• MystikIncarnate
    link
    fedilink
    English
    arrow-up
    21
    ·
    11 hours ago

    Well, the automotive industry has been working on making self driving a thing, and I recall when they first tried to tackle the problem of lane keeping.

    The first proposal was to embed magnets or similar into the road surface that the car could have a set of sensors for to determine if it was drifting left or right in its lane.

    Motherfucker, that’s just a virtual track for your dumb four-wheeled mini-train.

    It didn’t catch on, but AFAIK it was implemented in small areas as a trial and it performed adequately given the technology of the time.

    So I’m out here going, why the fuck are we pretending that vehicles are not just rail-free personal trains?

    • rumba@lemmy.zip
      link
      fedilink
      English
      arrow-up
      6
      ·
      4 hours ago

      That’s the end game for self-driving cars. They can drive close enough together to draft, efficiency goes way up. If a problem happens ahead, they communicate back so that pileups don’t happen.

      • buttfarts@lemy.lol
        link
        fedilink
        arrow-up
        3
        ·
        1 hour ago

        I am looking for intersections with two catepillars of drafting vehicles that slightly intersperse to cross each other at right angles by microtiming the gaps to avoid collisions by microseconds.

      • MystikIncarnate
        link
        fedilink
        English
        arrow-up
        3
        ·
        2 hours ago

        IMO, the big strength with self driving cars, if we ever get there is that level of car to car communication. The vehicle will be able to communicate ahead and see the best possible route, and where there’s congestion etc, then optimize the drive to avoid unnecessary delays.

        A big problem with human drivers is the tenancy for ghost traffic jams to occur. There was a test they did with about 10-20 drivers of all varieties put into cars and told to drive a circle track, following eachother. No other instructions were given. All they need to do was keep distance in front of them and everything would be fine, what was observed was that some drivers went more quickly than others, and would brake to a near stop when they came close to the person in front. In doing so, everyone ended up basically in stop and go conditions.

        IMO, that test exemplifies the problem with human drivers. Put enough of them on the same road and given enough drivers and enough time, traffic/congestion will create slowdowns that otherwise shouldn’t exist.

        Taking people out of the equation means that all of the cars can accelerate at the same time and travel in tight packs, so merges are effortless because the entire system is working together to ensure that merging vehicles are able to merge (allowing sufficient space for them to merge), and perhaps more importantly, the merging cars will match pace with the flows of traffic already traveling on the road. Those are the two main tenants of a zipper merge. Find space to merge into, and match pace with the vehicles in the lane you are merging into. Seems that a lot of people forget that last bit.

        So rush hour nonsense will at least be reduced.

    • MonkderVierte@lemmy.ml
      link
      fedilink
      arrow-up
      10
      ·
      edit-2
      7 hours ago

      The first proposal was to embed magnets or similar into the road surface that the car could have a set of sensors for to determine if it was drifting left or right in its lane.

      That’s a thing for forklifts since what, a century? Even better, they use a wire with a set frequency instead of magnets, it’s called wire guidance system.