THE MOST RECENT TOURNAMENT HAS CONCLUDED
More to come this fall
Micro Invaders is a strategic robotics game where multiple teams compete to develop the smartest AI, enabling two autonomous robots to perform manoeuvres in a fast-paced game of strategy. The month-long opportunity to hone your skills and practice culminates in a one-night robot showdown, where the AI-controlled mini-robots go head-to-head in hours of intensive tournament challenges.
Of the applicants, 12 teams will be chosen to compete and become certified AI Citizens in the Robot Uprising universe. The number of the teams is really limited due to Covid-19, so make sure to apply as early as possible. There will be additional activities, videos and games as part of the competition.
The finalists will receive golden tickets to our next large Robot Uprising event when restrictions relax, and they will get a chance to compete for the main prize (2000€) and, of course, ETERNAL GLORY!
Gather your friends, develop your unbeatable AI and compete in the most interactive AI tournament in Finland. Practice with custom-made arenas, climb the ladder and get connected to the biggest companies in the industry. The final tournament will be held in our custom venue in Pasila, provided by YIT and live-streamed to Twitch.
Micro Invaders is a strategic robotics game where multiple teams compete to develop the smartest AI, enabling two autonomous robots to perform manoeuvres in a fast-paced game of strategy. The month-long opportunity to hone your skills and practice culminates in a one-night robot showdown, where the AI-controlled mini-robots go head-to-head in three hours of intensive tournament challenges.
Of the applicants, four enthusiastic teams will be chosen for a one-day showdown and become the first D0WNF4LL certified AI Citizens in the Robot Uprising universe.
The finalists will receive golden tickets to the Robot Uprising Artificial Invaders event in 2021, where they will get a chance to compete for the main prize (2000€) and, of course, ETERNAL GLORY!
Gather your friends, develop your unbeatable AI and compete in the most interactive AI tournament in Finland. Practice in public arenas throughout Helsinki, climb the ladder through ranked games and get connected to the biggest companies in the industry. The final tournament will be held in our partner’s premises and live-streamed on Twitch.
Code your AI. Practice by using our simulator.
Week 2 & 3
Receive your robots!
Get ready to conquer!
Receive your robot
Week 2 & 3
Practice at home and train your AI model using our simulation
Show down tournament. Get ready to conquer the arena
Each team has two autonomous robots that are playing in a square battle arena. The arena is filled with energy cores (green spheres) that give points when a team managed to move them into their base. After all cores all collected or a game time ends, the team with the most points wins.
In addition to the positive energy cores, there are also corrupt cores (red spheres). If three corrupt cores have been scored into the base of the team, the team loses immediately. These corrupt cores will be delivered to the arena throughout the match and none will be present when the game starts.
The robots have to be completely autonomous, and they shouldn't be designed to destroy other robots. Therefore, only the team with the smartest robots will come out on top!
There are no restrictions on how the robots should be programmed. The only limit is that any manual controlling is forbidden.
You can use state-of-the-art machine learning techniques, or go with more traditional state-based AI. The robots can be controlled through WiFi, and thus, you can run the AI on your personal computer. If you want to go crazy, you could even rent some external servers to run your incredible complicated AI.
We will provide a simple example program that shows you how the control the robot with Python 3. We will also provide a simulator that has a premade structure for reinforcement learning. So, if you like to go with machine learning, the base is already done for you.
The robots don't have any sensors on their own, but there is a camera that streams top-down view of the arena. The video stream can be used to find locations of the robots & cores. The robots must have an ArUco marker on top of them, and the cores are coloured so that they can be easily detected from the image. We will provide OpenCV & Python 3 example that finds the location of the robots and cores from the video stream.
We will provide a simulation that can be used to test and train your AI. The simulation has a base structure for reinforcement learning, so you don't have to do everything from scratch if you would like to use machine learning.
The simulation is built in Unity, and it should work on Windows, macOS and Ubuntu. The reinforcement learning part is built with Unity ML-agents.
It's not mandatory to use the simulation but it is encouraged and we will provide support.
Micro Invaders is designed for hacker teams of 2 to 4 members. Individual applications are accepted, but they are encouraged to form a team and apply.
The competition requires excellent coding skills and ambition to conquer new challenges. We will provide a simple example program that shows how the robots can be controlled through WiFi. There are no restrictions about technologies, so you can freely use your favourite programming languages. Also, there are no restrictions about how the AI should be coded, therefore, you can use machine learning or simple state-based logic. In other words, if you are confident with your coding skills, you are good to go.
Unfortunately, according to the situation of the pandemic, we are not able to accept applications from abroad for this competition. But we encourage you to follow up on our upcoming challenges in 2021
Programming & Robots
Micro Invader provides the competitors with a pre-desgined robot by us. We will provide electronics and a 3D-printable robot model. The dimension of the robot is around 10x10x5cm.
The provided robot kit will consist of:
- 3D-printable body with tracks
- ESP32-based circuit board
- Two motors
Modification of the robot is not allowed. In other words, you are not allowed to buy more powerful motors or build a gigantic cannon for the robot.
The robots will observe the game arena through an overhead camera stream. In the competition, we are using a Rasberry Pi camera to stream the overhead view. When you are practising at home, a standard webcam works well.
Note: Updated on 05.09.
Robots should be completely autonomous during the competition. So, the robots should be able to win the match without any human intervention. If a robot breaks down, gets tilted etc. and the team wants to pull their robot out of the game arena during an ongoing round, the team points will be reduced. Otherwise, there are no technology restrictions.
There are multiple different ways to make your AI. You can use, for instance, reinforcement learning, or you can write a pure state-based AI. We will provide a Unity simulation environment that can be used to train robots with reinforcement learning. Furthermore, you can also use it as a testing platform for your AI.
The robot's heart is the custom ESP32 circuit board with a WiFi antenna. We recommend that you don't write AI logic to the circuit board – instead, you can send motor commands through WiFi from your computer. We will provide an example program that listens to motor commands from WiFi. So, you don't have to know anything about coding ESP32 boards; instead, you can focus on coding the most intelligent AI with your favourite technologies.
Game & Arena
The game is played in a square-shaped 1.5x1.5m arena with multiple cores/balls located at the centre. Two sides of the arena will function as scoring areas, one for each team. The robots will start the match from their team's side.
Teams get points by collecting green energy cores to their base. In the competition, cores are light and hollow balls.
There are two types of cores, energy and corrupt ores. Energy cores increase your points, and corrupt cores can eliminate your team. So, teams are trying to push corrupt cores to the enemy base and capture the energy cores.
One match consists of either 3 or 5 rounds. Each round lasts 2 and a half minutes.
When the time is up, or all energy cores are collected, the team with the most points wins the round. However, if a team receives three corrupt cores, the team loses the round immediately.
A team may lose points if their robot needs human assistance, at the discretion of the judges.
For instance, if a robot gets stuck and has to be moved by hand or if the robot breaks.