We are wasting too much time, energy and money on autonomous vehicles. We know what to do and it's not AVs.
According to Axios, global investors dumped $4.2 billion into companies working on self-driving cars (or autonomous vehicles, AVs) in the first 3 quarters of 2018.
$80 billion. For what? According to Volkswagen, we are still a long way away from true autonomy, and that "driverless vehicles have limited appeal and high cost." From Reuters:
The $4.2 billion figure doesn't include all the funds automakers are investing into developing their own new tech. A Brookings Institution report last year estimated that from August 2014 to June 2017, a total of nearly $80 billion was invested in the area by the auto industry and venture capitalists.
Autonomous cars require high-tech infrastructure, hugely expensive lidar and radar systems, as well as pricey deals with cloud computing and mapping providers, VW’s Thomas Sedran told Reuters on the sidelines of the Geneva motor show.
He compared full Level 5 autonomy to "a manned mission to mars."
You need latest-generation mobile infrastructure everywhere, as well as high-definition digital maps that are constantly updated. And you still need near-perfect road markings,” he explained. This will only be the case in very few cities. And even then, the technology will only work in ideal weather conditions. If there are large puddles on the road in heavy rain, that’s already a factor forcing a driver to intervene.
AVs might also be serious electricity hogs. Peter Fairley writes in IEEE Spectrum that "autonomous driving systems give cars eco-driving skills. But their computers and sensors could consume enough electricity to negate this green dividend." Consumption of fuel is increased by the clunky aerodynamics of the sensors and running all the computers needed to munch the scenery.
For the small and medium-sized equipment packages, going autonomous required 2.8 to 4.0 percent more onboard power. This went primarily to power the computers and sensors, and secondarily to the extra 17-22 kilograms of mass the equipment contributed.
AVs are going to be really expensive for quite a while, too. According to Sedran of VW:
...sensors, processors and software for so-called Level 3 cars already cost about 50,000 euros ($56,460). “We need the technology cost of the sensors to come down to around 6,000 to 7,000 euros,” Sedran said. “This requires quantum leaps in innovation in lidar technology, for example.” Even if this were achieved, the cost of high-definition maps and cloud computing add hundreds of millions of euros in annual costs for fleets of robotaxis or delivery vans, Sedran added.
Remember, we are not even talking about private autonomous cars, mobile living rooms like Steven M. Johnson envisioned, we are talking about filling our roads with Robotaxis. Who asked for this?
We talk a lot about dumb buildings here on TreeHugger, that rely on simple technologies like lots of insulation and good quality windows and eschew smart tech. The same goes for transportation: How about solving the last mile problem by having good sidewalks and safe intersections?
Then there is the e-bike revolution that is happening right now, where better batteries are letting more people go longer distances on more diverse terrains than ever before. How about investing in complete, separate and safe bike infrastructure? A huge proportion of the population of all ages, far more than could ever be served by robotaxis, could use this.
Then there is the dumbest and oldest tech next to walking, the 18th century solution, two steel rails. You don't need 5G navigation and quantum leaps in LIDAR, the rails point in the right direction.
Let's face the reality: We are not going to replace 95 percent of privately owned cars with shared autonomous vehicles anytime soon. Even if we are in the trough of disillusionment on the hype cycle right now, we have a long way to go to get to the plateau of productivity.
It's time to get real, to invest in proven technologies that can work for the greatest number of people in the shortest amount of time. And that is not the self-driving car.