DOA. That’s my prediction of how autonomous vehicles (AVs) will be accepted in the marketplace when they’re formally introduced. From surveys I’ve seen, the majority of people don’t really want self-driving cars. Of course, a segment of the population is verbally supporting AVs in surveys and opinion forums, but will these people actually buy one? I will be surprised if sales ever exceed 1% of the total auto market. And after all the publicity related to the two recent deaths related to self-driving cars, that anti-AV demographic will probably grow larger.
You’ve probably already heard of the recent death of a woman pedestrian in Tempe, Arizona by an Uber AV test Volvo. That’s one sure sign that AVs simply aren’t ready for the real world. It’s amazing that the LiDAR, radar, and cameras did not detect her. Was it the fault of the sensors, the algorithms, or what? My impression is that the sensor and AI technologies were better than that. Maybe the victim was actually at fault?
The test driver wasn’t paying attention either, and that may have been a factor. But it’s difficult for a backup driver to maintain a high level of attention for any length of time when the car is on autopilot. Hopefully we will learn from this tragedy. In the meantime, some but not all companies are suspending AV tests until some conclusion is reached.
Another recent death in California also indicates that the current AV systems may not be ready for deployment. This death was in a Tesla Model X SUV with the Autopilot in self-driving mode. It may have been malfunctioning, but it did apparently give the driver some warning that wasn’t responded to. Maybe the driver was at fault by not responding. I can predict that most AV owners will not be paying attention as they should be while riding. A government investigation will hopefully reveal the problem.
Still in Hot Pursuit
In any case, two deaths aren’t enough to stop the aggressive development of AVs. Auto manufacturers, technology, and taxi companies continue with their hell-bent effort to develop the definitive AV. Driving safety is the ultimate goal here, but the whole thing is really more of a competition to see who gets there first with the best features and capabilities. And patents. And riches.
Billions of dollars is being spent on something that most potential customers don’t seem to want. It strikes me as something we’re doing just because we can, and then thinking up the rationale later. In any case, there’s maximum momentum right now, and we should see some commercial self-driving vehicles in another year or so.
Most of us will probably not buy an AV. Instead, we’ll increasingly embrace the excellent advanced driver-assistance systems (ADAS) being incorporated in most new vehicles. These can and do reduce accidents. The addition of V2V and V2X radios are expected to further improve safety. In my opinion, the combination of ADAS and V2V/V2X plus an alert human driver is probably better than an AV to improve safety.
The whole point of the AV is to eliminate human error in driving. But will AVs really reduce accidents? The 2016 Federal data, the most recent available, says that over 37,000 persons died in a vehicle accident. AVs, as they say, will eliminate most of those accidents caused by human error… we hope.
The human is fallible, of course. However, I believe the AI and sensor technologies in the self-drivers are also fallible, but in a different way. As I see it, AVs will reduce some types of accidents, but will introduce another new offsetting class of accidents with conventional vehicles and pedestrians. Developers will do their best to design the best hardware and algorithms, and rigorous testing will get the bugs out. But as you know, nothing is perfect.
I can’t get over the feeling that we’re abdicating too many of our responsibilities to robots. I’m not against automation, per se, in applications like manufacturing and materials handling. But for some things, I question it. Are you ready to entrust your life not only to the complex mechanical and electrical structures of a self-driver, but also to unique software algorithms of artificial intelligence?
Automation makes us lazy and minimizes the need for personal thinking. Technology seems to be dumbing us down. But that seems to be our present and our future. I just can’t seem to get a better attitude about what apparently is inevitable.
In spite of the rosy outlook and hype of driverless cars, we will continue to buy our new SUVs with their ever-more-efficient internal combustion engines and ADAS. The AV technological developments in sensors and machine learning are spectacular and will find uses elsewhere. And, needless to say to most readers here, the whole AV movement is a big benefit to the electronics industry. That’s good for all of us. Maybe I should just shut up and get with the program.