4 ms·

I wrote the initial interface specification for the co-ordination system in Rio's open-pit mines. The mere fact that we were able to account for the state of ev

I wrote the initial interface specification for the co-ordination system in Rio's open-pit mines. The mere fact that we were able to account for the state of every moving thing in the pit makes this a completely different proposition to self-driving vehicles in an uncontrolled real-world environment. At a guess it's at least ten to a hundred times easier.

King-Aaron 8y ago
Eventually this could be put into place in the broader traffic ecosystem - if all the cars on the road were required to have a transponder sort of arrangement on board, you could have a central point of control, and perhaps tie in people's phones etc to locate pedestrians. I do want to note that I am aware that there is an unfathomably large amount of complexity behind that however.
- Schwolop 8y ago
  There's a movement towards this; look up "Vehicle-to-Everything" or "V2X". The idea is to enable things like one vehicle transmitting sensory information or intent to others around it, helping to prevent rear-end collisions, etc. My personal take is that it's well intentioned, but there will always be things that aren't part of the corpus of communicating devices that will then limit the utility. I suspect the communication of sensor data is the more useful part than some of the demos I've seen so far of, for example, a car sending an alert to a person's phone so they don't step into a driveway when it's reversing.
- brett40324 8y ago
  And when software on a car assumes pedestrians always have a mobile device on their person, charged and ready for a split second alert? IMO This would be an absurd design.
  - King-Aaron 8y ago
    It'd be a tool in the toolkit, in my hypothetical. I mean, by the time autonomous cars were common enough on the road, we might all be tagged with an rfid chip as a normal identification or something... I don't disagree that it's an absurd idea, but it's just a devils advocate position I'm discussing here.
- jcranmer 8y ago
  Because my dog will totally have his mobile phone on him when I'm chasing him down. Or my little kid chasing after his ball...

King-Aaron 8y ago

Eventually this could be put into place in the broader traffic ecosystem - if all the cars on the road were required to have a transponder sort of arrangement on board, you could have a central point of control, and perhaps tie in people's phones etc to locate pedestrians. I do want to note that I am aware that there is an unfathomably large amount of complexity behind that however.

Schwolop 8y ago
There's a movement towards this; look up "Vehicle-to-Everything" or "V2X". The idea is to enable things like one vehicle transmitting sensory information or intent to others around it, helping to prevent rear-end collisions, etc. My personal take is that it's well intentioned, but there will always be things that aren't part of the corpus of communicating devices that will then limit the utility. I suspect the communication of sensor data is the more useful part than some of the demos I've seen so far of, for example, a car sending an alert to a person's phone so they don't step into a driveway when it's reversing.
brett40324 8y ago
And when software on a car assumes pedestrians always have a mobile device on their person, charged and ready for a split second alert? IMO This would be an absurd design.
- King-Aaron 8y ago
  It'd be a tool in the toolkit, in my hypothetical. I mean, by the time autonomous cars were common enough on the road, we might all be tagged with an rfid chip as a normal identification or something... I don't disagree that it's an absurd idea, but it's just a devils advocate position I'm discussing here.
jcranmer 8y ago
Because my dog will totally have his mobile phone on him when I'm chasing him down. Or my little kid chasing after his ball...