Planned Release Date: 2027

About This Game

You teach. The AI learns to walk.

AI ROBO STADIUM is a game about teaching robots to walk, then sending them into competition.
You decide what to reward and what to punish. How to drive the 16 motors in the legs is something the AI figures out on its own through reinforcement learning.

Train them in the Lab

• Set reward and penalty levers such as "follow the command", "keep the body level", "don't fall", "save energy" and "move smoothly"
  
• Pick a practice ground from flat floor to steps and slopes, add pushes, and randomize the physics to prepare for the real thing
  
• Watch 16 robots practice at once while the learning curve climbs
  
• Take a model that is still in training and pilot it on the spot
  
• Save as many trained models as you like under your own names, and keep training any of them later

Compete with the AI you raised

• Race: run an obstacle course of steps, slopes and a slalom against a rival robot in a single shared lane. Bumping is allowed, and a robot that cannot get up within 3 seconds loses
  
• Sumo: a one-on-one pushing match in the ring. Knock your opponent off balance with punch, lift and wedge arms
  
• You steer, the AI walks. You give the forward and turn commands, and the model you trained turns them into leg movements

New parts mean new training

Swap in a LiDAR, an IMU, foot contact sensors, wheels or upgraded motors. Bolting on a part does not make a robot stronger by itself. It has to learn to use the new loadout first.

Your PC becomes a robotics lab

Training runs on your PC's CPU. It needs no GPU and no Python setup. You can keep racing and wrestling while it trains, and training continues even after you close the game.

Robots

• Gazelle: a quadruped
  
• Teeter: a biped with wheels at the end of its legs

The game is currently in Japanese only.

AI Generated Content Disclosure

The developers describe how their game uses AI Generated Content like this:

Some of the in-game UI text and store page text was drafted with the help of an AI assistant and then reviewed and edited by the developer. No AI-generated images, 3D models, music or voices are used. The robots are controlled by neural network policies that players train with reinforcement learning on their own PC. These policies output joint motor targets only. They do not generate images, audio or text.