RRT

A C++ implementation of RRT using Boost::geometry and matplotlib-cpp

Installation

git clone https://github.com/archit2604/RRT.git

Running the code

To plan a path using the RRT algorithm use the following commands. The obstacle co-ordinates and the search space is fixed.

cd ./RRT
g++ rrt.cpp -std=c++11 -I/usr/include/python2.7 -lpython2.7
./a.out

Once you run the file, it will ask for user inputs for various fields, choose the values from permissible ranges mentioned.

---

Variable taken from user and their definitions

• start_x (float) : x-coordinate of the start point

• start_y (float) : y-coordinate of the start point

• goal_x (float) : x-coordinate of the goal point

• goal_y (float) : y-coordinate of the goal point

• K (int) : maximum number of nodes in the tree

• step (float) : step-size for RRT

• g_sampling (int) : goal sampling rate

Conditions:

• start_x , start_y , goal_x and goal_y should be inside the search space and outside obstacles.
• K should be less than 500.
• step should be preferably a low value, to increase the possibility of getting a path.
• Set g_sampling to 0 to get an unbiased tree.
• If g_sampling is 5, the every 5 th sample is goal.

---

Changing the Obstacle Co-ordinates

To change the Obstacle Co-ordinates the line 18 in the file rrt.cpp needs to be edited.

int obst_array[][2] = {{2, 7}, {7, 7}, {6, 4}, {4, 4}, 
{4, 6}, {2, 6}};

Changing the Search Space

The search space is a square with (0,0) and (12,12) as the opposite edges. To change the Search Space the line 54-55 in the file rrt.cpp needs to be edited.

int Offset = 0;
int Range = 13;

To set the search space to any square with (a,a) and (b,b) as the opposite edges change the code to:

int Offset = a;
int Range = b+1;

---

Plots

Figure-1

• start : (0,0)

• goal : (10,10)

• K : 100

• step : 0.5

Without Bias ( g_sampling : 0 )

With Bias ( g_sampling : 25 )

---

Figure-2

• start : (3.5,5.5)

• goal : (5.5,7.5)

• K : 100

• step : 0.5

Without Bias ( g_sampling : 0 )

With Bias ( g_sampling : 25 )