AI
All projects are done in Unity using C#. Playable Web GL builds are in each GitHub repo.
Behavior Trees
Designed and implemented a custom behavior tree from the ground up. AI enemies will move within range, fire at the player, and get more ammunition completely independently.
​
GitHub Repo:
ML Agents
Basic study of machine learning using Unity ML-Agents and Python. The agent was trained to avoid obstacles and reach the end. Negative and positive reward barriers are left visible to show the training environment.
​
GitHub Repo:
Steering AI
Developed and implemented basic steering behaviors. Including but not limited to, Avoiding, Evading, Path following, Targeting, and Arriving.
​
GitHub Repo:
Steering AI Cont.
Implemented the abstract behaviors in a simple game-like setting. The player must dodge incoming missiles to survive.
​
GitHub Repo:
BOID Study
Implemented a dynamic AI boid for simulating basic flocking behavior. Boids follow a predetermined set of waypoints and will avoid obstacles. Cohesion and Separation strength values can be set on the fly to see their effects.
​
GitHub Repo:
Ballistics
Created an AI that uses ballistics equations to create a firing solution. Muzzle velocity and target position are given, and the firing vector is calculated. Can be easily modified to solve for the other variables. In this example, AI pick their own position and muzzle velocity semi-randomly to make a frog hop around.
​
GitHub Repo:
Nav Mesh
Demonstration of Unity's built-in pathfinding system Navmesh. The character navigates wherever the player clicks as well as avoids the obstacles with no separate steering behaviors.
​
GitHub Repo:
