The patent relates to security systems and more particular to wireless sensors used within such systems. It provides for automatic reporting of prognosis data from wireless mesh sensors.
Brief outline of the case
The opponent appealed the OD’s decision to reject the opposition.
The board set aside the decision of the OD and decided revocation for lack of sufficiency under Art 100(b) for the granted patent and Art 83 for all AR.
The case is interesting in view of the reasons for lack of sufficiency.
The sticking point between the parties is the interpretation to be given to the expression: “a cloud processor (46) of the cloud application (44) that determines a location of the sensor (12, 14) …“
The meaning of the term “determines” was the subject of dispute between the parties.
The opponent’s position
The invention according to claim 1 as granted referred to an apparatus comprising a cloud processor of the cloud application that “determines” a location of the sensor.
The patent did not describe how the cloud processor could determine the location of the sensor solely on the basis of the information described in the patent to be sent to the cloud processor.
Even if the cloud processor would be considered to receive an identifier of a specific sensor, the identifier as such did not enable the cloud processor to determine the location of the sensor.
Furthermore, a position of a control panel or of a secured geographic area did not correspond to a location of the sensor, because several sensors were assigned to the control panel and to the secured geographic area and the distance between these elements could be such that the potentially faulty sensor could not be easily located by a responsible person.
The proprietor’s position
The patent was not prejudiced by the ground for opposition under Art 100(b).
The term “determine” in the sense of claim 1 meant “ascertain or establish exactly by research or calculation“. Thus, the cloud processor in this sense ascertained or established where the wireless sensor was.
Since the cloud processor was located remotely to the sensors, it could not perform any physical ranging or measurement in order to determine the location.
The skilled person would be aware of that and exclude a corresponding interpretation of claim 1.
The only reasonable interpretation for the person skilled in the art was thus that the cloud processor determined the sensor location in the sense that it “ascertained” the sensor location, which included retrieving a location of the sensor from a database. Since the type of security system which was the subject of the invention would normally be thoroughly documented, such information would have been available to the cloud processor.
In order for the cloud processor to determine a location of the (potentially faulty) sensor, it could receive a GPS signal from the control panel or from individual sensors, if they were equipped with GPS modules.
Furthermore, “a location of the sensor” as recited in claim 1 referred to a location within a secured geographic area, and no higher accuracy of a sensor location to be determined by the cloud processor was required.
Consequently, the location of the secured geographic area could be considered to correspond to a location of the sensor to be determined by the cloud processor within the meaning of claim 1. Since geographic locations were static, there was also no difference between the current location of the sensor and a previously stored location of the sensor.
As regards “a location of the sensor” within the meaning of claim 1, another central point of the proprietor’s argument was that the sensors necessarily had to be in the vicinity of the control panel in order to communicate with it.
Based on this interpretation of claim 1, the proprietor considered that the location of the control panel processor sufficiently reflected a location of the sensors within the meaning of claim 1.
Finally, the proprietor argued that one way to carry out the invention was disclosed in paragraphs [0031] to [0040] of the patent.
The board’s position
The board agreed with the proprietor to the extent that it is not justified to interpret the above feature in an unjustifiably narrow manner.
The board also agreed with the proprietor that the required accuracy of a location of the sensor to be determined by the cloud processor is not defined in claim 1. An approximate or relative position of the sensor, depending on the specific application and configuration of the claimed apparatus, may well be considered to be within the scope of a functionally reasonable interpretation of claim 1.
This consideration on the part of the person skilled in the art is not entirely without limits and, in particular, is not entirely unrelated to the result envisaged by the claimed invention.
The objective of the present invention is not only to identify the potentially faulty sensor, but an important aspect is to locate it in such a way that a responsible person can take the necessary measures to remedy the potential fault.
For the board, it is not apparent why claim 1 should be interpreted more narrowly with respect to assessing the positional relationship between the control panel processor and the sensors. Claim 1 says nothing about the spatial relationship of the control panel processor and the sensors.
The board however held that the meaning of the term “determines” was undoubtedly unclear in the context of claim 1. The proprietor has conceded that the term implies at least some kind of activity or operation of the cloud processor that is performed to determine a location of the potentially faulty sensor.
What was actually disclosed?
For the board, it is not directly and unambiguously derivable from paragraph [0031] of the patent that the cloud processor receives information enabling it to determine a location of the sensor.
Paragraph [0032] merely discloses that the cloud application may include one or more computer programs running on a processor. It does not provide any explicit or implicit technical teaching as to what activity or operation the processor performs in order to determine a location of the sensor and what data or information is processed for that purpose.
Paragraph [0033] discloses, among other things, a notification sent to the responsible person that includes an identification of the sensor by location and time. This passage is concerned with the notification sent by the cloud processor to a responsible person and does not include any information about how the processor actually determines a location of the sensor.
Paragraph [0034] of the patent is concerned with the content of the notification to be sent to the responsible person with regard to a potential failure, but it is not apparent what the skilled person could infer from this disclosure about how the processor determines a location of the sensor.
From paragraph [0040] the person skilled in the art can derive that “the cloud” receives a notification of the potential failure along with the GPS location of the panel. The board considered that the skilled person would understand claim 1 to mean that the cloud processor determines a location of the (potentially faulty) sensor at least sufficiently accurately so that it can be found among the multiple sensors without the responsible person having to make considerable additional efforts.
In the absence of any further description in the patent of how a location of the (potentially faulty) sensor can be determined among the multiple sensors from the position of the control panel, paragraph [0040] cannot be regarded as disclosing a way of carrying out the invention according to claim 1. Similar considerations apply to a location of the secured geographic area.
In conclusion, contrary to the proprietor’s argument, the patent does not disclose a way of implementing the invention according to claim 1, in particular, how a cloud processor can determine a location of the sensor within the meaning of claim 1.
Common general knowledge
A further question to be answered is whether the person skilled in the art is able to identify, on the basis of common general knowledge, how the apparatus is to be implemented in such a way that the cloud processor determines a location of the sensor in order to achieve the desired effect, i.e. to find the potentially faulty sensor without undue burden, in order to avoid false alarm penalties that may arise due to a faulty sensor. The board was not convinced that this was the case.
Finally, the board noted in general that the scope of claim 1 encompasses a large number of technically meaningful embodiments of the apparatus, not all of which the proprietor may have had in mind when drafting the patent.
In particular, the apparatus of claim 1 covers a large number of different combinations of any kind of different sensors, threats and secured geographical areas, not only in terms of the quantity of these individual elements, but also in terms of their functional and relative positional relationship to each other.
It is an established principle that the mere fact that the scope of a claim is broad does not in itself lead to a lack of sufficiency of disclosure. The board further agreed with the proprietor that the onus is normally on the opponent to substantiate, by means of verifiable facts, serious doubts that the patent does not disclose the invention in a manner sufficiently clear and complete for it to be carried out by the skilled person.
What would have been needed
In the present case, however, the patent does not contain any description of how the cloud processor determines the location of the (potentially faulty) sensor, which casts serious doubts on whether the invention can actually be carried out by an average person skilled in the art without undue experimentation or inventive skill by following the instructions of the patent.
The cloud processor, in order to perform the claimed function, is dependent on external information received from other parts of the claimed apparatus. For the skilled person to carry out the invention, it would therefore be necessary to find a way of implementing the apparatus as a whole, depending on its specific configuration, in particular in terms of type and number of sensors, threats to be detected as well as the type and extent of the secured geographic area.
In particular, the skilled person would need to suitably identify, based on common general knowledge, a sufficiently precise location of a potentially faulty sensor so that a person responsible for locating the potentially faulty sensor can do so without undue burden.
They would also need to determine how the information about this location can be obtained within the apparatus, how it can be transmitted to the cloud processor, and how the cloud processor processes this information to finally determine the location of the (potentially faulty) sensor.
For none of the extensive tasks which should be carried out by the cloud processor, the patent or the common general knowledge provide the skilled person with guidance on how to construct the apparatus so that the cloud processor can satisfactorily perform its claimed function of determining a location of the (possibly faulty) sensor.
Comments
The present case illustrates that it is possible to object a lack of sufficiency without the need of bringing forward verifiable facts. It can be enough to show, in view of the breadth of the claim, that serious doubts arise that the patent does not disclose the invention in a manner sufficiently clear and complete for it to be carried out by the skilled person. What is claimed is clearly not plausibly or credibly disclosed.
It however remains an exception as lack of sufficiency cannot primarily only be shown on the basis of verifiable facts. In the present case, the onus of proof was shifted to the proprietor.
The boards have certainly more time than an ED or an OD to look at a case, but the high production pressure on examiners actually withhold them to enter such difficult discussions.
The present case is nevertheless one more in the collection of 50+ cases in which the decision of the OD is set aside.
https://www.epo.org/en/boards-of-appeal/decisions/t210005eu1
Comments
2 replies on “T 5/21 – A too broad claim can lead to a problem of sufficiency”
Dear Mr Thomas,
I agree that it may be difficult for examiners to spot sufficiency issues, because this requires a thorough reading of the description and a need to delineate the common general knowledge of the skilled person as of the filing date. I know from experience as an opponent that this may require a lot of work, particularly for fast-moving technologies. The current pressure on production certainly does not help. But to be honest, this was already the case when the pressure was not so high.
There are other reasons for which examiners are less than eager to raise Art 83 objections, at least in mechanical/electrical areas. First, an Art 83 objection has a poor optics : it may appear as applicant-unfriendly, since if the objection is justified, it cannot be cured without adding new matter. Second, examiners know that the applicant typically has an expertise in the specific field of the application much more specialised and sharper than their own knowledge, and may feel intimidated. I take from my exchanges within SACEPO with a member of DG1 in charge of supervising a area of technology that he had only seen two rejections based on Art 83 in his entire career.
The EPO actually relies on third parties to spot Art 83 issues either in opposition or by way of Art 115 observations. It is of note here that from the mid-2010’s, the EPO has encouraged Art 115 observations and this has had some success. According to Guideline E-VI 3, the scope has been broadened beyond patentability conditions i.e. beyond the literal scope of Art 115, to include Art 83, 84, 123(2) and more.
Another development of interest occurs in AI inventions, as sufficient disclosure emerges as an issue for training datasets and algorithms.
The EPO is not isolated in this practice. The recent PACTE reform of French law, while adding a review of inventive step by INPI, did not extend the review to sufficiency.
Dear Mr Hagel,
Thanks for your interesting comments.
I agree with you that ED are in general reluctant to raise objections under Art 83. This is particularly the case when the statements in the whole disclosure appear plausible and would actually need experimental data. This should however not hinder that in case of doubt, to raise an objection under Art 83. This depends widely from the situation and cannot be a general rule.
There are other situations in which the objection need not be an objection under Art 83, but can also be an objection under Art 84 or 56.
I refer here to G 1/03 and T 2001/12. T 2001/12 has been followed by lots of decisions confirming it. The conclusion from G 1/03 or T 2001/12 can be resumed as follows:
– If all the features allowing to solve the problem defined in the application/patent are present in the description as filed, but not in the claim, the objection to raise is Art 84.
– If all the features allowing to solve the problem are not present in the description as filed, but the effect to be achieved is claimed, then the objection to raise is Art 83.
– If all the features allowing to solve the problem are not present in the description as filed, but the effect to be achieved is not claimed, then the objection to raise is Art 56.
There is thus a link between Art 84, 83 and 56. It all depends on what is the description and/or in the claims.
If some additional features are needed to solve the problem, but are not to be found in the specification as filed, then the objection cannot be Art 84.
If experimental data are needed, those can only come in examination by third parties observations. As experimental data have a cost, I cannot see a third party carrying out experiments and filing them as third party observations. I would rather expect that the third party waits until grant and then files experimental data as opponent.
There are however situations in which experimental data are not required. This is for instance the case when a method of measurement of a parameter is not included in the description or the claims, unless the parameter is a well-known one, e.g. Vickers hardness.
Some boards do however consider that the absence of measurement method of a parameter does not deprive of the promise of the invention and therefore consider in opposition the objection to be an objection under Art 84 and hence not to be dealt with under G 3/14. Other boards consider that the absence of a measurement method boils down to a lack of sufficiency. I am rather inclined to follow the latter boards, than the first ones.
As soon as information, not present in the original specification is needed and this lack of information cannot be compensated by common general knowledge, then in general, we have a problem of sufficiency. I follow you without hesitation when you state that an objection of lack of sufficiency could only be overcome by adding subject-matter. This is to me a very simple way to determine in first approximation whether the objection should be Art 84 or Art 83.
If the measurement of a parameter depends from a norm, it is of utmost importance to originally define the version of the norm to be used. Norms can change with time, even between the priority and the filing date. Norms can also change after grant. Under Art 123(2) it is not possible to add the version of a norm after filing.
A typical example are viscosity measurements with Brookfiled viscosimeters. There are lots of variations possible, rotation speed form of the spindle, and without the exact configuration, the results can be quite different and not comparable.
Some norms are even secret and licensees are not allowed to disclose the norm, e.g. CD or DVD. Any invention improving CDs or DVD might face an objection of lack of sufficiency.
In my last directorate we had the chance to deal with perpetuum mobile. It is clear that an objection under Art 83 has to be raised when the applicant explains that he gets more energy out as he has put in the system, or that friction can be ignored. Most of this type of applications are actually not refused under Art 83 but under Art 123(2) for obvious reasons.
I also agree with you that for AI inventions, not only the algorithm but also the training data are necessary for the invention to be reproduced by the skilled person. As data are precious goods, the holder of such data might not be inclined to disclose those. I am therefore rather inclined to think that for this first reason, AI inventions will not inundate the EPO. Another reason will be G 1/19. Without a technical effect in real life resulting from the AI, obtaining a patent might be difficult.