Austin's $10 Billion Expanded Rail with Subway Plan Approved

[mhays]

Pretty simple. If tech can drive a car, maybe it can also drive a train that crosses streets.

[Busy Bee]

^ I suppose that's in the realm of the possible but automobiles are ALWAYS going to have a human operation override... ALWAYS... Which means there would still be an achilles heel in the safety alert "loop" if a person is in the control and fails to be notified for whatever reason of the oncoming train or for whatever reason purposefully ignores it. Even if such a system could be constructed it gets me a tad irritated that onus of any kind would be on the train, which ideally should operate in on a closed grade separated right-of-way. Trains don't run into vehicles, vehicles (and pedestrians) get in the way of trains.

[Busy Bee]

Originally Posted by mhays

Pretty simple. If tech can drive a car, maybe it can also drive a train that crosses streets.

Automated trains already exist, have for decades. What exactly would it do differently? Is it supposed to yield to vehicles in a crossing or something? I'm not being flippant, I honestly am unsure what it is you're pitching here.

[mhays]

Automated trains can't currently cross streets. That type of train needs a human driver.

[Busy Bee]

If what you are proposing is an automated control system for trains advanced enough to avoid collisions at grade crossings by either communicating with "smart" partial or fully autonomous automobiles and/or automatically braking when barriers are tripped by an unlawful crossing vehicle or pedestrian... then I guess I understand. It does seem like a bit of a roundabout solution to what is ultimately a problem with rail right-of-way not being fully closed and separated by grade with other modes. And like mentioned it doesn't address human error when an intruding vehicle disregards all pronpts and warnings. Sime sort of "big brother" override that takes control of the vehicle and governs its speed or braking is never going too happen in this country and many others, even if practical or technically possible. People don't even want [VMT]mileage tracked as a replacement for the MFT, they would freak out over Skynet controlling their freedom on wheels. As for the train avoiding the collision, I suppose that comes down to good old physics. Combined speed and distance may make potential collisions at grade crossings unavoidable.

[mhays]

The "bar" for trains ought to be high. But we're starting to allow automated cars, and the variables from a driver perspective might be much simpler for trains than for cars.

The reason to do it would be staff cost and the difficulty of hiring. I can imagine an automated system resulting in fewer incidents just by being fully attentive vs. human drivers.

[electricron]

Originally Posted by mrnyc

err, hate to break it to you tex, but many families pay that money and much more for apts and terracehouses and not all the walls are hollow.

Many do, i'll grant you that, but most do not.
https://www.statista.com/topics/5144...#topicOverview
"Everyone wants to feel like the king or queen of their own castle, and there is no other type of home that makes one feel this way more than a single-family house. In the United States, the majority of housing units are single-family houses – about 82 million out of the total 129 million occupied units in 2021. These homes are mostly owner-occupied, but a small share is rented. Most of the sales are of existing homes and just a small share is of newly built homes. In 2022, 5.8 million existing home sales took place, versus 644,000 newly built home sales."
Some math = 82/128 x100 = 63.5%
Therefore, almost 2/3rds of the homes n the USA were single family homes in 2022 (last year).
Some more math = 5.8 / 6.444 x 100 = 90.006
And 90% of the buyers of homes last year bought existing homes, not brand new ones.

[jmecklenborg]

Originally Posted by mhays

I wonder if the same self-driving tech that applies to cars will eventually apply to rail...maybe we won't need drivers even for systems that cross streets?

I thought of this years ago - that automated streetcars ought to be much easier to build than driverless cars. Obviously, the track eliminates many variables, but also the system's data could be fed by cameras/radar on every catenary pole and pointed down side streets.

The big problem with driverless cars is that they can't take advantage of perspectives other than their own, including those 100+ feet ahead. A driverless streetcar system could.

[ardecila]

Originally Posted by mhays

The "bar" for trains ought to be high. But we're starting to allow automated cars, and the variables from a driver perspective might be much simpler for trains than for cars.

The reason to do it would be staff cost and the difficulty of hiring. I can imagine an automated system resulting in fewer incidents just by being fully attentive vs. human drivers.

Agreed... an autonomous train or bus on a fixed route *ought to be* a lot simpler than an autonomous car designed to go anywhere.

I think part of the problem is that driving the vehicle is only half the job. Drivers/operators are also expected to perform the role of conductor and enforce fares, maintain safety/order, etc. AI can't really do that (yet).

I believe I mentioned before that the Dubai Tram is the one example of a (semi) automated surface LRT. It has drivers, but it has enclosed stations with platform screen doors and driver-assist systems that help the tram berth in alignment with the platform doors. But it also has its own ROW, and the cityscape of Dubai can't really be compared to American or Euro cities.

[mrnyc]

Originally Posted by electricron

Many do, i'll grant you that, but most do not.
https://www.statista.com/topics/5144...#topicOverview
"Everyone wants to feel like the king or queen of their own castle, and there is no other type of home that makes one feel this way more than a single-family house. In the United States, the majority of housing units are single-family houses – about 82 million out of the total 129 million occupied units in 2021. These homes are mostly owner-occupied, but a small share is rented. Most of the sales are of existing homes and just a small share is of newly built homes. In 2022, 5.8 million existing home sales took place, versus 644,000 newly built home sales."
Some math = 82/128 x100 = 63.5%
Therefore, almost 2/3rds of the homes n the USA were single family homes in 2022 (last year).
Some more math = 5.8 / 6.444 x 100 = 90.006
And 90% of the buyers of homes last year bought existing homes, not brand new ones.

yeah thx for looking it up, but anyone could have ballpark guessed 2/3 of home sales are sfh, thats no surprize. that 1/3 are not is the point. not insignificant.