3 ms·

It's just a change in safety mechanism. If a car is capable of stopping in a shorter distance (because it has faster reactions, and now weighs less) then the sp

It's just a change in safety mechanism. If a car is capable of stopping in a shorter distance (because it has faster reactions, and now weighs less) then the speed of many accidents is going to come right down. If a car is capable of ensuring a collision is directly head-on rather than partial clipping, the types of safety mechanism can change. Nothing is going to change over night, but there's not going to be a need for the plethora of different options to cope with different types of accident that we have now. This will then drive planners to change how roads work - e.g. segment heavy transport maybe - and it will become a virtuous circle. Smaller, light-weight transit will make electric much more economic and work better for large cities. It will be an excellent change for everyone except those wedded to their SUVs.

falcolas 10y ago
> a car is capable of stopping in a shorter distance Stopping distance has more to do with speed than weight, since reducing weight also reduces the frictional force against the road at equal proportions to the reduction of potential energy, whereas the impact of velocity on the energy to be dissipated is squared. Brakes also wear down, steering linkages break, lugnuts are improperly torqued, tires blow out, debris can impact sensors or power packs, roads can be compromised by loose sand or potholes, lightning can strike the vehicle... failure modes are plentiful. You also still need to account for bugs and sensor failures; both of which will result in crashes in "non optimal" ways. As an example, I had a gyro fail in my multirotor; it was a gradual failure which was not immediately obvious (and completely invisible to the controller), and resulted in a couple of pretty spectacular crash that the multirotor (even with human inputs) had no way to recover from. Sure, you can add redundant systems (ideally in batches of three), but at some point there will always be a single point of failure.
- nostoc 10y ago
  Stopping distance has mostly to do with two things : reaction time and weight. I expect an autonomous car to have a far better reaction time than a human.
  - falcolas 10y ago
    Reaction time, I can kind of get. But weight, especially when compared to velocity and the coefficient of friction - i.e. the road conditions - has negligible impact. Sure, it increases your traction when stopping, but it also increases the amount of energy which has to be dissipated by the vehicle to stop. Something that has a much greater impact than either would be the road conditions. Wet, dry, icy, sandy, gravel, oily... all have much more impact than either weight or reaction speed. Have a look at this review: http://www.caranddriver.com/features/the-power-to-stop http://www.caranddriver.com/features/the-power-to-stop The quality of the brakes had a much greater impact than the weight of the compared vehicles (a SUV had a stopping range of 355 ft, whereas a coupe had a stopping distance of 365).
- adrianN 10y ago
  Friction is not as linear as high school physics teaches: https://www.wired.com/2014/09/friction-isnt-always-think/ https://www.wired.com/2014/09/friction-isnt-always-think/
  - falcolas 10y ago
    I can't read that article (it thinks I have an ad blocker). Can you summarize the better model they have in mind?
rwmj 10y ago
Also the car doesn't get drunk, tired, distracted, angry, or drive too fast for the road conditions. I'm looking forward to being able to take my bike to the station without the risk of being killed by a lorry driver looking at his phone (something I see on probably 1 in 5 journeys).
- pixl97 10y ago
  Well, your car might actually get distracted and/or drive too fast for road conditions. Making sensors that work under all conditions is very difficult.

falcolas 10y ago

> a car is capable of stopping in a shorter distance Stopping distance has more to do with speed than weight, since reducing weight also reduces the frictional force against the road at equal proportions to the reduction of potential energy, whereas the impact of velocity on the energy to be dissipated is squared. Brakes also wear down, steering linkages break, lugnuts are improperly torqued, tires blow out, debris can impact sensors or power packs, roads can be compromised by loose sand or potholes, lightning can strike the vehicle... failure modes are plentiful. You also still need to account for bugs and sensor failures; both of which will result in crashes in "non optimal" ways. As an example, I had a gyro fail in my multirotor; it was a gradual failure which was not immediately obvious (and completely invisible to the controller), and resulted in a couple of pretty spectacular crash that the multirotor (even with human inputs) had no way to recover from. Sure, you can add redundant systems (ideally in batches of three), but at some point there will always be a single point of failure.

nostoc 10y ago
Stopping distance has mostly to do with two things : reaction time and weight. I expect an autonomous car to have a far better reaction time than a human.
- falcolas 10y ago
  Reaction time, I can kind of get. But weight, especially when compared to velocity and the coefficient of friction - i.e. the road conditions - has negligible impact. Sure, it increases your traction when stopping, but it also increases the amount of energy which has to be dissipated by the vehicle to stop. Something that has a much greater impact than either would be the road conditions. Wet, dry, icy, sandy, gravel, oily... all have much more impact than either weight or reaction speed. Have a look at this review: http://www.caranddriver.com/features/the-power-to-stop http://www.caranddriver.com/features/the-power-to-stop The quality of the brakes had a much greater impact than the weight of the compared vehicles (a SUV had a stopping range of 355 ft, whereas a coupe had a stopping distance of 365).
adrianN 10y ago
Friction is not as linear as high school physics teaches: https://www.wired.com/2014/09/friction-isnt-always-think/ https://www.wired.com/2014/09/friction-isnt-always-think/
- falcolas 10y ago
  I can't read that article (it thinks I have an ad blocker). Can you summarize the better model they have in mind?

rwmj 10y ago

Also the car doesn't get drunk, tired, distracted, angry, or drive too fast for the road conditions. I'm looking forward to being able to take my bike to the station without the risk of being killed by a lorry driver looking at his phone (something I see on probably 1 in 5 journeys).

pixl97 10y ago
Well, your car might actually get distracted and/or drive too fast for road conditions. Making sensors that work under all conditions is very difficult.