Autonomous vehicle standards won’t drive patent wars
Counsel at AV stakeholders, including Nvidia, AGC and two car companies, say standards are likely to emerge but will probably not lead to the same skirmishes seen in the connectivity space
One minute read: In-house counsel say standards could be established for various autonomy-enabling AV components, including mapping, infrastructure data processing and car communication. Whether those standards emerge will depend on a range of external factors, business drivers and market dynamics. But if standards do emerge, they are unlikely to lead to significant skirmishes because the pool of patent owners will be small the the relevant components will be cheap.
Standards are likely to spring up for various autonomy-enabling components of driverless cars, but they will probably not spur the same kind of negotiation or litigation skirmishes that have erupted between telecoms and automotive companies.
In-house sources at car makers and automotive suppliers tell Patent Strategy that interoperability and regulatory drivers could push the creation of standards and standard essential patents (SEPs) in area such as mapping, artificial intelligence (AI) and public infrastructure connectivity.
But the fact that these standards are likely to be developed by a handful of players, that the components should be cheap and that a lot of autonomous innovation will focus on software, and thus not be patent eligible, will diminish the chances of a new front emerging in the SEP wars.
“If IP is developed that becomes relevant to a standard for car autonomy, it will likely be in the possession of about two dozen companies, each of whom would need to take licences from one another,” says Jason Skinder, global patent counsel at a tier-one auto company in Michigan. “Ultimately, that will be resolved in some sort of cross-licensing scheme.”
Patent wars could break out over autonomous standards, of course, if a single SEP owner decided that it wanted to monetise its portfolio aggressively – although that is true of virtually any standardised technology.
“If SEP holders take a more reasonable approach and realise they can make a significant amount of money on hardware, that approach would be a much smarter way forward.” Eric Rogers at AGC Automotive Americas
But the head of IP at a car manufacturer in Germany says that such a scenario is unlikely because most autonomy-enabling components will not be expensive enough for either side to justify prolonged and costly litigation.
“The relevance of particular patents is one thing. But in the end, it all comes down to costs,” he says.
“If we take video coding as an example, most film and content streaming companies pay MPEGLA for access to high-resolution video streaming technologies. But there are no fights there because each patent doesn’t cost very much to license.
“If the price is too high for a technology, fights erupt as implementers try to avoid those costs. If the costs are fair, no one fights because it is simply cheaper to take a licence. And I do not foresee anyone fighting over the components for our cars.”
Eric Rogers, IP counsel at auto supplier AGC Automotive Americas in Michigan, agrees that cost will be an important factor, but points out that the hardware for autonomous driving is currently much more expensive than most telecommunications equipment.
He adds that a single autonomous driving sensor, such as automotive-grade Lidar, is currently priced at more than $1,000. “We are talking about much higher prices than for telecoms equipment. If SEP packages are going to be priced as a percentage of the hardware cost, that model is going to be a much tougher pill to swallow.
“If SEP holders take a more reasonable approach and realise they can make a significant amount of money on hardware, that approach would be a much smarter way forward.”
The car maker head of IP adds that his assumption is that if patent wars do come out of autonomous standards, they will be settled by those automotive suppliers that build sensors, and not by car manufacturers in any case.
“If there is a fight, we would trust our suppliers to solve the issue. There may be standards but I am not sure there is a huge risk of patent fights that will come out of them,” he says.
The industry might see an uptick in litigation surrounding autonomous SEPs if patent trolls pick up enough IP to assert against implementers. But Stephen Chandler, European general counsel at autonomous AI developer Nvidia in the UK, points out that it will be a while before enough universities and businesses start to sell their autonomous vehicle SEPs.
“Hopefully, any litigation that might emerge here would be the sensible type. It is a whole different type of discussion and negotiation when it comes to speaking with other innovators,” he says.
Mapping out standards
Standards are currently few and in their infancy. But with the potential safety concerns surrounding driverless cars and benefits of having said systems that are compatible with other vehicles and public infrastructures, it is not unreasonable to expect more standards to arise.
Skinder says that we might start to see standards established for maps, for example, because of the benefit of operating driverless cars in pre-charted areas. He explains that each driverless car company currently has its own mapping system to enable its vehicles to know where things are and thus better navigate the space around them.
“It is an area that is ripe for standards, but the competitive landscape might preclude those standards from becoming established for a while.” Jason Skinder
These maps are not necessarily interchangeable, however, because they are created with sensor sweeps that are specific to a particular company’s cars and the technology suite those cars use. One firm, for example, uses a combination of radar and Lidar and other technologies for its sensor sweeps, and the maps it generates would be incompatible with a different company’s cars if they used a different technology suite for their mapping.
“The benefits of standardising this mapping process is that each car company could benefit from the maps generated by another company,” says Skinder. “Any autonomous vehicle could therefore operate in an area previously mapped, whereas currently it must do its own sensor sweep before it can work in a particular area.”
He adds that it may take a while, however, for the incentives of mapping standards to outweigh the competitive advantage of having proprietary systems. While mapping standards would speed up the process of mapping areas, they would allow competitors to enter markets that they could not otherwise enter.
“It is an area that is ripe for standards, but the competitive landscape might preclude those standards from becoming established for a while.”
Another area where standards might arise, according to Skinder, is public infrastructure connectivity. There is an opportunity for cities to standardise their infrastructures to allow them to better communicate with autonomous vehicles.
Las Vegas, for example, implemented vehicle-to-infrastructure technology into its traffic-signal network to enable two-way communication between cars and traffic lights and traffic cameras. The technology should enable drivers – both AI and human – to identify better routes and traffic incidents.
It could also make autonomous driving safer in the future by enabling cars to tell if they might have to stop suddenly for a changing light. Companies already have systems that allow cars to detect changes in the infrastructure around them, but those have certain failure rates and simply cannot be as reliable as the infrastructure itself.
If companies want their cars to communicate with public infrastructures to improve safety, they might also want those cars to communicate with other vehicles to tell each other about traffic accidents or other safety-related issues.
The AI algorithms that run these cars will probably have to be standardised to a certain extent to allow them to decipher information sent to them by other cars. Chandler at Nvidia, one of the biggest players in autonomous brainpower development, says standards might indeed emerge for how you process data that comes from various sensors.
The head of IP at a car manufacturer in Germany agrees, adding that it is important for cars to communicate road defects or how to avoid traffic jams to one another.”
Autonomous driving is still a long way down the road – with in-house sources predicting wide-scale driverless car adoption by 2028 to 2030 – so these standards are a way off as well.
But, given the projected value of the autonomous car market, these potential standards and others are worth thinking about now. Even if we don’t see the same patent wars emerging in the autonomous arena as we are starting to in the connected space, standards will still be an important consideration for automotive companies as they look to develop their IP strategies.