 Simulation of MAV flying through tunnel (click to download) |
 Simulation of MAV flying through a near-Earth environment (click to download) |
Conducting experiments to test our control algorithms is an extremely time consuming process. It would be a much more time efficient process if we could model the optic flow sensors along with our flying environment to test our algorithms in simulation.
|
Simulation of MAV flying through tunnel (click to download) |
Simulation of MAV flying through a near-Earth environment (click to download) |
The simulations above were developed using OpenGL. A MAV measures the optic flow on the left and right side of the aircraft as it is flying at a constant velocity. The objective of the simulations is to (left) have the MAV maintain its flight path down the tunnel’s center by balancing the optic flow and (right) fly a relatively straight path while avoiding collisions with urban-like structures.
The optic flow is calculated using the 1D equation OF=V/D sin x - w where “D” is the distance from the sensor to the obstacle along the optical axis, “x” is the angle the optical axis makes with the fuselage and “w” is calculated using first-order approximation of rudder deflection to yaw rate.
Some things not considered in the simulation are an accurate dynamic model of our aircraft as well as the optics of the sensors. This makes the sensor's computation of optic flow 100% accurate which is not realistic. We are currently working on modeling the optics of the sensor. In the meantime, we adjusted our simulation by adding in some random noise to the calculation. With noise, the MAV still maintains a relatively straight path down the center of the tunnel.
Back to the Table of Contents
Back to the Table of Contents