Autonomous Driving in 2025
I’ve written this to document my ideas for where autonomous driving is heading and what companies and regulators need to focus on. For those unfamiliar with autonomous driving I make reference to the 5 levels of autonomy, these are graphed below, the rest should be pretty self explanatory. I will caveat this by expressing that I am an optimist to this technology, what I plan to identify are the ways companies should be focusing their attentions to get the technology to a level of regulatory approval, breaking down the psychological barriers of future consumers; and for safety to be a key focus area in what are the most risk prone development years. We have seen this many times in history, from steam engines exploding to plane crashes, where technology development has been stalled, therefor stalling progress.
Let’s start out by looking at the key players in the market in all autonomous systems operating in the outside world today.
Uber — Uber Advanced technologies Group found themselves in a bit of hot soup last year after killing a cyclist. This sent shock waves through the autonomous driving community that big tech companies were letting untested and risky technologies out on the open road.
The approach Uber has taken, like many of the other large tech firms, is to build fully autonomous systems which are able to navigate multiple scenarios and environments. This allows for the most direct approach to full automation. The company is operating level 5 automation and needs a driver present in the vehicle as a security measure.
Waymo — After spinning off from its parent Google, the company has taken a very similar approach to that of Uber in developing fully autonomous systems. Since beginning its pilot program in Phoenix, Arizona in April 2017 the company has recently begun charging for rides. The company chose Arizona due to its approach to regulation and their all round good weather giving the vehicles LiDAR technology the best environment to operate in (we do not want to see this running in any polar vortices just yet). The company has partnered with a number of OEMs to manufacture its vehicles. The company is operating level 5 automation and needs a driver present in the vehicle.
Cruise (GM) — Unlike Waymo, Cruise is all about doing the full stack. The company is pursuing a different market, currently that of SF, however, the companies are bound to clash heads in the coming years with the addition of Uber (pending further accidents) which will most likely lead to the biggest winner of this three horse race being the consumer. The company is operating level 5 automation and needs a driver present in the vehicle.
Tesla — Whether you’re in camp Elon or not we cannot argue that Tesla has been the driving force of vehicle electrification over the past ~15 years. The company has also made negative headlines more than once in the past few years #fundingsecured; Reading in the Passengers Seat While Driving, and my personal favourite, Tesla Driver Asleep For 7 Miles. These headlines are concerning for any driver and it seems that Tesla has a battle with its customers about how to use the autopilot.
I recently test drove one (research purposes, I wish I was a customer) and the autopilot in the UK needs to have a hand on the wheel and if it doesn’t not sense they’re there the autopilot will deactivate after 15 seconds. These are some of the small steps the company has taken towards safety. Tesla is still operating level 3 automation.
Navya — The French company has been building campus and airport type shuttles that operate in a controlled environment. The buses are the first instalment of what some very few members of the population have begun to experience.
If you park your car at at Heathrow you can jump into an autonomous shuttle to take you to Terminal 5, or if you work at Google one may take you from one building to another within the confines of the campus. These are able to operate in these spaces due to many factors, but primarily the regulation on private land is minimal and the pedestrians and risks associated are mitigated by the speed at which they move and the lower traffic density achieved in these environments. These are level 5 autonomous vehicles.
Nuro — For $5.95 you can now have your groceries delivered to the front of your house in Scottsdale, Arizona. The company has been in development for some time now and beginning to roll out near Phoenix (like Waymo, good weather/regulation for autonomous vehicles). The automated vehicle commutes between the store and the home delivering peoples groceries ordered online. We are beginning to see these more and more and how they will interact with our day to day activities holds strong promise for the future. A similar company is delivering Pizza at the George Mason University, Fairfax, as well as Amazon investing heavily in this space. Just plug your pin in and grab your groceries when the robot opens its doors.
Renault EZ-Pro — Renault has taken a slightly different approach and to help consumers have a better experience the company has included what they call a concierge. This employee is able to support the driving should there be obstacles or potential hazards as well as supporting the delivery with not only last mile but last inch. There is no date for launch of this from Renault, however, they have said that this will at least have the highest level 4 autonomy to allow for the concierge to only intervene if absolutely necessary.
The first three examples I give are the biggest players in the US, yes there are others here in the UK such as FiveAI and Oxbotica but these are much less funded than their counterparts across the pond. As mentioned they are all pursuing the same market, that which Uber has been exploiting for the past decade or so. The three companies are looking to roll out Level 5 automation in the coming years and some are closer than others. The race is on and the first to market is going to be in good stead to rapidly grow, however, the competitors arrival will only be a benefit for the consumer in the end.
The regulators are a key part of this process and the longer they take in deciding how and when these cars will be allowed to roll on our streets will be the only thing the investors of these tech giants should be worried about. The consumer will also need to get their head around what is going on here and it seems that people are still relatively hesitant to fully self-driving cars.
Tesla and other OEMs working with level 3 automation and up are taking the slower approach, this however is causing disruption in the market due the amounts of headlines these companies are getting by the few drivers misusing or depending too heavily on the technology which is causing the company unwanted attention. As cars begin to roll out additional features like lane keeping, and automatic braking the drivers become more and more dependant and begin to lose attention on the road. How they combat this? We will hopefully soon find out.
The last category, the delivery vehicles and low speed controlled environment vehicles is the one i’m keen to explore further in the short to medium term (5 years) this category has lots of interesting players in the market and the controlled routes at low speeds will be what gets consumers accustomed to autonomous vehicles on the open road. By getting into an autonomous vehicle on a designated route at an airport is much more doable psychologically for a consumer. This may be the key stepping stone required to have them become confident enough to head out onto the open road. Consumers are the key element of this — Silicon Valley can build all the technology in the world but if users do not feel comfortable around it, it will be useless. That being said, the future looks bright for autonomous driving — whether you believe it’s in 5 years or 20 years, we can all agree that the robots are coming.