EE106A PROJECT
GOMOKU ROBOT
GOMOKU X BAXTER
What is Gomoku?
Gomoku is a board game played with Go pieces (black and white stones) on a 15-by-15 Go board. Players alternate turns placing a stone of their color on an empty intersection. The winner is the first player to form an unbroken chain of five stones horizontally, vertically, or diagonally. (from wikipedia)
​
What is Baxter robot?
Baxter is an industrial robot built by Rethink Robotics. It has two-armed and an animated face. It is 3 feet tall and weighs 165 lbs without its pedestal; with its pedestal it is between 5'10" – 6'3" tall and weighs 306 lbs. It is used for simple industrial jobs such as loading, unloading, sorting, and handling of materials. (from wikipedia)
END GOAL
Our goal for the project is to build an AI-based robot that can play Gomoku with human players on a physical chessboard.
​
The project, in the end, should be able to determine by the vision sensors as well as the external Gomoku engine where to place its next step, according to the current board scenario and a robot controller that controls Baxter to place/mark a real-life stone/symbol on the board. Also in order to increase the functionality of the robot, a cheating detective feature will be built-in as well. If the human player doesn’t follow the Gomoku rules to make valid steps, the eventual system should be able to detect the invalid steps and make an attempt to reverse it. The robot should not make its next step until the human player takes a valid step.
AN INTERESTING PROJECT
Recalling the first game of the Google AlphaGo against the world’s top Go player Lee Sedol, the steps of AlphaGo were indeed taken by an actual human player. It’s interesting that in those games, a human player was manipulated by AI. Since the invention of computers, the computers have been manipulated by the human for more than half a century, but today they start taking it over due to their smartness. It’s awkward that someday in the future humans will be manipulated by computer AI in order to make it work. They should be able to finish the work by themselves, aren’t they? When we see the game between Lee Sedol and AlphaGo, we questioned myself that why AlphaGo is so smart but cannot take its own steps. If we were Lee Sedol, we’d like to play against AlphaGo directly, instead of a human player who was completely manipulated by the computer. In the labs, we met the Baxter. Maybe the Baxter can be physical chess game player run by chess game AI engine such as AlphaGo.
​
In order to make Baxter achieves the desired goal, firstly the webcam needs to scan the entire chessboard, interprets the location of the chessboard as well as the current locations of all Gomoku pieces on it. The built-in cheating detective will be able to determine if the human player has been taken his/her valid step. If the test has been passed, the Gomoku engine will be run and make the decision of the next step. In the end, the Baxter should be able to take her step on the physical chessboard.
IN THEÂ REAL-WORLD
We may add our appliance to the well-known Go playing AI, AlphaGo, which is only an AI program and relies on a person to play the stone on board. We can add the Baxter robot to the AlphaGo AI the robot can put and pick the stones to and from the board by itself without the help of another human. This addition will make it more modern and fancy.