3 December 2018

Patenting the Internet of Things

As the Internet of Things revolution gathers pace, we take a look at the challenges that are likely to face applicants seeking to protect new IoT-related inventions.


The Internet of Things is one of the technology sectors at the heart of the so-called Fourth Industrial Revolution (4IR), which also encompasses such diverse fields as robotics, AI (Artificial Intelligence), 3D printing and autonomous vehicles. A 2016 study by the McKinsey Global Institute estimated that by 2025 there will be around 26-30 billion IoT-enabled devices in the home and workplace. This proliferation of connected devices will undoubtedly drive numerous innovations, many of which will of course be eligible for patent protection, as researchers explore new applications and use-cases.

Unsurprisingly, the number of 4IR-related patent applications has grown commensurately with increasing R&D activity in this technology sector. In a study titled “Patents and the Fourth Industrial Revolution”, the European Patent Office (EPO) found that the number of 4IR-related applications filed at the EPO increased by 54% from 2013 from 2016, making it one of the fastest-growing technology sectors in terms of the number of patent filings over that period.

In this article, we will look at examples of some of the types of innovations that can arise when developing IoT systems, and discuss the key issues that need to be considered when assessing the patentability of IoT-based inventions.

IoT software patentability

As the number of connected devices and smart sensors increases, so too will the volume of data being generated by those devices. Regardless of whether that data is sent to the cloud for processing, or is processed at the network edge, the software that processes the data and generates some form of output can be eligible for patent protection in its own right, provided certain criteria are fulfilled.

To answer the question of whether a new software program is likely to be eligible for patent protection in Europe, it is first necessary to consider the function that the software performs, and in particular the problem that is being solved. It is well-established in European patent law that in order to be eligible for patent protection, a novel software program must provide a solution to a technical problem. Unfortunately, however, the law does not provide a single definition of what is meant by ‘technical’ in this context. Consequently, it is usually more constructive to discuss the meaning of ‘technical’ with reference to specific examples, rather than attempting to come up with a generic definition.

For instance, consider a hypothetical system in which the novelty lies in the fact that a blockchain ledger is used to track a chain of transactions relating to the same physical asset. A blockchain inherently contains information about all transactions in the chain, and can allow an audit trail to easily be constructed. In the case of non-performance of one of the contracts in the chain (e.g. failure to deliver the agreed goods), the blockchain audit trail can allow quick and easy identification of point along the chain at which the fault occurred, which ordinarily would be a time-consuming and complex process under current systems. However, from a patentability perspective, an EPO examiner might well object that this constitutes a solution to a business problem, rather than being a solution to a technical problem.

Conversely, consider the example of an asset performance management (APM) system in which an AI algorithm receives data from smart sensors which are monitoring a piece of equipment in an industrial facility. The sensors could measure parameters that provide the algorithm with information about the current physical state of the equipment, such as the surface temperature, humidity levels, vibration, noise, and so on. The AI algorithm then processes the sensor data to attempt to predict when failure of the equipment is likely to occur, so that an engineer can be despatched to repair or replace the part in question before failure occurs. In this way, it is possible to avoid a situation in which the component fails unexpectedly, potentially damaging other components and resulting in further system downtime while a replacement part is procured and installed. If the only novelty lies in the AI algorithm itself, and the new algorithm is able to predict when a failure will occur with greater accuracy than existing solutions, then the system incorporating the novel AI-based software may well be patent-eligible on the grounds that it solves a technical problem, namely how to improve the reliability of the system as a whole.

It is also worth briefly mentioning the potential impact of 5G on the patentability of IoT software solutions. The next generation of mobile communication networks, referred to as fifth-generation or ‘5G’ networks, will enable machine-to-machine communications at significantly lower latencies than can be achieved with existing 4G technology. Particularly in industrial IoT (IIoT) applications, where latency is often of critical importance, existing solutions have typically been limited to data-gathering and analysis without performing real-time control, due to the relatively high latency of 4G (or older) networks. It is anticipated that the lower latency of forthcoming 5G networks will enable the next generation of 5G-based IIoT systems to control industrial processes in real-time.

This is of particular relevance when it comes to the patentability of software, since under both UK and European patent law a technical effect is normally present if the new software has an effect on a process carried on outside the computer, for example, controlling machinery on a production line. Of course, a software program which controls a 5G-based IIoT system will still need to meet the usual requirements of novelty and inventive step. However, the fact that a physical process is being controlled should help to satisfy the ‘technical effect’ requirement, and ought to mean that patent applications directed to these types of software-based solutions will have one less hurdle to overcome.

IoT hardware patentability

The second of the two examples mentioned above, namely an APM system which can more reliably predict the failure of a piece of equipment, can also be used to illustrate types of hardware innovations that are likely to be patentable in an IoT context. In the example described above, the only novelty lay in the software algorithm that is used to process the data. However, it is equally conceivable that the reliability of the AI prediction may be increased by modifying the hardware to improve the quality of the data that is fed into the algorithm. Here, if the novelty lies simply in substituting a more accurate temperature sensor for the sensor that would normally be used in existing systems, and the improved type of sensor is itself already known, then it is likely that this would be considered to be an obvious modification and therefore would not constitute a patentable invention.

On the other hand, if the invention lies in the realisation that relocating the sensor to a different position on the equipment will provide a temperature reading that gives a better indication of the likelihood of failure, then this modification may well not have been obvious. For example, experiments may have revealed a surprising correlation between the temperature at a particular part of the equipment and the observed failure rate. If so, then it should be possible to secure patent protection for a system comprising: (i) the equipment itself; (ii) an IoT temperature sensor disposed in the new location; and (iii) a server configured to receive temperature data from the sensor and use the data to predict when the equipment will fail, even though the individual hardware components may themselves be known. Also, in some cases it may be possible to obtain independent protection for (i) and (ii) together, without having to include the server.

Of course, in addition to improvements at a system level, there are clearly many opportunities for patentable inventions to arise at the component level. Improvements to hardware components such as energy harvesting mechanisms, power supplies, batteries, wireless transceivers, antennas, sensors, actuators, and so on, may well be eligible for patent protection in their own right.

IoT use-cases: point solutions to specific problems

As more and more devices become IoT-enabled, the number of potential use-cases will increase. At the recent IoT Solutions World Congress 2018 in Barcelona, it was often remarked that the IoT sector should not be seen as a single market from a commercial perspective, as the range of potential IoT use-cases are too diverse to be considered as one all-encompassing market. Instead, delegates were encouraged to think in terms of ‘a thousand million-dollar markets’ as opposed to one billion-dollar market. However, a consequence of the fragmented nature of the IoT sector means that there are numerous opportunities for start-ups and SMEs to succeed by offering point solutions to specific problems.

Returning to the APM examples discussed above, it will usually be the case that the invention has been developed with a particular application in mind. For example, the APM system may have been designed to predict the failure of mechanical bearings in a steel rolling mill. However, it may be theoretically possible to apply the same principles to other types of systems in which similar bearings are used. Consequently, there may be a temptation to draft patent claims broadly without specifying the environment in which the component is operating.

However, in reality there may not be any commercial need to protect these other theoretical use-cases. For example, failure of the same type of component in other systems might be relatively inconsequential, in which case there may not be a commercial market for an expensive IoT-based APM solution. Drafting claims too broadly at the outset can complicate matters during prosecution, as otherwise unrelated prior art may be cited. Also, it may be more difficult for a patent examiner to identify the technical problem that is solved by the invention when the claim scope is too broad, resulting in objections that would not otherwise have been raised if the claims had been drafted more narrowly at the outset. Therefore in cases where a business model has been built around a single use-case, as is often the case for start-ups, applicants may wish to consider drafting their claims narrowly to focus on the specific application that is relevant for their business. This can help to improve the chances of success for the patent application, and to reduce costs during prosecution.


As the IoT sector matures and competition in the marketplace becomes more intense, it will become increasingly important for companies to protect the IP in their commercial products. Without proper protection in place, competitors will be free to introduce their own copycat products, resulting in a negative impact on market share and profits. However, by being careful to spot potentially patentable inventions at an early stage when developing new IoT solutions, and by drafting claims that provide a clearly-defined solution to a specific technical problem and therefore support a strong inventive-step argument, companies can build a patent portfolio that will allow them to maintain their competitive edge.