What is a Robocar?

In writing about robocars -- self-driving cars -- I have a particular vision in mind. It is not an uncommon vision, but neither is it the only one.

Here are the key elements of this vision:

1. No new infrastructure: The robocar is able to drive on unmodified city streets, shared with cars driven by people. No new infrastructure is necessary, though some forms of cheap infrastructure (particularly non-physical infrastructure like networks and databases) may be built as desired to make things even better.
2. Autonomous: The robocar works almost all the time by just being given a destination. The occupants do not have to monitor it, though it's not out of the question their advice might be used in unusual non-urgent situations. That means they can always operate safely, but may sometimes need human judgement on navigation or getting out of an odd situation. Thus vehicles can also operate without people in them, to deliver themselves, recharge themselves and park/store themselves. While they might find uses in communicating with other cars or with central city or street computers, this must not be essential to their operation.
3. Safe: Robocars don't get approved for the road until they can demonstrate a safety record a fair bit superior to human drivers, and in fact safer than sober, alert drivers.

There are a few other attributes which are highly valuable, but not absolutely essential:

• Taxis: Some vehicles will be designed to be easily hired out to others as taxis. Eventually this leads to a city of "on demand" vehicles where a cell phone can summon a vehicle in a few minutes and eventually within a minute.
• Mostly elecrtric/non-gasoline: The use of alternate power systems, most probably electricity, opens up new options in car design and allows vehicles to enter buildings.
• Minimally regulated: We want fast innovation with competing companies selling to early adopters. Vehicles will be certified for safety but beyond that the regulators should stay out of the way.
• Social, productive and comfortable: Single passenger cars will be common and act as an office on wheels. Multi-passenger cars will offer face to face seating and a social environment, changing transportation from a waste of time to useful time.
• Connected, but only barely: The vehicles will use external data resources, but primarily through their HQ, not by taking the unacceptable security risk of communicating directly with other cars or infrastructure.

Some may view the safety point as a cheat: I am defining a robocar as a vehicle that's demonstrably safer than human drivers. That diverts an otherwise important question of how safe the vehicles will be and how they will become that safe. I discuss this issue a fair bit, but my main goal is to explore, "what happens when we can make the safe robocar?"

Other Visions

There are other visions of a robocar future, or the path to it.

• Many, for example, think the right way to go is special infrastructure, with vehicles that only self-drive on special highways and lanes, either apart from human driven traffic, or sometimes with it. This philosophy is in decline.
• It is often proposed that we begin with a convoy technology, where a lead car, driven by an expert driver, directs a virtual train of many vehicles which follow the lead. This is easier to do, but I am concerned that mistakes here crash 40 cars, instead of one.
• Some hope to put the robocars on the road early, before they can demonstrate the safety record, possibly by having human oversight. Indeed, a number of people have built robocars and let them drive them on ordinary streets while they watch, ready to correct mistakes. I have also seen proposals for robocars that need regular human supervision, but don't need 100% of the attention of a human. In some visions the human can even be remote, watching video over a network. Many military ummanned vehicles operate this way.
• An early robocar form has more in common with PRT. It runs only on dedicated guideways. The Heathrow Airport Terminal 5 parking shuttle is an example of this, already in operation.
• I regularly encounter different views on whether people will insist on owning their own cars, hire taxis all the time, or own a car but rent for special needs. My vision allows all these modes and it may be difficult to predict what will dominate.