Posted: July 5th, 2021
UK Patents Court holds electronic interface patent invalid, by Graham Burnett-Hall, Marks & Clerk
The UK Patents Court has issued its judgment in an infringement action brought by Add2 Research and Development Limited against dSPACE Digital Signal Processing & Control Engineering GmbH and dSPACE Limited, in which Add2 alleged infringement of European Patent No. 1 163 622 B1. Judgment was handed down on 17th June 2021.
A large number of issues were covered by the judgment. The most important were whether Add2’s patent was infringed by dSPACE’s SCALEXIO system and whether the patent was valid. dSPACE did not seek revocation of the patent – by the time of the trial the patent had expired – but dSPACE raised validity as a defence to the infringement claim. As will be explained below, the judge, Mr Justice Meade, held that specific aspects of the SCALEXIO system would have infringed the patent but that nevertheless the infringement claim failed, as all the relevant patent claims were invalid.
The patent claims in issue were not the ones originally granted. Add2 sought permission to amend the claims and did not contend that the unamended claims were valid. As proposed to be amended, the claims had particular relevance to interfaces for use in simulation systems, in particular, “hardware-in-the-loop” (HIL) simulation.
Many mechanical systems are controlled by electronic control units (ECUs). This has particular application in the automotive sector: a modern car may have dozens of ECUs controlling all aspects of the car, for example, fuel injection or operation of the electric windows. In order to test the functioning of an ECU during development, the ECU could be connected to a test vehicle. However, it is often preferable to simulate the vehicle, or the relevant parts of the vehicle. This requires a simulation system with a processor that can run a mathematical model of the device being simulated in real time, and inputs and outputs that can receive and send signals from and to the ECU. In practice a system may have dozens or even hundreds of signal channels carrying signals between the simulation system and the ECU.
The signals received by the ECU need to be in the form that the ECU would normally receive in the real world. This may require conversion of the signals as generated by the simulation system, for example the voltages, currents and loadings, so that they match those experienced in the real world. Similarly, the signals produced by the ECU may need to be converted to signals of the type that the simulation system can handle. For example, voltages may need to be adapted from those typically found in an automotive context to TTL level voltages. Such conversion and adaptation is known as signal conditioning.
The key claim of Add2’s patent concerned “a signal interface circuit for providing an interface between a simulation system and a system under test”, with a circuit portion that provided an analogue interface and a circuit portion that provided a digital interface, where the circuit portions could be enabled and disabled so as to reconfigure the interface. The signal interface circuit also had to include features relating to a load, which in a HIL system would replicate a real-life component or could even be a real component, such as an injection valve. In particular, the load had to be selectively connectable to a high or low power rail. It was common ground that these claim features would be understood by the skilled person as relating to a “single ended load”, where one end of the load was connected to an input terminal and the other to the desired power rail.
dSPACE’s SCALEXIO system was a modular system that comprises a large variety of hardware boards and units, a number of which may be combined in a specific system. The boards include input/output (I/O) channels that are suitable for a variety of different signal types. Of particular relevance to the infringement claim, some channels are intended to be used with analogue signals, other channels with digital signals. There are also “Flexible In” channels that can be used to take both digital and analogue measurements of a signal produced by an ECU.
The infringement arguments
By the time of the trial Add2’s infringement case concerned just two of the Flexible In channel types, designated “Flexible In 1” and “Flexible In 2”. There were two infringement issues for the Court to resolve. First was a point of construction: did the “signal interface circuit between the simulation system and the system under test” of the claims include the signal generation and measurement components of the I/O boards or was the claim concerned with the signal conditioning circuitry alone? Add2 argued that the former interpretation was correct.
dSPACE’s position was that the patent, which only described a single embodiment of the invention, was solely concerned with signal conditioning. This mattered because those parts of the Flexible In channels that provided the signal conditioning for analogue and digital signal measurements were always enabled, whereas the patent claim required the circuit portions to be selectively enabled and disabled.
However, if the I/O components were included in the interface, it would be possible to program the SCALEXIO boards so that either (or both) of the analogue and digital signal measurements would be relayed to the real-time processor of the simulation system.
The Court held that the former interpretation was correct: the signal interface circuit of the claims could include the I/O portions of the signal channels. The fact that the patent only had one specific embodiment did not mean that the claims were not to be understood more generally.
The second infringement issue concerned the load and the meaning of “connectable selectively to a high or low power rail”. The Court held that this did not require that the load must be capable of being switched between a connection to a high power rail and a low one, whilst remaining in situ; the claim also covered situations in which the load would have to be disconnected entirely before the connection to the power rail was changed.
There was also a complex indirect infringement issue, concerning the failure insertion capabilities of the SCALEXIO system. This turned out not to be central to the judgment.
The main validity issue was obviousness: dSPACE contended that all the claims in issue were obvious over, first, an article (“Woermann”), which described a HIL system known as CARTS and, second, over one of dSPACE’s own prior art boards known as the DS1103.
The CARTS system described by Woermann included a VME bus computer, with I/O CPUs responsible for signal generation and signal measurement, and an “I/O subsystem”, from which signals were sent to and received from the ECU. The obviousness argument centred on the understanding of what was disclosed by a component in the I/O subsystem described as a “switch matrix” and whether this was under software control or only comprised manual switches. Add2’s position was that the skilled person would believe the switch matrix to be a manually controlled device.
Add2 further argued that the skilled person would be held back by a fixed “mindset” that, for any HIL system, the interface between the simulation system and the ECU would always be built from scratch. The judge rejected this argument. In any event, the “mindset” argument did not even get off the ground in the context of Woermann, as the switch matrix of Woermann, whether computer or manually controlled, provided the flexibility of being able to adapt the interface.
As to whether the switch matrix disclosed by Woermann was computer or manually controlled, the judge considered that Woermann was not sufficiently explicit to be able to determine that is was under computer control. However, the judge found that this would have been obvious to do, noting that Add2’s expert very clearly accepted this under cross-examination. It followed from this finding and from the common general knowledge that all the patent claims in issue were invalid.
dSPACE’s separate validity challenge based on the DS1103 board failed, the judge holding that the argument was overly elaborate and based on hindsight.
There were a number of peripheral issues that the Court had to decide, including added matter and insufficiency arguments and pre-action conduct. On pre-action conduct, the judge noted that there had been a meeting between the parties prior to proceedings being commenced, at which dSPACE had tried to explain why there was no infringement but was told by Add2’s director that it was pointless to do so, but the judge nevertheless concluded that dSPACE should have provided more information about their products at the pre-action stage.
However, the judge also noted that this was not a general finding as to what defendants have to do at the pre-action stage of a patent dispute – for example, it could be appropriate to withhold information if early discussions make it clear that the patentee’s claim is speculative.
Of particular interest to legal practitioners will be the finding that an assignment of the patent in suit to the Claimant was void. Before alleging infringement, the patent was held by a trading company, Add2 Limited. The directors of Add2 Limited transferred the patent to a dormant company, Add2 Research and Development Limited (the Claimant), for no consideration. The two companies were linked through common ownership and common directorship. dSPACE argued that the assignment should be characterised as an unlawful return of capital to the shareholders of Add2 Limited and was therefore void under common law. The consequence of this finding would mean that, unless the deficiency with the assignment were corrected, the Claimant would lack the standing to bring the claim.
Mr Justice Meade agreed: although the director and shareholder concerned had acted entirely honestly and although Add2 Limited was solvent at the time of the assignment, that did not change the fundamental analysis, which was that a valuable opportunity had been given away by Add2 Limited for no value. This amounted to the taking of an asset out of Add2 Limited by voluntary distribution and the relevant assignment was void.
In this case, however, the original patentee, Add2 Limited, had subsequently gone into liquidation. The liquidator executed a confirmatory assignment transferring any ownership Add2 Limited still retained in the patent to the Claimant. Mr Justice Meade held that this remedied the deficiency on the basis that the liquidator was not subject to the same common law rule on the distribution of a company’s assets.
It is not unusual for a holder of an intellectual property right to consider assigning IP to a separate legal entity before bringing infringement proceedings. This case demonstrates that considerable care will need to be taken to ensure that appropriate consideration in return for any transfer of intellectual property is provided.
A copy of the judgment can be found here.
[Marks & Clerk Law LLP represented dSPACE]