Tesla has yet to start testing its robotaxi without driver weeks before launch

If we create dedicated tracks for these buses we could make them go even faster, and automating them would be so much simpler: you wouldn't even need steering! Someone should look into this.

The answer is, whenever there's enough money to be made there to support a self driving company's expenses to develop and operate it. These programs cost billions to develop, and public transit is likely not a profitable market niche at the moment.

Also keep in mind that the driver also is responsible for handling fare evasion and ensuring the passengers stay at least somewhat ruly, which a self driving vehicle might not be able to do.

I’ve seen buses have to threaten to cause an accident to force someone to make space for them to merge. That’s gonna be a risky bit of code.

> It seems like a much easier thing to do. They drive on fixed routes and do not have to handle complicated parking situations.

- transporting people has higher legal requirements

- "complicated parking situations" is IMHO not really the problem, most parking situations are relatively straight forward. The problem is complicated driving situations. The fixed routes make it less likely to have, but the bigger size makes it more likely. Bus drivers needing to spontaneously adopt there route is rare but a thing, but having a hugely long car which isn't allowed to drive backward is making that more likely

- people, Bus drivers aren't just driving the bus but also tend to handle the passengers in various ways. The simplest cases are ticket payment (not always) and people asking questions (common especially with older people). More subtle ways include things like giving people the feeling they are not alone in the bus even if they are the only passenger (i.e. subtly reducing bad behavior) telling a passenger that they can't take a bicycle, motorcycle or for the supper crazy kind on fire grill into the bus. Or stuff like calling the police if they notice a passenger getting threatened or assaulted. etc. This category kind applies to any "at same time shared autonomous vehicle" is one of the more "open questions" of autonomous public transport. Depending on country, culture and are I think the difference can be anything from irrelevant to wupsi we have a huge problem.

- Insurance, a lot of autonomous care providers try to doge responsibility, i.e. they drive but you are responsibility for how they drive in most situations. That isn't exactly inviting for bus companies sure there probably is some insurer which is willing to handle it, but then the moment you hit the legal gray area between care maker or insurer has to pay things will likely get messy and that gray area is AFIK currently still too big.

so IMHO less of a pure technical and more of a legal/people issue

Deployment might be easier but the opportunity would also be a lot smaller. Taxis have somewhere close to a 1:1 relationship between driver and number of passengers. The financial impact could be huge. Buses are probably somewhere in the 30:1 ratio (obviously this varies widely), the cost of the driver relative to number of passengers carried is much less.

The trouble with buses is that they're very big and dangerous; the worst case scenario is very, very bad, and bus drivers are better trained and more regulated (in particular around age, sight, health conditions etc) than normal drivers; the standard expected is higher.

Here's what can happen when things go wrong: https://www.dublinlive.ie/news/dublin-news/wellington-quay-b... (a bus mounted the pavement behind a parked bus and hit a group of people, killing five). In a similar accident with a car, at bus-like speeds, you _probably_ wouldn't expect deaths.

Why would you do that, a bus driver costs $10/h and the bus costs several $100k even if it's NOT self driving. The cost of the driver must be miniscule in comparision, not to mention if the self-driving bus and associated insurances will cost many more factors of $100k...

Having ridden on busses countless times in SF, autonomous busses seem way off. The situations that busses have to navigate can be incredibly difficult, and the vehicle is not as nimble as a medium sized automobile. Still, improvements could be made for the autonomous busses, like giving it a dedicated lane like they did on Van Ness. At that point though, why not have a light rail? Autonomous train systems have existed for over three decades. https://en.wikipedia.org/wiki/List_of_driverless_train_syste...

I talked to one of the higher ups at the VBZ which is the transport company of the city of Zürich Switzerland.

They do not see it happening anytime soon not even for light rails (trams). The primary issue is during busy hours the driver has to take calculated risks or they would not get anywhere. People will just block you. You cannot get an AI to do that or convince the public that they may be run over by an AI. How aggressive so you set the AI? Where do you draw the line?

What they do see is having AIs assist drivers where it will do emergency breaking or hint at dangerous/inattentive people some of which they already have in the new trams but they can switch it off of they need too.

> I wonder when buses in public transport will become autonomous.

Once you don't have to pay a human driver salary, it won't make sense to use such large and expensive vehicles as buses.

> It seems like a much easier thing to do. They drive on fixed routes and do not have to handle complicated parking situations.

Route-finding is the easy part of self-driving, it was solved basically as soon as we had GPS maps.

The hard part is avoiding hitting anyone, and anything, on the road while do you that.

I'm not at all sure either way about parking. We have parking sensors, but if parking is a "solved" problem or not, I do not know.

> I wonder when buses in public transport will become autonomous.

Test projects, but anyway:

https://www.letsholo.com/oslo

https://www.toi.no/transport-and-behaviour/autonomous-buses-...

Fixed routes make the easy part easier, you need a really good gps for autonomous driving. However having driven VW autonomous vehicles, the hard part is that driving means interacting with other drivers. For starters, just standing there and switching on hazard lights is usually a not disastrous idea.

Disclaimer: I work for Moia though no longer with AVs.

In my previous city (near Paris), we've had one experimental autonomous bus (exactly one, as far as I know) online for years. I have no idea whether it was considered a success or a failure, but it seemed stuck a bit in limbo: it worked, but I guess not well enough that the city planned to expand its use.

China has had driverless buses for some time:

https://youtu.be/89djfMaQWZw

“The future is already here – it's just not very evenly distributed.“

It's much easier, secure and cheaper to pay a bus driver, it's a political issue not a technical one

I'd be pretty surprised if China wasn't already figuring this out.

There aren't even many autonomous trains, although you would think this is the easiest to do. In Tokyo the Yurikamome started operating about 30 years ago (it is fun to ride at night on the front seats when Odaiba is almost empty), and in Germany there are two autonomous subway lines in Nürnberg. I wonder why transition is taking so long, considering that the shortage of train drivers will become much worse very soon.

Switzerland had for 5 years between 2016 and 2021 a few autonomous regional buses in test drives, with only one incident (a broken glass). At some point the test was stopped and the results buried, without any explanation or followup.

I think it's rather likely that Waymo has not recouped its initial investment, and even if you completely ignored that I'd be surprised if it doesn't currently still operate at a net loss.

The whole situation is artificial and propped up by external funding right now. Making this a viable industry is going to take a lot more time, if it's possible at all.

It is only natural that capitalists want to play it like that. Put the burden of safety on society, etc.

This is why we should have laws that say that even a small software update requires testing by an independent organization for at least several months.

"Move fast and kill people."