3rd year project help!

Hi there! Im currently looking to get a head start on my 3rd year uni project (doing a Computer games programming degree, c++ based) and basically im just looking for some general advice to help me refine/narrow down my ideas! I have about 12 weeks off before i return to uni! So, anything i manage to get done now would be a GREAT head start, and this project is worth 2 modules out of my total 6 we do in 3rd year (so its one third of my 3rd year grade!!) Im currently well into the 1st class grade bracket based on my second year grades, but since this is worth such a large part of my 3rd year i need to get it right! The main worry im having is that although we have free reign to pretty much investigate and develop a project in any area (related to games), we need to write an accompanying 30,000 word dissertation! (apparently thats like 50 pages according to tutors... although im sure its far more!) So i need a topic ill be able to write lots about. Currently im open to ANY interesting ideas however i do have a couple of my own, its just that they seem a little bit ... well.. not focused enough, and i dont want to waste my summer coding something i have no chance of writing a decent dissertation about. (*takes breath* - sorry about the longness of this post!) My first idea was Ai related. Something along the lines of what ive seen in the Kynogon demos at http://www.kynogon.com/products/demos/index.html. I found particularly fascinating the 'bodyguards' demo. It made me think of doing something similar along the lines of Team/Squad Ai, with some focus on threat evaluation (identifying potential avenues of attack, windows etc). The other topic i found fascinating were the two 'company of heroes' articles in Ai Game programming Wisdom 4, 'Company of Heroes Squad Formations Explained' by Chris Jurney (Kaos Studios), and also the article 'Post processing for high quality turns'. Both were extremely enjoyable to read (recommended if you havent!). I particularly thought the idea of 'leapfrogging' between locations providing cover by deviating slightly from the actual path, and then adjusting speed slightly to Maintain formation, was excellent! However both these articles are fairly short... well when considering i need to write 30,000 words. The problem is i dont want to focus on one and then not be able to write anywhere near enough on the topic! The only kind of solution i have is to target a project in a very component based/incremental approach. I need some kind of way of having a project whereby i can atempt one topic/area of interest, and then if i cant write enough about it, i can add another. Do you think for example i could attempt the basic bodyguard simulation , then extend it to add cover based leapfrogging (whilst protecting the hostage) ? (important to note that one is a third person perspective, and the COH leapfrogging obviously is an RTS so i dont know how well it would translate across...). Im pretty stumped!!! My main worry is that if i approach it this way, my ideas my seem to seperate, disjointed, and forced together! However thinking of ONE idea i can do the whole thing solely on seems immensly difficult. Oh and to top the pile of random ideas off, ive been reading about influence maps forever. They seem like perhaps they would be a great way to do risk assessment. All i know for certain is i need to hone down the idea and get a good plan before i begin!! Any suggestions (new or refinements, or areas perhaps totally new areas you guys/gals would like to see investigated!!) would be So appreciated it is untrue. Although this would be an AI project, the sad but true fact is that i would LIKE to have some at least basic textured models /lighting. Thats no problem for me, but it is a problem when i come to the fact that ill probably need some kind of animations. (shooting, crouching, walking , running , perhaps blind fire etc). Its obviously gonna be alot nicer to showcase to an employer that way. However i have no Animations, cant animate, have no real experience animating (besides Direct X skinning tutorials) or any idea where / how i can obtain them lol. I would say im an intermediate C++ programmer. Last thing i covered was interface/implementation classes and factory functions for portable code. So obviously i know the pre-requisite stuff too! Cant think of an easy way to express what my level is! (any ideas?) I have decent experience with some AI (mainly just basics - pathfinding, path smoothing, FSMS, targeting - 94% on my pathfinding assignment worth 70% of the module ftw! covered A*, Dijkstra, breadth, depth, best first) and DirectX( 91% on last assignment implemented real time env maps, spotlights, accompanying shadowmaps, parallax mapping, cell shading (basic)). However with regard to AI i havent really covered much more than what was mentioned before! I do feel i pick up new things quickly though. Oh also i do have about 9 months to do this... so even if topics are difficult (lets face it, wheres the fun in something too easy!) i expect to be able to spend time solving them, and to learn new techniques. right. rant over. (slightly random all over the place rant too!). REALLY appreciate the effort of anybody who has read this .BIG thanks. Any suggestions/comments/discussions GREATLY appreciated!!