12 hours Ago
Cheri Beranek, President & CEO of optical-fiber management and connectivity company Clearfield . A 2023 EY National Entrepreneur of the Year. A future of autonomous driving (AD) may be nigh.
Not to sound too dramatic, but the prospect is pretty exciting. A shift in car capabilities toward AD will drive the development of more U.S.
smart cities , integrating more external touchpoints—roadside sensors, cameras, traffic lights, road signs and other infrastructure—to automate intersections, manage real-time traffic and detect pedestrians, cyclists and other road users. As the network grows, so do the mobility opportunities— robo-taxis and robo-shuttles for those unable to drive, and autonomous trucks and last-mile delivery for streamlined shipping. Fully autonomous driving networks (ADNs) would improve road safety and eliminate the human error behind over 90% of vehicle accidents.
The EU estimates its regulations to advance advanced driver assistance systems (ADAS) and AD could prevent 140,000 serious injuries and 25,000 deaths by 2038. Sound like a future worthy of becoming a reality? Fortunately, the steps needed to get there are reasonable, and technology conditions are prime. To scale, however, these applications need a secure and reliable network bringing these touchpoints together for low-latency, real-time performance—the closer, stronger and more balanced, the better.
Only a few obstacles remain, but from where I'm standing, it looks like we might already know the way through them. Apple’s Update Decision—Bad News Confirmed For Millions Of iPhone Users Today’s NYT Mini Crossword Clues And Answers For Thursday, September 19th ‘Agatha All Along’ 2-Part Series Premiere Review: I Am Cautiously Optimistic We're Closer Than Ever To Getting There Cars today are more connected than ever. Most come equipped with some ADAS to reduce driver input and enhance the driving experience.
Connected cars use sensors, LiDAR and cameras to detect and communicate with other vehicles and IoT devices. In-vehicle software then leverages these connections to provide services like collision warnings, automatic braking, lane assist and cruise control. From Level 0 (no automation) to L5 (a fully autonomous state), most cars that claim self-driving today are L2 or L3—conditional systems that require driver input for more dynamic tasks.
In Huawei's first independent testing of an L4 system, it received a score of 3.8 . Its CEO believes that, with supportive telecom foundation models and digital twins to test and improve, the company is set to achieve L4 , which can predict and fix problems before they happen.
McKinsey has forecast that up to 70% of new vehicles sold in North America and Europe will reach L3 or above by 2030. At the same time, global calls for carbon neutrality and vehicle fuel efficiency and emissions regulations have the auto industry targeting decarbonization . A study published in Transportation Research Part D: Transport and Environment suggests autonomous electric vehicle (EV) adoption could reduce total greenhouse gas emissions from transportation by up to 34% by 2050, and customers want that future.
According to MarketWatch, EV sales are up 60% year over year. McKinsey research found that among surveyed car buyers, 42% have reported wanting their next car to be an EV, and 51% said they could see themselves switching to a fully autonomous vehicle. Obstacles In the Way At the same time, more connected cars are sending more driver data over fiber optic networks for backhaul into the cloud for processing than ever, from approximately 4GB up to 40GB daily —enough to fill the largest available computer storage in five hours.
Wi-Fi, 5G and cloud computing can meet the needs for pilot projects and AD development, but as more cars connect in more ways, they will produce more and increasingly diverse data. EV charging stations and other IoT will also be transmitting vast amounts of data along with a broadening range of industries leveraging the technology, not to mention our connected handheld devices. Generative AI and machine learning can cost-effectively manage such vast amounts of data to support AD at scale , but the massive energy and heat generation needed for those calculations could overload the centralized cloud computing infrastructure and resources, causing delays or service downtime.
Safety is the biggest consumer apprehension AD stakeholders will need to address, but relying solely on Wi-Fi to send so much complex data can cause bottlenecks and life-threatening delays. It also creates new issues about privacy. To accommodate AD applications and support industry growth, the network will require more fiber, security and decentralized processing for a more optimized load.
Solutions Are Already Being Solidified To ensure ADs develop in accordance with consumer safety standards, regulatory support will be critical. The Federal Motor Carrier Safety Administration intends to launch a regulatory approach to AD systems in late 2024. More regional and global entities establishing standards will guide better public and private coordination of network touchpoints and build consumer confidence in AD safety.
Alongside AD growth, the simultaneous advocacy of public officials to include ADAS capabilities in existing driving regulations can lead to a higher penetration of ADAS functions . Auto industry stakeholders will redirect their sales and business strategies. Manufacturers, software providers, and smartphone or device app companies will look to partnerships and R&D for competitive AD offerings and payment plans that garner higher take rates.
Increased public and private interest will set the stage for the rollout of AD and ADNs at scale, but telecom companies must first establish a connectivity infrastructure to manage all of the data. Edge servers may provide a solution. Connected over fiber networks and operating at the site of data production, edge servers can run on local surplus energy, prioritize real-time communication needs and filter out personal data before sending it to the cloud for greater data privacy.
Sending preprocessed data also results in more economical and energy-efficient computing, leaving the cloud free to expend resources on deeper learning between datasets from multiple edge servers—improving AD communication and general AI learning models. Industry experts are currently testing a hybrid edge and cloud computing infrastructure's ability to support AD growth, and stakeholders should be watching. If successful, telecom organizations would do well to consider future-proofing their networks with edge servers leveraging existing or new fiber optic networks in advance of rising demands.
McKinsey predicts ADAS and AD could grow the passenger car market by up to $400 billion by 2035. Self-driving technology is closing in, and consumers are ready—now is a good time to consider ways to seize the opportunity. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives.
Do I qualify?.
