Presumably driverless taxis would need road pricing as fleets of full size saloons carrying even 5 pax would take up far more space than a 10 metre long double-decker carrying 50 to 90 pax. On the other hand, if uber drivers were replaced by AI, would they be programmed not to park on cycle lanes/bus stops/roadwork narrowed road junctions? I ask because, around 1320 today I saw EXACTLY that; an uber taxi parked on the cycle lane on Store Street (under M'cr Piccadilly Station), it then came out of Store Street, "cruised" to London Rd/Fairfield junction; stopped, let a passenger out on the junction including taking time to remove luggage from the boot, before turning left onto Fairfield Street and stopping on the 219 bus stop. Of course, a 219 just happened to becoming along and the driver had to pick up around 10 pax in the middle of the road, not only because this uber taxi was parked alongside the shelter, but another private hire taxi was parked right on the bus stop flag, so she couldn't pull up beyond the stop to avoiding holding up the traffic.
Good points Dentonian; my view is that tricky issue with driverless technology will be whether the same roadspace can be shared by driven and driverless vehicles; since the drivers of the driven vehicles - especiall in the case of Uber-type cars - may tend systematically to exploit the safety features of driverless cars to assert priority at road junctions. If so, riding in a driverless car would consist of a series of emergency stops; as manually-driven cars continually pull out of side roads, overtake and cut in, turn right across the traffic stream, and do U turns in the middle of the road. Much urban driving is 'give-and-take'; but having vehicles pre-progarmmed only to 'give' may well result in counterpart manual drivers adopting n 'only take' approach when in the presence of a driverless vehicle. Which in turn implies more rigorous enforcement of road traffic lawsl; or otherwise simply banning uber cars altogether from urban areas where driverless taxi services have been implemented. In my view, these issues would be much less problematic if only bus services are driverless. Designated bus routes can be supervised throughout with CCYV; so that dangerous behaviours from manual drivers can be detected and punished.
All of which is a special case of the general principles for the economic and functionality of driverless vehicles. Much of the more extravagant claims of proponents of driverless technology rest on confusing these underlying principles.
Commentators talk about the five stages of autonomous functions in road vehicles (though confusingly there are actually six).
stage 0: the standard car with no specialised autonomous functions;
stage 1: autonomy in particular individual processes or actions; e.g. cruise control, ABS braking, parking sensors;
stage 2: autonomous capability for complete funcitons involving multiple processes; e.g automatic parking, automatic emergency stopping;
stage 3: fully autonomous, hands-off capability for most trips in their entirety; but requiring manual control as a continuous back-up for occasions that the autonomous control technology is incapable of dealing with.
stage 4: fully autonomous, hands-off capability for the entirety of every trip within the 'Operational Design Domain' of the technology. Maybe; only along certain roads; only in the absence of severe weather; only as directed by permanent road signs and signals (hence lacking the capability of responding to contrary verbal instructions from a designated police officer or traffic official; with the necessary additional capability of distinguishing such designated persons from bystanders)
stage 5: fully autonomous, hands-off capabilty for any trip within the region of coverage; in all domains and in all circumstances where a manual control vehicle would be capable of functioning. At stage 5 - and only at stage 5 - driverless vehicles are produced without manual controls, even for backup.
The inescapable factor determining the economics and potential of driverless technology, is that while the big step-up in functioning is from stage 3 to stage 4; the big step-up in technology and cost is from stage 4 to stage 5. Most commentators now accept that stage 3 functioning is so potentially hazardous that it is likely to be legally banned. A stage 3 'driverless' technology that hands itself back to manual control at exactly the moment when driving attention and skill is most needed, is likely worse than useless. But in technological terms; stage 4 operation is simply stage 3 operation with all the bugs removed; and the predictable hazards fully accommodated in advance. It is universally agreed that almost no features of stage 4 tecnology could be re-used for stage 5 operation. Consequently, almost all current investment programmes are only directed to achieve stage 4 funtioning; with the unspoken assumption that stage 5 will be unachievable (or if not, will be legally inhibited; so that those investiing in stage 4 technology will be protected from losing their investment).
But the key to stage 4 is the prior identification of the 'Operational Design Domain'. In practice this will mean that only some roads on the network will be legally capable of driverless operation; your driverless taxi will take you from door to door along these roads; but beyond them, you will need to use a driven vehicle (or walk). Maintaining a road within the Operational Design Domain will be costly - Google resurvey every road they drive down every day with manually driven vehicles, and it is not unlikely that will be the legal minimum. Which is one reason why driverless buses are likely to be favoured economically over driverless taxis; it is much easier and cheaper to maintain the city bus routes (several hundred miles perhaps) rather than an entire city road network (tens of thousands of miles).
In all of this the issue or marginal costs and benefits will be key; will it be worth the extra cost of maintaining an entire city road network in the ODD, for the additonal functionality provided by driverless taxis? will it be worth the massive technological cost upgrade of moving from stage 4 to stage 5 operation; simply for the occasional additonal functionality of driverless trips in thick snow (say); or down a dirt track?
