Behaviour Blending
Goal
Create a tool that can be used by artists to manipulate the behaviour of crowds by changing some parameters.Implementation
-
Study how Reinforcement Learning can be used for crowd simulation. This was an effort to blend two simple behaviours, the flee and seek behaviour. The two behaviours were trained separately using Reinforcement Learning. Afterwards, we used the user defined weights to blend the outputs of the two policies.
Results
- Required a different approach.
Interaction Fields
Goal
Use a force field to guide crowds.Implementation
- This was an effort to study how an interaction field that exists in the environment can be used to guide crowds. We used Reinforcement Learning to train a force field instead of individual agents, which guides multiple agents from a starting point to a goal.
Results
- It is more efficient to train individual agents.
Object-Oriented Programming Projects
Battleship
Implementation of an adapted version of the Battleship game where the user plays against the computer.World of Ants
-
Implementation of a simulation that takes place in a 2D environment where ants, termites and wood exist following the given rules.
AI Projects
Tools: C, Java, Python
Data Structures and Algorithms
-
Implemented several algorithms that uses data structures like Arrays, Linked Lists, Stacks, Queues, Binary Trees,
AVL Trees etc to solve different kinds of problems.
Database Project
-
Implementation of a database for questionnaires that are developed by multiple companies and showcasing the results in a webpage. The project was a team effort. Unfortunatelly, the access to the database is no longer available.
Tools: Microsoft SQL, PHP, HTML
Visual Computing Projects
Tools: C++, OpenGL, Python, OpenCV