CheckJump: An Approach to Real-Time Pathfinding for 2D Grid-Based Platformer Games
Ibnu Athaillah (a*), Moch. Kholil (b), Rafika Akhsani (c), Ismanto Ismanto (d), Heri Waspada (e), Muchamad Saiful Muluk (f)

Department of Multimedia
Akademi Komunitas Negeri Putra Sang Fajar Blitar
Jl. Dr. Sutomo 29, Blitar 66133, Indonesia

Abstract

This research proposed a solution for game developers that are building a 2D grid platformer game that requires characters that are able to intelligently find a way to navigate levels through various movements. This research is based on an unannounced platformer game that is currently under development and uses the defined input handling and jump mechanics. The pathfinding task is done by utilizing the A* algorithm and movement sequence with Finite State Machine. Vertical movement predictions can be made with precision by using the built-in Physics system to replicate physics-affected movements outside of the real-time frame. This solution is tested on varied levels, each with a different size and complexity. In terms of capability, NPC has been shown to be capable of performing a sequence of horizontal or vertical movements in order to reach its target. In terms of performance, there isn^t any noticeable impact on memory usage. However, the number of extra frames required for mapping is proportional to the number of cells in the level. Fortunately, this is a task that can be done sparingly without impacting functionality. The pathfinding process itself, including the NPC movement, does not have any noticeable performance impact. This technique has the flexibility of real-time adjustments to changes in the level layout and can be used as an option for developers looking for a platformer pathfinding solution.

Keywords: Platformer Game- Non-Player Character- Pathfinding-

Topic: Artificial Intelligence and Data Science