Processing and storing data, such as visual data, is key to the success of self-driving cars, allowing safety systems to recognise and respond to changes in their environments. However, as these systems become more sophisticated and autonomous vehicles (AV) become more accessible to the wider consumer market, the onus is on electronics manufacturers to produce components and technologies that can process ever-growing amounts of data.

In particular, the storing and processing of visual data is critical in autonomous vehicles, and not just from a safety perspective. Currently, there are a number of systems being trialled, which could enable a vehicle to sense where it is in relation to its surrounding, including LIDAR and RADAR, or visual image recognition. With no clear forerunner, all three systems are currently undergoing trials by some of the industry’s leading AV manufacturers.

Gigabytes of data

Autonomous visual recognition systems are especially data heavy, largely due to the number of images that the car is required to capture in order to operate safely. For example, the amount of data in a single-frame image can be up to several tens of megabytes depending on the resolution of an image. As an autonomous vehicle safety system would typically take several tens of these images every second, it is clear that the amount of data that would need to be stored and processed is significant, with some autonomous vehicles producing up to one gigabyte of data a second.

To handle this data current autonomous vehicles can perform in excess of ten trillion calculations per second, far exceeding the processing power of standard computers.

Until this stage in the developmental lifecycle of autonomous vehicles, manufacturers such as Tesla and Google have had to install large, unsightly boxes on their vehicles, to provide the infrastructure required to capture and manipulate the data required. Obviously, in an industry where aesthetics and design are just as important as performance in the mind of the customer, the size and configuration of these systems is sparking pockets of innovation.

Many automotive manufacturers have been developing chipsets for some time, however these have tended to find use in other areas of in-car electronics, such as engine management systems. To date, most of the research and development around processing data for autonomous vehicles has been left to companies more synonymous with chipset manufacture. With automotive manufacturers (AMs) largely out of the picture when it comes to developing this technology, this has left a gap in the market for other innovative businesses to fill.

BlueBox

One such example is NXP, a semiconductor manufacturer which has developed ‘BlueBox’ (pictured above) a ‘plug and play’ autonomous vehicle chipset which can be fitted to cars. Approximately the size of a shoebox, this product contains all the data storage, sensors and processing tools required by an autonomous vehicle. For AMs, this effectively makes it possible to build the majority of the vehicle, including the design and mechanical elements, and fit the box under licence.

One of the main challenges for chipset manufacturers seeking a foothold in the autonomous vehicles’ market is ensuring that their products meet the required electrical and fire safety standards. The BlueBox is already compliant, so further testing is not necessary after it is fitted to the autonomous vehicle.

For many electronics manufacturers, the rise of autonomous driving has provided opportunities to diversify into different markets. For example, Nvidia, which made its name manufacturing high-end GPUs, however, some of its products and systems are finding a second use in the fast-developing field of visual data processing for autonomous cars. This approach of applying technology into new, often unrelated sectors is a common practice and moving into the AV space and this could bring new challenges for innovators from an intellectual property perspective.

IP strategy

Technology transfer, as it is known, often signals the need to review an overarching IP strategy. For instance, in the case of semiconductor manufacturers, patent applications may have been written in a way which applies to the IT sector. Before applying it more widely, a strategy may be required to extend the scope of patent protection to ensure that the applications cover the use of the technology in these new market opportunities.

Furthermore, in view of the ever-increasing legislative requirements there is a chance that any new patents secured could become ‘standard essential’, which means it may be possible to license the technology to other manufacturers in the sector, in exchange for potentially-lucrative royalty payments.

The development of AVs is providing many opportunities for innovative electronics businesses striving to claim a stake in a fast-growing market.  However, whilst the quest to develop smaller and faster components and processing systems is exciting, they must be sure to protect their innovations as they go.

https://www.electronicsweekly.com/news/design/eda-and-ip/comment-autonomous-driving-seizing-tech-transfer-opportunity-2019-04/