Verizon and Cisco test virtualized network for deploying autonomous delivery vehicles and robotaxis

The proof-of-concept project shows that this approach could be a safe and cheaper alternative to roadside units, according to the companies.

Cities could use cellular and mobile edge compute to deploy delivery bots and robotaxis instead of installing roadside units to power the network, according to new research from Verizon and Cisco. The two companies recently completed a proof-of-concept project in Las Vegas to test the idea.

Instead of deploying enough roadside units to build a vehicle-to-everything network, cities could use existing cellular networks, provided they have a public MEC infrastructure. Using existing cellular networks would speed up the rollout of autonomous vehicles, but some government agencies have concerns about the service limitations of that approach.

The Cisco and Verizon test proved that the combination of Verizon’s LTE network, public 5G Edge with AWS Wavelength and Cisco Catalyst IR1101 routers meet the latency thresholds required for autonomous driving applications, according to the companies.

Krishna Iyer, director of systems architecture at Verizon, said in a press release that the project shows the strength of mobile edge compute platforms for connected transportation innovation with much more streamlined architecture.

“This test is a huge milestone in proving that the future of connectivity for IoT applications can be powered by cellular,” he said.

Mark Knellinger, lead transportation solutions architect at Cisco, said in a press release, that this is “huge for roadway operators in that it relieves them of the massive expense of deploying and operating a dedicated vehicle-to-everything environment.”

SEE: Top 5 autonomous car roadblocks

Using a virtualized network for AV communications could be a cheaper alternative to installing roadside units. A 2014 study from the U.S. Department of Transportation found that the average installation cost for a single dedicated short-range communication station ranged from $13,000 to $21,000. These stations are a core component of vehicle-to-everything infrastructure. The cost includes hardware, installation labor and design and planning. The authors noted that equipment costs dropped over the course of the experiment and that prices may fall further as the market grows and as tech specifications stabilize.

Components of a vehicle-to-everything infrastructure

Autoweek describes “vehicle-to-everything” as the all-encompassing term for a vehicle’s connected communications that can help the car navigate the environment and send information back to the communications network. This includes processing real-time traffic information, reacting to changing road conditions and recognizing road signs and warnings.

According to Autoweek, the term also includes other communications, including:

• Vehicle-to-vehicle: Exchange information wirelessly with other vehicles on the road.
• Vehicle-to-infrastructure: Share and receive information with infrastructure elements such as connected cameras, streetlights, signs and lane markers.
• Vehicle-to-pedestrian: Communicate with cyclists and walkers to improve safety.
• Vehicle-to-network: Connect and share data with data centers, road infrastructure and other cars.

Getting multiple stakeholders to agree

Another report on the costs of vehicle-to-infrastructure construction illustrates the complexity of these projects. The report lists these stakeholders involved in a V2I deployment:

• Road operators
• Mobile network operators
• Suppliers/technology providers
• Vehicle OEMs
• Service provider
• User of the service solution (drivers and fleet operators)
• Academic research

Agreeing on a communications architecture is obviously a key component of these projects. This paper analyzes the costs of a cellular network-based communication system, which could be 2G, 3G, 4G and/or 5G, and two direct communication plans that use dedicated spectrum, 802.11p and PC5.

The study concluded that an infrastructure based on the mobile communications network is appealing and cost-effective, but there are also concerns about whether that kind of network can support the safety critical services. The report authors found that “a risk-averse view is often taken by ROs where guaranteed performance for such services can only be currently delivered via direct communications systems.” The concern is that these services “tend to be limited and do not necessarily guarantee the ability to maintain such levels of service across the entire road network.”