Seems so simple, unless you factor in that humans have auditory tracking and can hear the drone following them and act accordingly by going off trail and into places that the drone simply cannot navigate or reliably identify them in due to multiple physical and logic processing limitations based on input sensory parameter limits of various sorts.
To the prepared and knowledgeable person they are a limited threat. But yes, to the typical unprepared person who doesn't know what are their likely design limitations these are terrifying concepts to have to deal with.
I wonder how well they could do at chasing someone down a wooded path that has assorted fine wire and string lines strewn across it? If the sensory system can detect them what sort of added complexity is required to process those fine stringline obstacles out Vs spiderwebs and like that it could fly through?
I cant find my post now, but back when i did it, the drone had to navigate a building, which was infact a mock up built in the school hall with rooms. We put windows in and some were open some close with clear glass, and one broken but with a hole in the middle. Using the simlink software we had at the time, the libs were basic but even back then flying through a broken windows easy, detecting a closed window even easier.
The stereo camera vision was the new bit on the drone, apart from the actual self decision making side of things, on top of that it had a laser range finder and ultrasonic sensors from a car bumper system. Moving objects are really simple to track, modern motors and light drones you wouldnt hear in 'stealth' mode. Best avoidance would IMHO be stand by a tree and face it, dont be tempted to move, get behind the drone and take it out. or signal jam, but my own drone used non standard frequencies because of noise, i used 868mHz and 969 MHz.
In reality it likely you would be taken out from 60,000 feet above and 30 miles out. But lets say they go the small killer route, then nets are good. Dead simple leave a branch in the direct path and one leave one side open, as it closed in push something to that side, the drone has no option but up into a netting and the motors catch.
this kind of stuff was what we spent weeks talking about, most of it because in our scenario what would you encounter in say a tower block fire or the twin towers? The idea behind the drone was search and rescue without sending a crew in unless needed. So we planned for buildings damaged by fire, it was a school project done for a university competition, it ended up as my project against MIT, my team dropped out leaving me after 3 weeks on my own.
Thats when the uni stepped in and lent me some student help twice a week to build stuff and do the simulink things i couldnt do. But the actual program steps and all the electronic design was mine, it wasnt that hard in reality. The big problem i had was money, the school budget was £50 i added £300 i got some equipment donated, we went up against MIT etc with budgets in the $000's and came second.
our vision system was better, there drone faster but the applications were different. Plus as i got told the following year...... a school kid beating 3rd year university students, just isnt on.
I think most of all of this is posted up on here, i have the actual drawings and schematics somewhere i might post up. Pommie gave me the direction for the maths for self flight, he answered a different question but turns out same maths involved. loads of people on here gave me ideas i put into it. The drone itself died, but one day........ When i get a glow engine to power electric motors, i will come back to this.
then its there drone against mine
.
The technology itself could really make a difference to people, as always its down to the intent of the user that makes the difference. if you think you could shoot one down, go on you tube, they do this as a sport with machine guns the lot!! Its actually really hard to shoot one down if the pilot or drone knows your trying to do it.
