Earlier this year, Cisco set out a number of predictions for “connectedness” in the automotive sector: Namely, it said the horizon for driverless cars is likely to be 5–7 years, as by then it will cost more for us to drive our cars than to let them drive us. Cisco also noted that traffic management will emerge as a primary “killer app” for the Internet-of-Things.
We have likewise touched on traffic management concepts in these pages and would tend to concur that, indeed, roadway applications are very ripe for the picking as an Internet-of-Things venture, if for no other reason than to improve the quality of life for daily commuters. However, when it comes to the tipping point of driverless cars becoming a mainstream reality due to cost, we at KORE would choose to protract the timetable a bit. Even with the technologies bearing down fast that would shore it up, it still may end up being farther away than cost considerations alone would lead us to believe.
To be sure, the “Driverless Car” as a concept has been with us since at least 1970, commensurate with the opening of Epcot Center in Orlando, Fla. Among those initial ideas, the notion was essentially a “fly by wire” approach, where the vehicle would follow unseen tracks using magnetic or induction technologies. These approaches are now commonplace in large industrial logistics and manufacturing environments, where sleds of parts move from one place to another on defined routes.
We also now see, essentially, pilotless aircraft; cockpits generally manage almost all matters of flight from origin to destination – EXCEPT for close contact, on-the-ground movements. In the analogous driverless car scenario, recent telematics advances are certainly making great strides in controlling open road driving – lane departure, active radar cruising, automatic stopping and even GPS linked to maps for cornering. These are huge leaps forward, but they are open-road systems, where little supporting infrastructure is needed or exists. Close contact, on-the-ground movements do and will continue to require humans at the controls.
So, while many building blocks are already in place in the automobile to make it “driverless,” in order to really get to a true, self-driving environment, we still have to establish (1) substantially more vehicle-to-vehicle as well as vehicle-to-infrastructure communication; (2) substantial leaps forward on security of these communications to prevent intrusion; (3) notable regulatory changes for liability, insurance, and the like; (4) wholesale human behavioural change for adoption and (5) notable investment in the fixed infrastructure for sensors, with proper hand-offs, standardised interfaces and many further considerations to be mapped.
Even then, we must bear in mind that the average age of on-road automobiles in the USA hovers around 11 years, so these systems will only be effective for a tiny proportion of the population as a whole. We’re not going to retrofit all those cars, so it becomes as much a question then of economics as it is of technology. Who is going to pay for that infrastructure, what are the tangible benefits to be realised, and when?
We do of course believe that the many “Driver Assistance” sensing and in-cabin infotainment applications will continue to evolve: these are technological increments and are important differentiators for the auto industry. Tesla is already changing the perception that a car can be far closer to a computer on wheels than the mechanized analog system it’s always been. But bridging this to an unaided, driverless experience is not 5 years – maybe not even 15 years – away (a fact that may actually be comforting to the petrol heads among us!).