more on autonomous cars – how long could it take?

A few weeks back a friend and I were talking about the prospects of a certain connected car product, and got into a broader conversation about autonomous cars and V2I systems, and when they may come to fruition and where. I was feeling, well, grumpy, and posited that we would both have grandchildren bouncing on our knees before we saw a city-scale (DMA scale), robust, broadly adopted connected car / V2I system *in the United States*. Upon further thinking we suggested top-down city-states – Dubai, Singapore – as perhaps uniquely suited to implement a full connected car / autonomous car infrastructure first.

Let’s pause to define robust, broadly available V2I – here, I’m defining it as DSRC wireless infrastructure, not just deployed as infrastructure, but with the requisite radios also available in the cars, across makers and models. This is what would be required for low-latency, cross-model information between cars. And this is a major reason why leaving autonomous driving or even quasi-autonomous driving (let’s use NHSTA Levels 3 and 4 for discussion purposes) to the market may not be enough. That, and who will pay for the infrastructure? A city? The government?

LTE networks are already deployed and thus leveraging LTE would save the capex of rolling out local wireless infrastructure through a hodge-podge of federal and state money. However, for latency reasons, doing so would likely necessitate the ability to hairpin at the LTE base station, rather than go back to a switch. Also, in this scenario, do all cars need embedded LTE radios? Could you rely on driver smartphones with LTE to share information between cars? In that case the phone would need to further interoperate with the car to share information for autonomous driving purposes. I struggle with this concept. It feels like a lot of variables need to cooperate, and reliably.

A relatively recent innovation in autotech is the hybrid engine, which, admittedly, is not new at this point as Toyota alone has been working on them for 20 years. This by itself should tell us the rate at which new autotech innovation can enter the overall consumer vehicle fleet.

Screen Shot 2015-07-22 at 10.29.40
The overall US market for automobiles for consumers is between 16-17M units each year. As the above chart shows, hybrids – all models combined – sell about 500,000 units per year, or about 3% of the total market. Is this the rate at which autonomous cars would penetrate the market? By the way, this leaves aside the issue of regulatory regime. California’s rules temporarily promoted hybrid / EV ownership as a way to get the carpool lane – could similar rules promote autonomous car ownership as a means of getting the self-driving car lane?