Observation Denoising in CYRUS Soccer Simulation 2D Team For RoboCup 2024

TL;DR

Partial and noisy observations in Soccer Simulation 2D hinder decision-making. The paper proposes a denoising framework that uses predictive modelling and intersection analysis to refine object positions, forecasting ball movement via velocity and extrapolating unobserved players using per-player-type tables. The approach iteratively updates estimates as new observations arrive and reports an average improvement of 5.37 cm in opposing-player coordination compared with a baseline. This work demonstrates a practical improvement in observation accuracy that can enhance kicking, passing, and interception decisions in RoboCup SS2D.

Abstract

In the Soccer Simulation 2D environment, accurate observation is crucial for effective decision making. However, challenges such as partial observation and noisy data can hinder performance. To address these issues, we propose a denoising algorithm that leverages predictive modeling and intersection analysis to enhance the accuracy of observations. Our approach aims to mitigate the impact of noise and partial data, leading to improved gameplay performance. This paper presents the framework, implementation, and preliminary results of our algorithm, demonstrating its potential in refining observations in Soccer Simulation 2D. Cyrus 2D Team is using a combination of Helios, Gliders, and Cyrus base codes.

Figures (1)

• Figure 1: The observer is denoising its teammate's position based on our denoising model. The left image is the previous cycle of the right image. the red polygon shows the prediction of our model, while the black sector shows the information that we receive from the server. The blue dot shows the position of the teammate without any denoising, and the red dot shows the position of the player with our new model. In the left image, the blue dot is hidden under the red dot.