AutoChess
Automated Physical Chessboard & Reinforcement Learning Agent
State Perception
Sensor Grid Tracking
Engine Reasoning
Minimax / RL Evaluation
Motion Planning
Collision-Free Path Solver
Physical Actuation
CoreXY Magnetic Gantry
Overview
AutoChess bridges physical tactile gameplay with digital chess computation. The system combines an embedded sensor grid for move recognition, a CoreXY stepper-motor magnetic gantry under the board to physically glide pieces across the surface, and an intelligent chess engine evaluating optimal moves in real time.
The Problem
Playing digital chess lacks tactile physical engagement, while playing physical boards alone provides no interactive opponent without manually moving the opponent's pieces.
The Solution
Engineered an autonomous physical chessboard combining an electromagnet mounted to a dual-axis CoreXY gantry beneath the board with real-time sensory tracking and an intelligent chess engine pipeline.
Technology
Hardware & Robotics
- Arduino
- Raspberry Pi
- CoreXY Gantry
- Stepper Motors
- Electromagnet
Software & Intelligence
- Python
- Python-Chess
- Stockfish API
- C++
State Perception
Sensor Grid Tracking
Engine Reasoning
Minimax / RL Evaluation
Motion Planning
Collision-Free Path Solver
Physical Actuation
CoreXY Magnetic Gantry
Key Capabilities
Physical Move Actuation
Sub-surface CoreXY gantry with an electromagnet smoothly guides captured and active pieces without physical human intervention.
Board State Sensing
Magnetic reed switch / Hall effect sensor grid detects piece lifting and placement in real time.
Intelligent Engine Integration
Interfaced with computational chess engines to compute difficulty-adaptive countermoves.
Obstacles Faced
- Preventing piece collisions when sliding pieces across crowded board squares by implementing a pathfinding algorithm that navigates pieces along square boundaries.
- Eliminating magnetic field crosstalk across neighboring squares to ensure accurate state sensing.
Key Takeaways
- Hardware-software synchronization requires defensive state validation and debounce filters on physical sensors.
- Spatial motion planning algorithms must account for piece geometry and boundary tolerances.
Future Improvements
- Online multiplayer synchronization over WebSockets allowing remote opponents to play on physical boards.
- Voice command integration for hands-free move declaration.