In Part III of our AV blog series, we explored the potential of autonomous vehicles in the agriculture and construction industries. We also discussed the challenges that are preventing them from taking off in the real world.
Chief amongst those challenges was safety, especially as it relates to deploying a fleet of vehicles on site.
It is extremely rare to have a single machine operating in silo. For that reason, communication between all vehicles being used on site is an absolute necessity—regardless of whether they’re the same machine or not. Having the ability to communicate speed, location and heading are all impertinent to managing a fleet of vehicles for certain operations.
This is where vehicle-to-vehicle (V2V) communication comes in. While still nascent, V2V is quickly being developed to enable vehicles to wirelessly exchange information.
A technology at an inflection point
V2V communication technology is nascent and continuing to be developed, but it is not new. Global V2V communication market has already exceeded $10 billion as of 2021, and is projected to exceed $50 billion by 2028, according to QualiKet Research.
Figure 1: Global Vehicle-to-Vehicle (V2V) Communication Market Size (US$Mn 2017-2027
Certain branches of V2V technology are fairly mature. One such example is Dedicated Short-range Radio Communication (DSRC) which enables transmitting basic messages between vehicles and is mature enough to function in the Safety Pilot. In fact, the National Highway Traffic Safety Administration (NHTSA) is proposing to issue a new Federal Motor Vehicle Safety Standard that would require all new light vehicles to be equipped with DSRC devices, enabling V2V communication.
However, there are still a number of technical challenges within the V2V sphere that remain unsolved. Conquering these challenges is critical to bringing a fleet of autonomous vehicles to the industrial world.
Figure 2: V2V Communication Challenges by ResearchGate
Unplugging safety-rated protocol and network
Compared to its wired counterparts, a wireless autonomous or remote-controlled vehicle requires a different and more complex network topology. Moreover, the underlying types of connectivity—Bluetooth, Wifi, and LTE network—require a different network topology to operate the vehicles and robots.
One company rose to the challenge: Fort Robotics, which raised $13 million Series A earlier this year led by Prime Mover Labs. Fort offers remote controls with built-in safety features, such as certified emergency stop and drop sensors, as well as a wireless communication system that ensures reliable safety commands and secure data pipelines. Operators are able to command a fleet of vehicles and robots while ensuring they could put a stop to all the machines if they want, when they want, and without fail.
Solving the curse of distributed knowledge
A multi-vehicle system requires mutual awareness amongst the vehicles. Each vehicle needs some knowledge about other vehicles on-site, in addition to their local knowledge.
This awareness is sometimes based on the provision of a shared knowledge base for the entire team. The knowledge base, which typically would be implemented as a distributed replicated knowledge store, contains the concepts, objects, and relations known to the vehicles, as well as the fused perceptions of the state of the execution environment.
This is a very resource-consuming and a performance-critical element of the V2V protocol;
the vehicles must continuously access thousands of objects and relations within the shared knowledge base in order to process and make sense of local information.
While much research has been focused on knowledge representation techniques (i.e. teaching the vehicle and robot to adapt to the environment), little has been published specifically on distributed knowledge bases for multi-vehicles systems.
Breakthrough technologies that successfully solve the distributed knowledge problem in a resource-efficient manner and for performance-critical elements could be a good investment opportunity to look further into.
Mobility, latency, and data transmission speed
The high speed nature of autonomous vehicles make latency and data transmission speed a much more acute problem. Even microsecond delays have the potential to cause crashes or accidents.
When a vehicle communicates with a hub, it is communicating with known elements in terms of location and network topology. However, in the context of direct V2V communication, vehicles face other problems due to high mobility, unpredictable network topology, and intermittent connectivity between vehicles due to constant changes in their velocity and direction. Two vehicles could also be moving in opposite directions of each other, reducing the amount of time needed to establish connection links with other vehicles, requiring fast data transmission.
Long-range communication could also create low latency and reliability among vehicles if a vehicular communication network is warranted, which would result in information loss and transmission delay. This calls for faster, short-range communication frequency, such as radio (or 5G), or hybrid architecture that provides both cellular and short-range communications, efficient radio resource management, and effective information dissemination.
Specialized middleware support
A multi-vehicle system, like any other distributed system, benefits from middleware that hides the complexities of distributed computing in heterogeneous environments and, thus, eases the job of the developer of a distributed application. In general, the middleware needs to simplify the application design by making transparent the low-level details of the underlying hardware, software, communication, sensing, and actuating activities.
There currently does not seem to be a specialized middleware for a fleet of vehicles, which could be an interesting investment opportunity also.
We still have a long way to go
Autonomous vehicles have come a long way from science fiction. Both hardware and software components are now relatively mature. Big Tech companies are sitting on a treasure trove of data that made it almost impossible for new, underfunded entrants to compete. Regulation, however, is still playing catch up to technology. This opens a vast opportunity for autonomous vehicles to find applications within unregulated terrain, particularly in outdoor and industrial use cases.
As the number of autonomous vehicles grows, they inevitably need to interact. The challenges within the V2V communication represent yet another treasure trove of investment opportunities. We still have a long way to go.
And for a venture capitalist, a long way is generally a good way.