Technology Category
Search and Optimization: Sampling/Simulation-based SearchHumans and AI: Human-Aware Planning and Behavior PredictionMachine Learning: Evolutionary Learning
Evolving Neural Controllers for Xpilot-AI Racing Using Neuroevolution of Augmenting Topologies
🤖 AI Summary
To address the control challenges posed by high-fidelity spatial physics simulation—including inertia, friction, and gravity—in the newly introduced racing mode of the Xpilot-AI platform, this paper proposes an enhanced Topological NeuroEvolution of Augmenting Topologies (NEAT) method tailored for adaptive navigation on dynamic tracks. Our approach innovatively extends NEAT to a multi-agent parallel evaluation framework with dynamic circuit-structure evolution, jointly optimizing both controller topology and connection weights to enable autonomous emergence of human-like driving policies. Experimental results demonstrate that the evolved controllers achieve up to a 32% reduction in lap time on complex tracks, exhibiting robust high-speed cornering, real-time speed modulation, and other advanced racing capabilities. These findings validate the effectiveness and generalizability of our method for end-to-end racing control under stringent physical constraints.
Technology Category
Application Category
📝 Abstract
This paper investigates the development of high-performance racing controllers for a newly implemented racing mode within the Xpilot-AI platform, utilizing the Neuro Evolution of Augmenting Topologies (NEAT) algorithm. By leveraging NEAT's capability to evolve both the structure and weights of neural networks, we develop adaptive controllers that can navigate complex circuits under the challenging space simulation physics of Xpilot-AI, which includes elements such as inertia, friction, and gravity. The racing mode we introduce supports flexible circuit designs and allows for the evaluation of multiple agents in parallel, enabling efficient controller optimization across generations. Experimental results demonstrate that our evolved controllers achieve up to 32% improvement in lap time compared to the controller's initial performance and develop effective racing strategies, such as optimal cornering and speed modulation, comparable to human-like techniques. This work illustrates NEAT's effectiveness in producing robust control strategies within demanding game environments and highlights Xpilot-AI's potential as a rigorous testbed for competitive AI controller evolution.
Problem
Research questions and friction points this paper is trying to address.
Develop high-performance racing controllers for Xpilot-AI
Optimize neural networks for complex space simulation physics
Improve lap times and racing strategies using NEAT
Innovation
Methods, ideas, or system contributions that make the work stand out.
Utilizes NEAT for neural network evolution
Adapts to complex Xpilot-AI physics
Enables parallel multi-agent evaluation
🔎 Similar Papers
No similar papers found.
