Can human-centered collaborative driving quickly add a bit of happiness to the time we spend in cars?
Getty
It’s been over a year since the death of Elaine Herzberg in the tragic Uber “driverless” vehicle crash, which is perhaps the point at which the hype around driverless vehicles started to cool. Now that we seem to have passed the peak of inflated expectations (Gartner Hype Cycle), the open question is how long it might take to reach a stable market for self-driving vehicles.
In a TEDx talk last year, I talked about many of the complexities to be overcome as part of an evolution of autonomy on our roads and why we may not be heading towards a revolution of robots in our driveways as soon as some may think. In a 2015 TED talk, Chris Urmson talked of the Google Self-Driving Car team’s commitment to keeping his 11-year-old from needing to get a driver’s license. Fast-forward four years, in a recent Slate article, Chris now mentions that it may be another 5 to 10 years before we see limited self-driving deployments in select cities, and 30 to 50 years until we see self-driving vehicles everywhere. Quite a shift in talking points for one of the key leaders in the automated mobility space.
Perhaps the newer, more realistic time frames being discussed are an illustration of the complexities that lie ahead. Self-driving vehicles may be the future, but it is looking increasingly likely that it will take decades, if not the best part of a century, to fully transform how we live and move.
The technology and software needed for self-driving vehicles will certainly continue to mature. However, robots at the helm may never be perfect. Society may look to aviation safety as an acceptable benchmark versus Elon Musk’s current “better than human” mentality for the technology.
How society sets a safety threshold for self-driving cars might have the greatest impact on the time needed to reach a stable market (widespread technology adoption with revenue exceeding costs). There is currently no good estimate on how safe self-driving technology needs to be to facilitate market acceptance. Many people have a fanciful view of seemingly perfect (safe) robots, but from an engineering perceptive, perfection will be difficult. Initially, self-driving vehicles will crash, perhaps reducing risk in some situations (distracted driving, etc.) but also creating new risks in other areas (sensor failure, etc.).
Even if self-driving vehicles could significantly reduce fatalities and major injuries associated with the undertreated health crisis on our roads, the frequency of robot error might just be difficult for society to bear. For some reason, human error is just more palatable. This ethical dilemma is overshadowed by discussions of the trolley car problem (i.e., saving oneself or someone else). Regardless of such unlikely hypotheticals, if a set of crashes were to set off alarm bells like the Boeing 737 MAX, it might take even longer for society to see the benefits of self-driving technology.
In some sense, meeting a safety threshold will all come down to cost. Higher levels of safety demand more sensing, more computational capabilities, greater investments in engineering and more complex functional prove-out, while at the same time likely resulting in a restriction of self-driving technology to more limited uses cases. If the price point for automated driving is higher than manual driving (a fact that will certainly be true in initial deployments and perhaps for years after), companies pushing the technology will encounter consumers unwilling to pay a premium for less – and limits in the subsidies that investors are willing to bear in exchange for future promise.
How long does society need to continue to endure an unprecedented level of tragedy on our roads waiting for an unproven technology, with so many unknowns and an increasing timeline to come to market? Could the explosive quest to develop self-driving vehicles, fueled by a promise safe, low cost, seamless mobility be a path longer than current investors can bare? Alternatively, should we increase investments in technologies to augment driver capabilities that might enhance comfort and increase safety in a more near-term timeframe?
Perhaps our hyper-focus on self-driving vehicles has gone far enough. Taking a page book from aviation, collaborative driving where humans play a role with automation in an engineered system is looking more and more promising. With automation assisting drivers to a greater degree, enhancing safety while reducing the day-to-day stresses of driving could more easily be the first of many stepping-stones on the long path to self-driving and driverless vehicles. Many of the automation-fueled features needed to accelerate these changes are here today and can be enhanced through stronger human-centered engineering that, while complex, offer the potential of benefits far sooner than driverless mobility.
While most of the world’s automakers are working on both self-driving, driverless, and collaborative driving solutions, attention on self-driving and driverless vehicles seems to dominate the headlines. Is this too much focus on the wrong solution? Economics teaches us that sunk costs should be ignored in future decision making, could it be that rebalancing investments and doubling down on reinventing human-centered collaborative driving is the better bet?
