the layperson the results of statistics – see Figure 19. Such explanation neither represents technical data nor supports the user in performing a technical task. Thus, it cannot produce a technical effect (see T 1670/07 – Shopping with mobile device/NOKIA, reasons, point 13).
Therefore, the Board decided that the subject-matter of the claims was not inventive.
You can read the full decision here: T 1148/18 (Travel process prediction/THE AQUA ENTERPRISE COMPANY) of December 15, 2022, of the Technical Board of Appeal 3.5.01.