We need project guidance to find the right sensors that can make a robot with wheels move when someone is standing with their back turned to it, and stop when said person turns around. Since we are students, we ask for the more affordable answers possible, please!!
To detect generally whether the person is turned towards or away from the robot you would need to use a camera and a powerful processor running a complex computer vision program. An Arduino Uno will never be powerful enough for that sort of application. You would probably want to use a Raspberry Pi or other SBC since this would be difficult to do even with the most advanced Arduino boards.
What might make this possible on the Uno is if you had a way to make detection of which way the person is facing much more simple. An example would be if they wore a pendant with an IR LED that sent a signal to an IR receiver on the Uno. When they are facing away from the robot the IR LED is obscured. I think that would be the easiest solution. I'm sure there are some other variations on this sort of concept as well.
Secret trick to this is in the feet of people..
When people are facing the robot, their feet will be closer than their legs and nearly as wide.
When sideways, their feet will be 2x+ as wide as their legs and offset in direction they are facing
When facing away, their feet will almost be same distance away as their legs.
So now you just need PIR sensor to say it’s human, and two sweeping laser range sensors, one scanning just above ground level, other say 4 inches up.
Yes it’s not a good as a face recongintion system, but it would be something you could do with a arduino.
What if the subject is in a wheelchair?
AWOL:
What if the subject is in a wheelchair?
Yea suppose it’s not a ADA compliant sneak attack rover without additional coding..