Connectivity is arguably the single most important automotive development of recent decades. OnStar, the first widely available connected vehicle technology, was announced only 20 years ago. Today this technology enables features such as Over the Air software updates, traffic and parking information and is a critical enabler for the deployment of automated vehicles, all of which will be more important as vehicles evolve to become “computers on wheels”. Up until now two-way connectivity has been enabled almost exclusively by cellular networks, sometimes using an occupant’s phone, often using a phone embedded in the vehicle. Now the US federal government is proposing mandatory deployment of a short range (<1km) communication system for vehicle to vehicle (V2V) and vehicle to anything (V2X) communications, primarily to support safety applications. This is being done with the support of the automotive industry, and similar proposals are underway in Europe and elsewhere. The government is advocating a Wi-Fi-based technology (802.11p), but cellular companies are advocating, wait for it, cellular technology. It is instructive to look at some of the issues in this debate which provides deep insights into the needs of automotive safety systems and the business of building cars.
First, a little background, the scenarios the V2X technology are intended to address can generally be classified as short range and low latency. Applications such as Intersection Movement Assistant, which arbitrates among the vehicles at, say, a four-way stop, or a red light warning (or possibly control) system where a signal broadcasts the phase information to a car so it can predict if it will enter the intersection on green or red. The technology has the capability of supporting all sorts of vehicle interactions and addressing many types of accidents, but automakers will need experience with the technology in the field before implementing more complex applications. The key data element transmitted is vehicle position information, the communication system provides high reliability for the messages to get through, low latency, and security, which is extensive but will not be addressed here. The Wi-Fi-based system has been undergoing development since 1999 by the automotive industry in conjunction with various governments around the world.
One of the guiding principles of safety systems is that they have to work. They not only have to work the day the vehicle is produced, they have to work throughout the lifetime of the vehicle. It’s not acceptable for your brakes to stop working after 20 years, or your airbags to stop working after 10 (or when the warranty runs out). The same expectation exists for communication based safety systems, they will need to work 15 or 20 years after a car is built. This must be possible with basic maintenance and no expensive upgrades. Furthermore, there is no precedent for some sort of a fee to keep your safety systems working. Imagine an annual bill from the vehicle manufacturer to keep your ABS operating. Not likely. What this means is that a communications based safety system built into a car in 2020 will need to work with a car produced in 2040, and any costs need to be captured at the point-of-sale. The vehicle manufacturer can’t go back and get more money if they incorrectly estimated the required communications and support costs.
This has implications for the V2V technology we are deploying today. First of all, the technology will not be state-of-the-art. Cellular technology is continually evolving to support new applications and a customer’s willingness to replace the phone every two years. Cellular communications technology that is built into a vehicle, with the vehicle’s three-year development cycle, will be obsolete before the vehicle even hits the road. At the end of the vehicle’s 20 year lifetime, any built-in communications technology will be ancient, but it still needs to work. Vehicle to vehicle communications technology will need to be stable over 20 or more years. Fortunately, the latest and greatest technology is not needed, just “good enough”. Think back to your cell phone from 2000- would it work if you charged it up today? No. That technology is long gone, but a MY 2000 car is still operational (and still more expensive than a brand new phone). A phone from 10 years ago uses a different technology than a phone today, but the central cell towers translate between the various technologies and protocols. Vehicle to vehicle communications use peer-to-peer networks, unlike the cellular system there is no “tower” that can translate. Everyone needs to speak the same language.
Short range communications technology is also used for Vehicle to Infrastructure (V2I) applications which will be talking to traffic lights, roadside sensors, and possibly tolling systems. Roadway infrastructure turns over even slower than vehicles and often has lifetimes of 30 to 50 years. Keeping transportation communications independent of the broader cellular system allows the cellular system to evolve, and also provides assurance that the transportation system will work in the event of natural disasters or other times when the cellular system may fail.
Now imagine a for-profit entity, such as a cellular carrier, in the middle of this mix. A vehicle manufacturer will need to guarantee the service for the lifetime of the vehicle, but carriers don’t guarantee compatibility for 20 years, let alone a price for the service. Anyone with a cell phone will attest to this. Someone must absorb the risk of future price increases and the continued existence of archaic V2X technology 15 years from now. The cellular industry is highly incentivized to provide new technologies every year, and the carriers are very motivated to get as much revenue as possible for their bandwidth. V2V safety protocols are very chatty (albeit over a very short range), and are justified by providing a public good, they probably don’t make sense for a carrier on a strictly economic basis. Cellular carriers are continuously complaining about the lack of spectrum for commercial services and from a purely dollars per gigabyte perspective any carrier will prefer to use the V2V bands for revenue-generating applications such as with 3-D augmented reality and holographic games.
The cellular network is a great accomplishment. The next generation, 5G, promises many advances, not least in the speed and latency of communications. Some of the newly developed technologies would likely improve the performance of individual V2V communications links, but this is not needed as the currently proposed system is “good enough”. In addition, while the individual links may be improved, the overall system needs to be evaluated in an environment with thousands of vehicles, which takes time. There are proposals in place to enable direct vehicle to vehicle communications using the cellular protocols. However these are all just proposals. No one has come close to demonstrating the latencies required by some V2V applications when using a cellular system with vehicles in different cells, nor has anyone describe the business model that carriers would need to support direct, ad hoc, vehicle to vehicle communications where there is no cell tower collecting revenue. No carrier has articulated a business model that makes sense for them to support the government’s proposed safety protocols, and with justification they are not interested in supporting the public good unless there is cash in it, although this does not dissuade the component suppliers who would like to see their hardware mandated into every vehicle.
The current system design advocated by the government is “good enough”. Very much like the early specifications for the Internet, it leaves the applications generally open and proscribes merely that vehicles provide a basic set of data about their basic characteristics, location and direction of travel. This data will probably get much more accurate over time, but it is safe to assume it will continue to describe a vehicle’s location. The basic data needs will evolve very little over time. Probably, in 20 years, vehicles will need to use two radios, one for each of the old and new communications technologies. By then these will be software defined radios and part of a much larger communication suite, but still be providing basically the same information.
This slow evolution of transportation safety communications is not a ideal, but it is a necessary byproduct of the fact that vehicles are expensive and do not get replaced on a regular basis. Some visionaries see that with shared vehicles utilization may be significantly above today’s 4%, justifying the much higher turnover of vehicles. Maybe we will all choose to upgrade our vehicle’s communications hardware every few years. This is speculation. Even then, we still have the problem of the slow turnover of transportation infrastructure. Decoupling the relatively slowly evolving, real time, short range and tactical, transportation communications from the long-range strategic communications that is well supported by the cellular network enables both systems to be optimized for their given business and technological requirements. It is the business requirements that are the often ignored drivers of the industry’s proposed solution. In an ideal world we could all upgrade our cars every two years, but in this world it’s nice to know that my 20-year-old car will still start and get me to work every morning.