The Leges Motus project began as a simple multiplayer game of laser tag, set in 2D zero-gravity arenas, written in a custom engine in C++ with SDL and OpenGL. The game combines simple, fast-paced gameplay with team strategy and tactics in several game modes, with three different weapons available to enable several play-styles. Later, we researched and implemented AI features in the game.
The AI for Leges Motus is primarily a fuzzy logic system (Klir and Yuan 1996), similar to a utility-theory-based AI (Russell and Norvig 2016). The agents calculate their strategy based on a hand-authored combination of multiple perceived game state features. The AI also requires a unique pathfinding solution. Because of the zero-gravity environment, agents cannot change direction in mid-air. Therefore, we use raycast sweeps to calculate a connected graph of possible jump points on each wall, enabling us to perform an A* pathfinding search (Hart et. al. 1968) on the resulting graph.
Leges Motus is available for Mac, Linux, and Windows as a binary for the stable version, 0.4, and the code is open source for both the stable version and the unstable, refactored version 0.5, which is missing some UI and game mode features. Development is currently on indefinite hiatus.