Neuro-Evolving Robotic Operatives | Friday Game

Back in 2003, there was a GameDev conference in Texas USA. The focus of the conference was artificial intelligence, and consequently several people from the Neural Networks Research Group at the Texas University Department of Computer Science were invited to make presentations on academic AI research with potential game applications. After some brainstorming, ideas and conversations NERO project started development.

NERO is a machine learning artificial intelligence game that combines creation, imagination, knowledge, experience and real time strategy elements. To play NERO you have to read a lot, but eventually you will get repaid by the creative feeling that this game gives you. There are some very comprehensive tutorials in game, that will take you by the hand and lead you to more and more advanced stuff.

The NERO game takes place in the future as the player tries to outsmart an ancient AI in order to colonize a distant Earth-like planet. Read the full NERO story. The player has to train and use agents for different tasks depending on game mode. For example if you play a capture the flag scenario, you’ll have to train and use agents that are good at attacking and capturing flags, while also making sure that you have trained enough defensive agents. All this happens inside a 3D sandbox-map where the agents of your and the opposing team interact and fight each other. Watch the following video to get the idea:


For every time you are doing good, you are rewarded by evolution potential. Your agents evolve to perform better in what they are trained to do. Training for complex tactical behaviors will require a player to think out and implement a shaping plan, leading the robots through a series of sandbox scenarios that guide them stepwise to the desired battlefield doctrine. Once the robots are able to handle multiple turrets, walls are included into the environment, and the turrets are allowed to move. Finally, after the robots have been trained through all of these incrementally more difficult tasks, they are deployed in a battle against another user’s trained team to see how effective the training sequence was.

Currently, NERO is almost dead in terms of development as all developers are now working for OpenNERO!

OpenNERO is based on NERO, but it is open source and currently under development. This projects aims to be used more as a platform for research and education in Artificial Intelligence rather than being a game. As a gamer you will either love this game with all your heart, or completely hate it. For those of you who will like it, there are also tournaments held every year to test your team against the rest of the world!