#3D pathgeneration and pathfollowing for UAV in VREP This repository contains python code, that can be ran in combination with the Roboter Simulation Platform VREP. We used this Platform to simulate Pathfinding, Pathsmoothing und Pathfollowing for an UAV in some different areas(one small test area and a bigger area showing 2 buildings of our univesity)

TODO Clean this REPO!!!!!

Ubuntu 14.04 VREP (VERSION) Spyder3 (for comfortable development)

1. Install latest VREP version

2. Clone the repo

3. Open VREP and open the scene named "vrep/columns_and_blocks.ttt"

4. Open spyder3

5. In Spyder, run the UAV_main.py to start the mapdatageneration, pathgeneration and pathfollowing algorythm

   Step 1:

   Generates an array with the mapdata for the current scene in VREP

   Step 2:

   Reads the starting position and the goal from the scene

   Step 3:

   Finds and interpolates a Path

   Step 4:

   Starts the pathfollowing

   Step 5:

   TODO: After the goal is reached, it shows the calculated path in comparison to the real path the UAV was flying in 2D.

• The function path_3d(pos, loop) included in the file "PathGenerationCubic.py" takes as an argument an array of 3D points (pos) where the path should go through and a boolean (loop) which defines if the trayectory is closed or open. The return of the function is an array for each coordinate of the path and the time (path_x, path_y, path_z and path_t).

• The function calc_vec_field_fast(path_x, path_y, path_z, cube_size, cube_step) included in the file "FieldGeneration.py" defines a 3D matrix CUBE for the field of size cube_size and with step size cube_step. For each point inside this cube it calculate the best aproximation vector to the path using the methodology explained in the paper: TODO. The return of the function is the calculated vector field and some helper grids for displaying this field in mayavi.

In the file "example.py" we define 3 points in a closed loop configuration:

Pos0 = array([4, 8, 3])
Pos1 = array([14, 12, 17])
Pos2 = array([14, 4, 17])
Loop = True
Pos = array([Pos0, Pos1, Pos2])

path_x, path_y, path_z, path_t = path_3d(Pos, Loop)

and then calculate the vector field with:

vector_field_3D, Xc, Yc, Zc = calc_vec_field_fast(X, Y, Z, 20, 1)

Finally this script uses mlab (Mayavi) to show the resulting path, the vector field and the striplines to observe the possible paths from every point in space.

To create a different path it is only necessary to modify the example file (example.py) with diferent paths and cube sizes.