Originally Posted by
cwatters
If Amazon end up making 20,000 flights an hour (see above) then they will need to be exceptionally reliable to avoid having a lot of crashes. One crash per million flights equates to about one crash a week. Would most be over cities? Would that be acceptable?
The crashes won't happen in the calm sunny days that Amazon show in their advertising and presumably in which all their testing has taken place. It will happen in the bad weather, wind, heavy rain, hail showers, snow, sleet etc.. In cities the Venturi effect between buildings can be extreme with a benign 15kt wind becoming random funneled 50kt gusts. Anyone who has flown helicopters in cities will be aware of these issues, an automated UAS with less than 30kt max speed has no chance.
There are going to be days or even weeks when the UAS delivery system will at best not work at worse will deliver dead UASs and packages to unwilling recipients. One person killed by an Amazon UAS and all the investment will be wasted. I cannot see this delivery UAS idea passing even a cursory safety assessment. That is before we get to the security acceptance of anonymous autonomous UAS flying 5lb packages around major cities like DC, New York or London.