Simple Planner | Robot Programming 2023/2024

Instruction on use will follow.

Briefly

This project is intended to build a simple planning system using ROS and RVIZ.

The main node that handles everything is the pose_handler_node. Recalling that an evaluation function f(n) is the sum of a cost function g(n) and a heuristic function h(n), there are 4 parameters that this node will get to be correctly launched:

• Alpha controls the weight given to the cost function
• Beta controls the weight of the heuristic function
• Dmax is a parameter to handle values of the distance map
• Mode let you select between 3 algorithm to use: A* (i.e. f(n)=g(n)+h(n)), UCS (Uniform Cost Search, i.e. f(n)=g(n)) and GBFS (Greedy Best First Search, i.e. f(n)=h(n)), in particular, selecting alpha and beta different from 1.0, you will choose the weighted version of each of these algorithm.

Execution

Once you are in the project folder, you need 4 different terminals windows to run the program, and the commands needed to be executed are:

1. roscore essential to allows communication between nodes
2. rosrun rviz rviz to execute RVIZ
3. rosrun map_server map_server src/display_map/src/diag_map.png 0.05 to use the pre_built map_server node that allows you to send a map over the topic "/map" to RVIZ specifing the image of the map and its resolution

4.  catkin build
    source devel/setup.bash
    rosrun pose_handler pose_handler_node 3.0 1.0 0.5 A* -DDEBUG
   

   in order to:
   • build the project
   • source the setup.bash script
   • execute the node with the following parameters: alpha, beta, dmax, mode.

Once you run the last terminal, you need to setup RVIZ in order to subscribe to published topics, in particular, you need a MAP component that subscribes to topic /map and a PATH component that subscribes to a topic /path. After that, you can choose via UI buttons of RVIZ the initial guess and the final pose and let the algorithm do the job. Once the algorithm completes the search, a path will be displayed on RVIZ.

Example of execution

Using

• Alpha = 3.0
• Beta = 1.0
• Dmax = 0.5

the following are the execution using every mode available.

With UCS (i.e. only cost function f(n)=g(n))

With GBFS (i.e. only heuristic function f(n)=h(n))

With A* (i.e. summing heuristic and cost function f(n)=g(n)+h(n))