Autonomous driving us over the edge: Are we connected-ready?
Data and the traveling of data via connectivity will enable new mobility machines to communicate with each other, humans, and the outside world. On one hand, connectivity will enable the transfer of data on vehicle performance, safety, and road conditions. On the other hand, the more commercial avenue, data will be streamed about the passenger’s entertainment preferences, daily commute and personal choices.
But connectivity is not something of the future. It’s already a reality as telematics and smart phone technology have shown us. It’s becoming ubiquitous and essential; an extension of today’s mobility, as expected by the consumer.
The Director of the Mcity Mobility Transportation Center, Greg McGuire, currently works with researchers at the University of Michigan who are exploring projects related to mobility and connectivity. He paints the picture of the data being transferred from V2V (vehicle-to-vehicle) and V2X (vehicle-to-infrastructure):
At 8.00 am, 70% of the vehicles drove passed McDonalds and instead visited Burger King. Or during rush hour, at around 5.00 pm there were more inflow of cars in the gas stations in the urban center than in those on the highway. In principle, you could see all these phenomena and quantify it by means of V2X capability. All this information could represent monetization opportunities.
But autonomous vehicles developers aren’t the only ones focusing on connectivity. In fact, industries whose core know-how is telecommunications and networks, are the ones embracing new solutions for the automotive industry’s transition into digitalization.
Barry Einsig, Global Automotive and Transportation Executive at Cisco Systems, shared with me the current efforts to make an intelligent transport system as an-end-to-end solution:
First one of its class, Cisco’s connected roadway network enables an SUV (sports utility vehicle) to be driving down an avenue and already be connected to the next traffic light. The current system, in the market for the last six years, helps engineers understand and deploy connectivity networks. They can optimize security requirements and data exchange capability.
The company currently works with Hyundai (and other carmakers) doing penetration testing for cyber security in the car. It also works with service providers (AT&T, Verizon) to figure out what’s needed to solve connectivity roadblocks. But what connectivity requirements are incumbents working towards? Is their customer, the self-driving companies, fully aware of what will be needed for cars to perform on the road?
According to Barry, Cisco is working on setting up a 5G network system inside the car for WIFI access. It’s a good thing that communication networks are driver-agnostic. The technology could be implemented to existing cars on the road; with or without a human driver. Nonetheless, will the upgraded G-version be able to handle all those 1s and 0s generated by the car’s operating system?
On paper, it appears that connectivity for Level 4 and Level 5 self-driving is pretty straightforward. We will need low latency and as much bandwidth as possible for all the millions of bytes to be streamed continuously back and forth. Surely, our twenty first century networks supporting our Instagram video frenzy and Netflix viewing could handle use cases like data streaming in/out of an autonomous car.
Maybe not.
Austin Oehlerking, CEO and Founder of Boxbot, the self-driving startup (in stealth mode), stresses the importance of connectivity for his specific last-mile delivery use case. He dives into his product’s requirements:
The street bots launched today will either need to be remotely operated or be 100% automated. And since the latter is still years away, for the former to work, high-quality connectivity is required. Connectivity to a network is critical for remote camera control, remote analysis or situation and remote planning.
For robots that will deliver packages only an ‘edge case’ may be required. This means a localized network with a nearby data station. In principle, these robotic creatures wouldn’t need to rely on an upgraded city infrastructure but rather a smaller network hub within their geofenced radius of operation. Nevertheless, the real-time operation remains a challenge.
Within the context of connectivity, I was surprised to learn how much a delivery robot differs from a ridesharing use case. Companies that are planning on rolling out their robo-taxis as early as 2020, are aware of the limitation on available infrastructure and communication networks to transfer data from/to cars on the roads. Waiting for the perfect scenario of smart roads and sophisticated networks may take too long. Therefore, autonomous cars today are being designed to be completely self-contained, meaning the vehicle sees, decides, and responds on their own without any outside input.
According to a recent article by TE Connectivity (TE), advanced autonomous driving will only be possible with more reliable, redundant, and real-time networks. Vehicle safety systems will generate high volumes of data, which will be continuously sent to the cloud. This data transfer will happen bi-directionally as the vehicle will receive information from other vehicles and its surroundings. Data transfer bandwidth will increase to 24 Gbps (from a 150 Gbps for present in-vehicle networks) for Level 3 driving and above.
Currently, the LTE cellular network, used by our smart phones, has a latency of 30–40 ms, making it less reliable and slow for self-driving applications. You probably experienced times when driving through a tunnel or leaving the metropolitan area, and losing all connectivity (bye bye bye signal bars). This is why, the new and more robust 5G cellular network, with lower latency and higher data rates of 10 Gbps, represents sweet words for self-driving cars.
With such major challenge on the horizon, what’s being done in the mobility ecosystem to ameliorate the pain?
Besides the network gurus, startups like EdgeConnex are trying to close the gap in the market. The company places data centers and cells closer to the end user’s point of access. Partnering with telecommunication providers, their value proposition is to deliver digital content at the ‘edge of the network’ with high performance and lower latency at any location.
Tesla, for instance, innately electrified and with a developed self-driving program, already has in place charging stations that could eventually become connectivity and data downloading centers. This versatility gives the carmaker an upper hand in data infrastructure over its competitors. It’s like they had a master plan all along.
We’ve come far from the times of slow dial-up internet service (missing that old raspy modem sound). Regardless, the path to driverless connectivity is still treacherous and dark. Cars need to collect data externally, process it, interpret it and react, all in a matter of seconds. Realistically, today’s data transfer capability isn’t fast enough for autonomous operating systems to rely on, if faced with a life or death driving decision.
Let’s hope that 5G will indeed be our networking panacea.
***
I hope you enjoyed this post related to my book, Autonomousity: Autonomous Vehicles & Emerging Business Models. You can check it out via this link: https://www.amazon.com/dp/B07QDM7HTX
I’d love to continue the conversation about self-driving cars! You can either leave a comment in my Medium page or connect with me via email at BejaranoAPaula@gmail.com or LinkedIn.
And if you’re in the San Francisco Bay Area on May 15th, make sure to check our “AI & The Automation of Moving Goods and People” speaker event.
