Technical Designer

Lockheed Martin: Project Everglades

Project Everglades was my Senior Design project at UCF, the Capstone of my Undergraduate Computer Science program. As part of a 5 person team consisting entirely of students majoring in Computer Science, I served as a programmer and Sponsor Liaison to the team's Lockheed Martin sponsors, Rebecca Broadway and Tanner Lindbloom. Project EVERGLADES was an ongoing project spanning multiple years with me and my team working on one of the project's many iterations. The primary focus of Project EVERGLADES is to serve as a showcase of Machine Learning both for growing interest in students towards the field, as well as for internal teaching of the subject.

Role: Sponsor Liaison, Programmer
2022

3D Game Representation of one of the Results of Project Everglades after my team's iteration.

3D Game Representation of a base. The 3D game world was created by the Florida Interactive Entertainment Academy (FIEA) a year prior to my team's iteration of Project EVERGLADES.

3D Game Representation of the simple grid board of the game. At project start, there was not yet diversity in the map and only a static field was consistently used.

Project EVERGLADES is a simulation where two opposing teams fight to control a map by commanding drones. Typically, the two teams are both controlled entirely by an AI. Through Machine Learning, AI can be trained to make smarter decisions in this game, and the ultimate competitions are between two trained AI.

A visualization of the different nodes of the game. This tool was created by my team to better understand the way the game worked. We created the ability to generate different 'fair' maps, with this one being a Bell Curve map.

A Non-Bell Curve Map, showing the extent of customization allowed to the game board.

Supplemental Resources:
The following is the exert for the goal of this project's iteration:
There are two primary objectives for this iteration of Project EVERGLADES. The first is to expand the game board, allowing for nodes on multiple elevations and connect them realistically. This means creating nodes on the at different elevations for a 3D board. The other objective is to allow players and AI to create custom drone squadrons composed of any drone unit type players desire, using a balanced cost system.
Though not required, Lockheed Martin would like to see increased playability, improved targeting capabilities in drones, and an implementation of electronic communications between said drones. In its current state, the foundations for the game have been set without much regard for playability and making the game fun. As such, a larger focus on bettering user experience is desired to make the game more user-friendly and engaging.
In terms of the current state of drones, they do not exhibit intelligent behavior when targeting enemies, so coordination in these attacks is desired. Communications are intended to model real-life electronic communications between drones, with the possibility for cyberattacks to jam enemy communications. Drones will be given fresher AI to allow them some self-agency in the case that such a cyberattack occurs.

One of the Presentation Slides, showcasing the roles delegated out to the team.

One of my various roles was to overhaul the complexity of the drones of the game. Prior to the project, there was one type of drone, which did not add much complexity to the strategy of the game.
By expanding the drones to have different types and specialization, a new and interesting challenge was given to the AI, 'Discovering how to best use unique units'. This makes better replication of real life combat where different units and groups have different strengths and weaknesses due to training or equipment.
I worked on creating new systems to handle multiple drone types with different attributes, as well as systems to handle squads consisting of multiple unit types.
↑Back to Top