Intel, meanwhile, talks about 4 terabytes a day flowing from connected autonomous vehicles and wireless industry advocates praise the vehicle-to-vehicle connectivity that will be enabled by C-V2X and 5G cellular networks. The question remains, though, are cellular wireless connections necessary for autonomous operation?
Even more salient than this question, though, is the issue of satellite connectivity. To get a perspective on this, it is worth assessing trends in the automotive wireless market and the implications for autonomous operation.
First of all, bringing wireless connectivity to cars is a fundamentally unnatural act for the average car company. The product life cycle of a car — 15 years or longer — extends beyond the life cycle of a particular wireless network topology, which is constantly evolving. On multiple occasions, car companies have been blind-sided by cellular network upgrades (analog to digital, 3G to 4G) that have led to the shutdown of legacy networks and the disconnection of cars — with resulting class action law suits, at least in the U.S. — or expensive retrofitting of existing vehicle fleets.
Secondly, car makers are committed to ensuring the safe functioning of their vehicles under all circumstances. Wireless carriers have no such requirement or commitment. The understanding is that the quality of wireless service is dependent and vulnerable to changes in demand.
For these reasons, multiple car companies are preparing to launch vehicle connectivity systems with multiple Subscriber Identity Module (SIM) devices to ensure higher levels of service quality and signal reliability. Multiple German auto makers are specifying so-called “dual-SIM dual active” (DSDA) 2-SIM configurations while some Chinese Original Equipment Manufacturers (OEMs) are proposing 3- and 4-SIM solutions. General Motors’ Cruise Automation autonomous vehicle subsidiary, for example, makes use of a module with four cellular wireless carrier connections.
The truth is that wireless connectivity will be essential to the operation of autonomous vehicles. Not only must maps, traffic, weather, road hazard information, and software algorithms be regularly updated or processed in real-time, the entire industry is facing a long-term requirement for remote control.
The only technology that can deliver a universally reliable connection at high speed and low latency is satellite. In fact, given that most autonomous vehicle developers are planning fleets of such vehicles serving public transportation objectives, the ability of satellite technology to broadcast vital information or commands to vehicles en masse is a huge advantage over cellular.
The reality is that cellular and satellite are ideally suited to work side by side to enable, enhance, and assist in the development and deployment of autonomous vehicle technology. The satellite networks best suited to the task will include OneWeb and SpaceX and the enabling antenna technology is already available.
For quality of service, reliability of signal and volume of data — particularly in a broadcast scenario — satellite technology is tough to beat. The immediate opportunity is for application in testing and development scenarios where cost is less of a concern.
As volumes scale up on the path toward deployment, the rationale for satellite technology will only improve. The conventional wisdom is that autonomous vehicles must be designed to operate in the absence of a wireless connection. The reality is that connectivity for autonomous vehicles will become a regulator requirement.
The debate over the need for wireless connections to enable autonomous operation will likely go on. The reality is that wireless connections between vehicles (for collision avoidance), with infrastructure (for integrating with traffic signals), and with control centers (via cellular and satellite for monitoring and remote control) will be an essential element of the evolution of automated vehicle technology. VS