Chess and artificial intelligence

Chess and artificial intelligence

An independent project led by Ioannis (John) Kimissis, Professor of Electrical Engineering at Columbia University, on the development, visualization and marketing of a smart chessboard using a number of components, detailed drawings and complex calculations, presents all possible technical problems that need to be solved.

The chessboard operates on sensors of a rheostatic switch (an electric switch operating under the influence of a magnetic field), LEDs and a microcontroller with a magnet and is designed to be high-quality and affordable. The main idea of the design is to create a system that is easy to reproduce and interactive for the user. Using rheostatic switches to determine the position of chess pieces (self-design of the pieces), the LEDs will demonstrate the moves on the computer. Figures moved manually will be highlighted on the computer screen, then the computer makes a move, which in turn is indicated by another LED. In addition to the main function of the chessboard and the software, it can be programmed in such a way that potential errors (incorrect moves) are displayed using another system of light indicators.

The scale and ambitions of the project require constant problem solving. One of the central problems I encountered at the beginning was that the low cost and low power consumption microcontroller – ESP32 – has only 34 contacts for sending important information, however, 256 contacts are required for all rheostat switches and LED indicators to work simultaneously. At first I thought that my ambitions were wrong, that it was impossible to create such a system in such an affordable way; that I needed to make it bigger or more expensive. But I went back to what Elon Musk calls “thinking from the first principles” (that is, “the act of decomposing a process into fundamental parts that you know are true, and building from there”) and compiled a matrix of rays for each rheostatic switch using schematic diagrams. I found the best solution for the RGB matrix – after reviewing the type of technology used, I used ARGB-ribbon lamps, where each lamp can be separately controlled using a non-pixel scheme (I independently learned how to apply this code using a downloadable scheme). By breaking down each potential problem into its component parts, I was able to rethink the various elements that make up the project as small manageable dilemmas that need to be solved.

The problem was also providing power, since I needed to control all the lines, LEDs, Bluetooth, while maintaining a voltage balance. The power supply was too bulky and inconvenient to meet all these requirements and remain affordable. Similarly, the LEDs could be controlled using a driver, but that would increase the cost by $20, and I was determined to stay within the total cost of $50. Again, returning to the creative and logical requirements of the “first principles”, I managed to find alternative solutions – in particular, to develop a new electrical circuit that used two power circuits (voltage and current). By breaking down the task into its component parts, I was also able to revise the project as a whole. Maintaining a detailed spreadsheet with data on all components, references, catalog names, manufacturer and production numbers allowed me to assemble different components and provide a clear diagram so that all elements were properly soldered in place. The feeling of triumph in the face of many problems only added to the reward from the whole project.

Leave a Reply

Your email address will not be published. Required fields are marked *