What happens when you give a machine the ability to perceive, decide, and act in the world – on its own? This is the question Autonomous Agents explores.

https://www.youtube.com/watch?v=61kGid-peqU

Autonomous Agents is a Level 8 module at TU Dublin’s School of Computer Science, shared with Game Design students in third year. It’s a hands-on, studio-classroom based course where students build autonomous systems that move, think, and act – from classic flocking simulations to cutting-edge agentic AI and robotics.

The module begins with fundamentals: how agents sense their environment, how they move (seek, flee, flock, swarm), and how they find their way through space using algorithms like A*. From there it progresses into decision-making architectures — finite state machines, behaviour trees — and then into modern agentic AI: large language models as agents, tool use, planning, and prompt engineering. We have several cool robots, and a motion capture system available to us for classroom use and projects:

Assessment is: a lab test demonstrating hands-on proficiency (20%), a take-home project building something like an artificial life simulation, a holographic autonomous character, or an autonomous robot (50%), and a written exam (30%). Everything lives on GitHub.

Reading includes Ian Millington’s AI for Games, Chip Huyen’s AI Engineering, and Craig Reynolds’ seminal paper on steering behaviours — alongside Isaac Asimov’s I, Robot and Philip K. Dick’s Do Androids Dream of Electric Sheep?

If you’ve ever wanted to program a robot that thinks for itself, power a simulated being with an LLM, or simulate a crowd of thousands — this is the module for you.

Studio Classroom runs on Tuesday mornings in CQ-240:

09:00-10:00 – Self directed
10:00-12:00 – Instructor lead
12:00-13:00 – Self directed