Skip to Content

How connectivity will accelerate the deployment of autonomous mobility solutions

Oussama Ben Moussa
28 June 2023

Why connectivity in autonomous vehicles and robots matters

Across industries, big bets are being placed on business models that leverage autonomous vehicles and robots. The trend seems irreversible – from fleets of drones in our skies to deliver goods, autonomous cars on our roads, driverless trains on our railways, and robots in our factories and warehouses.

It is not hard to see why – the potential economic and operational benefits are huge. According to Pedro Pacheco, senior director analyst at Gartner, “Autonomous drones offer a lower cost per mile and higher speed than vans in last-mile deliveries. When they deliver parcels, their operational costs are at least 70% lower than a van delivery service.”

While  most focus is rightly placed on AI to power autonomous decision making, connectivity is often overlooked. Yet connectivity unleashes the potential of autonomous vehicles and robots to receive information about their environment, acting like a second brain. In this blog, I will focus on connectivity’s role in deploying autonomous mobility solutions and explore this in both a non-safety and safety-critical context.

The four dimensions of the engineering challenges of autonomous mobility

While the range of application areas varies, from a technological perspective, four primary dimensions shape the nature of the engineering challenges of autonomous vehicles and robots:  

  • Mobility scenarios context: the diversity of mobility scenarios, for example weather conditions, road/premise characteristics, or driving rules
  • Vehicle or robot context: the complexity of the architecture and software, the volumes of data generated and managed, and the computing power close to the real-time processing, acceptance, and human interaction with the autonomous system
  • Communication context: Interaction with the environment, for instance other vehicles, local infrastructure, edge, cloud; requires executing real-time information from different sources or updating software over the air
  • Regulation and standards: A safety-critical system must comply with local regulations and converge on standards to accelerate the deployment of common tools and processes.

Safety-critical applications and connectivity

Imagine an autonomous shuttle crossing a complex intersection. Since the shuttle’s sensors cannot provide coverage of the whole intersection and because the intersection is not a completely controlled environment, random events, such as a cat crossing the intersection, could be missed.

Installing a camera and lidar in areas not covered by the shuttle’s sensors would make it possible to detect unreported work, an accident that has just occurred, or an object falling from a vehicle – The perception system makes it possible to detect and flag such situations.

The question then becomes: “How can we take this into account?” The information must be processed and sent to the shuttle as soon as possible to be considered in the trajectory planning to stop, decelerate, or direct the trajectory toward a safe lane. For this, connectivity is key. In this case, we need an offboard intelligence system installed in the infrastructure with vehicle-to-infrastructure communication.

In other complex situations, such as a highway with a high-speed bend, an offboard perception system could be installed on the bend. The perception system comprises sensors, including cameras, lidar, or radar equipped with software breaks. As the embedded perception in the vehicle is not sufficient to interpret all the hidden events, a cloud-based intelligence can serve as a second brain to interpret the scene.

REMOTIS is a remote intelligent autonomous transport system, a real-time self-driving prototype managed by a remote server through 5G communication. Remotis is a disruptive concept for connected vehicles that makes possible the demonstration of new distributed architecture and collaborative services and functions based on 5G technology.

Even better, a fusion of the onboard and offboard perception can help to anticipate dangerous situations or support autonomous vehicles that are unable to stop safely. Here, 5G technology can be useful regarding the latency and volume of data to be transported between the embedded system on the vehicle and the nearby infrastructure / cloud.

Mercedes Benz, at the forefront of technology in the large-scale deployment of the ADAS and AD systems, has recently been granted first authorizations to market ADAS level 3 (semi-autonomous driving) in Europe and the US! Mercedes team said “Car-to-X technologies enable a completely new form of information exchange.”

The E- and S-Class is already equipped for the next step: for vehicle-to-infrastructure communication. Use cases are aiming to inform if the crossing is blocked or in case of a poor visibility…the information passes through a secure server before being shared. The volume of information becomes more important when the infrastructure participates in the perception. Thus, the Federal State of Hesse in Germany became the first region to deploy a traffic light and vehicle communication with warning functions, offering services participating in the protection of drivers, pedestrians but also maintenance workers!

Another example of Safety-critical application is Automated Valet Parking, where the connectivity is mandatory to ensure and secure the service. Bosch Said: “Parking your car yourself is a thing of the past!”. The Automated Valet Parking system generally uses a combination of wireless communication based on the vehicle communication standard (DSRC) and cellular communication dedicated to vehicles (C-V2X).

However, parking a car automatically from a drop off zone in the parking garage to a free place require a continuous collaboration between the vehicle and the infrastructure. The complexity of the maneuvers and the distance between the vehicle and a possible obstacle such as a child emerging from an area outside the field of vision of the onboard system implies that the processing and communication chain must be less than 100 ms to allow the vehicle to brake on obstacles at 1 m. The network must be secure and providing uninterrupted service throughout the parking procedure.

Non-safety critical applications and connectivity

Consider the case of a maintenance application. Pipelines transporting all fluid types, such as oil, gas, and water, often degrade over time. Monitoring and repair can be challenging, especially as these pipelines run across large geographies and often remote areas. Imagine a vehicle with shape/dimensions adapted to this context, moving autonomously and equipped with sensors supervising the ducts and detecting anomalies. This information must be communicated in real time, and/or the data must be unloaded from time to time due to limited storage capacities on the vehicle. Connectivity can ensure continuity of service and high performance. The connectivity solution must be adapted to situations where coverage via a 5G network, for example, is unavailable. A road-side unit ensuring local vehicle-to-infrastructure communication can provide the capabilities to upload data and store it in the cloud.

Capgemini’s concept of a miniature autonomous vehicle named TITAN is equipped with sensor fusion and V2X communication. It can access hard-to-reach zones to ensure the control mission of the pipeline.

In the above use cases, safety-critical or not, the objective isn’t to bridge the embedded systems with edge or cloud by the latest generation of connectivity, but rather to do so with an adapted solution that takes usage, network availability, performance, and cost into account. Designing systems to onboard a technology isn’t the best approach.

The right approach is to select a technology solution from that wide range available today that makes it possible to deploy the right communication system for each use case. Connecting the world becomes a real scaling accelerator.

TelcoInsights is a series of posts about the latest trends and opportunities in the telecommunications industry – powered by a community of global industry experts and thought leaders.

Author

Oussama Ben Moussa

Head/CTO Autonomous Mobility, Digital & Software Engineering (D&SE), Capgemini Engineering
“Oussama Ben Moussa is the founder of the group’s ADAS academy, a training and skill development structure that addresses the needs of dedicated specialists to design ADAS and autonomous cars. He has applied for more than 10 French and European patents related to energy storage, engine performance, depollution, driver comfort, and innovative driver assistance systems.”