Homework 3

Due: March 23, 2006 (midnight). No late submissions accepted.

For this homework you will implement a search-based agent that is capable of computing a path from the agent's starting location to the goal location, and then follow this path to the goal. The specific requirements are enumerated below.

We will be using the same version of the Urban Combat Testbed (UCT) you already have (v0.5). I have updated the agent program to provide a template for a search-based agent. You can download this agent and supporting files from here. See the README.txt file included in this package for details about the contents.

We will be using the same type of goals as the previous homework; namely, getting your agent to a specific x,y,z location. However, now this task will be accomplished by searching for a path to the goal, rather than making more local decisions about the next action to take.

The search-based agent program is currently a template that defines a graph structure, reads in an XML description of the uct_apartments world (same as uct_rekjavik) which describes the open areas of the world and their interconnections, and then continuously tries to find a sequence of connecting areas starting from the area where the agent currently resides and ending at the area containing the goal location. Your job is to complete the program by implementing the following procedures:
- createGraphFromXMLInfo() - fills in the worldGraph data structure with the information read in from the XML file data/uct_apartments/uct_apartments-ssps.xml. See the procedure printXMLAreaInfo for an example on how to access the XML data.
- findArea(x,y,z) - returns the area containing the point (x,y,z). Each area is defined by the vertices of a three-dimensional polygon, so you will need to implement an algorithm for iterating over the areas and finding the one that encompasses the point.
- findNextAreaOnPath(a1,a2) - computes a sequence of connected areas from area a1 to area a2 and returns the area right after a1 as the next area on the path. You will need to implement some auxiliary procedures here that search the graph for the path and set the predecessor fields of the graph structure along the way to remember the path. A standard shortest-path algorithm will suffice here. In the event no such path exists, the agent will revert to a simple refelx agent. You can use the reflex agent included (same as the one provided for HW1), or introduce your improved reflex strategy from HW2.
- computeGoal(area,goal) - sets the x,y,z of the goal structure to an appropriate location within the given area (e.g., centroid). This procedure is used to choose a specific location to move towards within the desired area, which is next in the path toward the final goal location.
Since we only have XML data for the uct_reykjavik world (now called uct_apartments), your agent will only be tested in this world. As before, you should assume that the agent can start at any position and that the goal location can be any reachable position in the map. You may assume a tolerance of 10. Your agent will be evaluated according to its speed and success at reaching various x,y,z locations from various starting x,y,z locations. So, you should test your agent accordingly to be sure it is robust for these types of tasks.
Document the design of your agent, including the original elements of the search-based agent and those added by you for this homework, in a separate document (MSWord, PDF or text). You should also describe the strengths and weaknesses of your agent in regards to the navigation task described above.
Submit all source code and design document files in one zip file to me (holder@cse.uta.edu) by the above deadline. Also include in your submission a README file that describes what files are in your submission and any special instructions for building and running your agent. Most likely, you will submit an agent.cc file, design document, and README, and the agent.cc file can be compiled in the agent directory just as the original one. In addition to correct functionality and satisfaction of the above constraints, your submission will be graded based on good programming style and documentation.