##Intelligent Robot Systems 2016-2017
This code was created in order to allow experimentation towards developing specific modules of an autonomous simulated vehicle that performs full exploration and coverage of an unknown a priori environment. Next, you can find instructions on how to setup the code in your PC / laptop, the description of the code, as well as the description challenges.
The perfect score is 20 points (10 pts = 1 degree)!
At the end you must submit a compressed file containing this folder with your changes and a report on the selections you have made (up to 10 pages).
Challenge 1 [2 pts]: Laser-based obstacle avoidance
You must fill the part of the code that calculates linear and rotational velocities using the LIDAR values. The objective is for the robot to wander, but not collide to obstacles. Please add the code here and here.
Notes: The robot must have a maximum absolute linear speed of 0.3 m/s and maximum absolute rotational speed 0.3 rad/sec.
In order to check this functionality turn this to False.
Challenge 2 [1 pt]: Path visualization
This task is about making the path visible to the RViz tool. Please add the code here. The self.robot_perception.resolution and self.robot_perception.origin parameters may be useful, thus you must understand how the robot perception module works / stores the data.
In order to test it, the exploration mode must be enabled, thus turn this to True. This value will be kept True for all the remaining tasks.
Challenge 3 [2 pts]: Path following
This task is about producing the correct velocities for the robot to follow the produced path. Please add the code here. The known parameters are the global robot pose acquired from the SLAM algorithm and the next subtarget.
Again, the robot must have a maximum absolute linear speed of 0.3 m/s and maximum absolute rotational speed 0.3 rad/sec.
Challenge 4 [1 pt]: Path following & obstacle avoidance
This task is about combining the path following and obstacle avoidance velocities using a strategy like motor schema, subsumption or a hubrid one. Please fill the code here.
Again, the robot must have a maximum absolute linear speed of 0.3 m/s and maximum absolute rotational speed 0.3 rad/sec.
Challenge 5 [1 pts]: Smarter subgoal checking
This task is about making the subgoal checking routine smarter. Right now it checks if the next subtarget has been approached. But what if the robot reaches the end or a later subgoal without visiting the next? Please fill your code here.
Challenge 6 [5 pts]: Smart target selection
This task is about finding a smart approach of the target selection problem. Remember that this approach must work well in a-priori unknown environments which may be quite heterogeneous in structure (e.g. wide, narrow or both). The available tools you have are the robot's pose, the so far explored map, the coverage field containing information on what is already covered by the robot, the Brushfire field of the OGM, the skeleton of the OGM and a topological graph.
It is not necessary to use all of them. If you don't, remember to erase them in order to speed up the exploration, since as you will notice, more available information requires more time to calculate.
Also there is a possibility of the path planning to fail for specific targets that are close to obstacles or that they exist in a "strange" topology. In that case, you must write code to deal with these cases (e.g. select a more naive target selection method). Bear in mind that the robot should always have a target to reach in order for the exploration and full coverage to end!.
Please alter this code.
Extra Challenge 1 [3 pts]: Path optimization / alteration
The path planning module produces length optimal paths using an A* algorithm, along with a uniform PRM graph. This doesn't mean that the length optimal paths will result in less coverage / exploration time.
Investigate ways to alter the path, aiming at increasing the coverage rate, inserting code here (easy approach), or directly altering the path planning code here (hard approach).
Extra Challenge 2 [3 pts]: Algorithmic optimization
You will notice that a serious amount of time is being spent in calculations concerning the target selection. Feel free to alter whatever code you want in order to optimize it! Optimizations already exist using the Cffi library, where C code is being executed within Python. You can follow this approach or improve Python code.
Extra Challenge 3 [2 pts]: Surprize me
Open challenge! Do something that I do not expect!
How to test different maps