Decision Tree

Decision Tree

Input (filtered -> noise):

1. Distance of the distance sensors -> 8 sensors
2. Human info -> distance, horizontal angle and vertical angle

Cases:

1. No human detected:

• human info fails to detect human
• Start to search for human
• use last human direction and rotate for in that direction for a given degree

2. Human is close:

• fly back
• check walls -> use range sensor node information -> avoid walls
• if obstacle detected behind fly in direction of horizontal correction

3. Human far away:

• Do nothing but the default case

4. Defaul Case:

• done in every case
• adjust yaw based on human direction
• correct position based on range sensor information
• if adversarial corrections are required -> stop