Soccer-playing Robots Research Articles

Deep learning applied to humanoid soccer robotics: playing without using any color information

The goal of this paper is to describe a vision system for humanoid robot soccer players that does not use any color information, and whose object detectors are based on the use of convolutional neural networks. The main features of this system are the following: (i) real-time operation in computationally constrained humanoid robots, and (ii) the ability to detect the ball, the pose of the robot players, as well as the goals, lines and other key field features robustly. The proposed vision system is validated in the RoboCup Standard Platform League, where humanoid NAO robots are used. Tests are carried out under realistic and highly demanding game conditions, where very high performance is obtained: a robot detection accuracy of 94.90%, a ball detection accuracy of 97.10%, and a correct determination of the robot orientation 99.88% of the times when the observed robot is static, and 95.52% when the robot is moving.

Improvements of SYCU Humanoid Robot

This paper summarizes and describes our SYCU-Legendary team in terms of humanoid robot design and the progress that we have made. Aiming to win a higher ranking on the upcoming 2020 RoboCup Humanoid Kid-Size competition, some changes made on our robot soccer players, such as mechanical design, electrical configuration, sensors and the control system, are introduced. The robot intelligence is improved by optimizing the computation. Also, information about our past participations and honours in humanoid robot soccer championships held in Montreal of Canada, Sydney of Australia, respectively, are presented.

Performing the Kick During Walking for RoboCup 3D Soccer Simulation League Using Reinforcement Learning Algorithm

Nowadays, humanoid soccer serves as a benchmark for artificial intelligence and robotic problems. The factors such as the kicking speed and the number of kicks by robot soccer players are the most significant aims that the participating teams are pursued in the RoboCup 3D Soccer Simulation League. The proposed method presents a kicking strategy during walking for humanoid soccer robots. Achieving an accurate and powerful kicking while robots are moving requires a dynamic optimization of the speed and motion parameters of the robot. In this paper, a curved motion path has been designed based on the robot position relative to the ball and the goal. Ultimately, the robot will be able to kick at the goal by walking along this curve path. The speed and angle of the walking robot are set towards the ball with regard to the robots curved motion path. After the final step of the robot, the accurate and effective adjustment of these two parameters ensures that the robot is located in the ideal position to perform the perfect kick. Due to the noise and walking condition of the robot, it is essential that the speed and angle of motion to be measured more accurately. For this purpose, we use a reinforcement learning model to adjust the robots step size and so does achieve the optimal value of two abovementioned parameters. Using reinforcement learning, robot would learn to pursue an optimal policy to correctly kick towards designated points. Therefore, the proposed method is a model-free and based on dynamic programming. The experiments reveal that the proposed method has significantly improved the team overall performance and robots ability to kick. Our proposed method has been 9.32% successful on average and outperformed the UTAustinVilla agent in terms of goal-scoring time in a non-opponent simulator.

Advancing Research at the RoboCup@Home Competition [Competitions

RoboCup (https://www.robocup.org/) is an annual international robotics competition, founded in 1996, that aims to promote robotics and artificial-intelligence research. The initial challenge was for a team of robots to play soccer: it was the first competition that used soccer games to promote research and education. The goal of RoboCup is as follows: By the middle of the 21st century, a team of fully autonomous, humanoid, robot soccer players shall win a soccer game, complying with the official rules of the Federation Internationale de Football Association, against the winner of the most recent World Cup.

A team of soccer robots for RoboCup competitions in SSL league: system and algorithms

Introduction: The international RoboCup soccer tournaments for robots have been held since the mid-1990s and have attracted a lot of attention from universities all over the world. Their goal is to develop scientific and engineering research in the areas of robotics, artificial intelligence, computer vision, navigation, group and multi-agent interaction of robots and mechatronic devices. Purpose: To create a new generation research bench which assists in developing, training and evaluating soccer robots, as well as developing basic algorithms for managing an individual soccer robot or a group of them. Results: A stand was designed to develop and analyze algorithms for control over a group of robots and navigation in a highly dynamic environment with centralized control organized as follows: the computing center receives data from a video surveillance system, uses the obtained information about the positions of the robots and the ball and about the state of their sensors to calculate control commands (velocity vector, angular velocity, impact signal) which are then transmitted to the robots for execution. Basic control algorithms for robot soccer players, such as: moving to a point, turning on a point, movement around a circle, obstacle avoidance, taking a pass, goalkeeper actions, hitting a set point, as well as a game model for two field players acting against a goalkeeper were implemented and used to test the stand performance, to record the qualification video and to prepare documents for the participation of St. Petersburg State University team in the RoboCup Small Size League final, 2019, in Sydney, Australia.

Unpredefined ball detection algorithm for humanoid soccer robots

Humanoid soccer robots need to adapt to the conditions in human soccer games such as detecting a soccer ball that does not have predefined characteristics such as a definite color and that may blend in with the playing field. For such conditions, the problem cannot be solved by classical detection strategies based on a single colorblock. In this study, the ball color is split into a specific color and a shared color. Two rounds of labelling are used to generate a color lookup table. Color-blocks obtained by pixel-level segmentation are used in a marco-pixel clustering method based on a connecting relationship graph to generate several ball candidates. The best ball object is estimated via the membership function by fuzzy logic. Tests show that the method is able to detect unpredefined balls even in a very disturbed environment and at large distances from the robot and is also able to avoid confusion with the border lines and other robots on the field without excessive computing requirements. The calculations can reach a high framerate of 15 frames per second. This strategy provides an efficient detection method using strictly limited computing resources for robot soccer players. 摘要 为适应人类赛场环境, 仿人足球机器人需能够识别未预先定义足球, 此时足球不再具有固定的与场地鲜明对比的颜色, 因而不能通过传统的单色块识别策略解决。该文将足球颜色划分为专有色和共有色两类, 采用两轮颜色标记方法生成颜色查找表; 在图像像素级联通分割获得色块的基础上, 提出基于图连接关系的宏像素聚类算法, 从而得到若干足球识别候选对象; 再利用模糊逻辑方法中的隶属度函数从候选对象中获得最佳足球识别结果。实验表明在不具备高性能计算硬件平台的情况下, 算法能够在存在大量干扰, 远近距离大幅度变化的条件下准确识别出非预先定义的足球, 避免与边线、机器人等对象的混淆, 且达到每秒15帧的计算速率。该算法能够在严格受限的计算能力下达到高效的足球识别能力, 从而为参赛机器人提供了一种崭新策略。

Distributed Static and Dynamic Circumnavigation Control with Arbitrary Spacings for a Heterogeneous Multi-robot System

Circumnavigation control algorithms enable multiple robots to rotate around a target while they still preserve a circular formation, which is useful in real world applications such as entrapping a hostile target. Specifically, four quantities are involved: the circumnavigation radius, the angular speed, the height and the phase differences among robots, which are termed spacings in this paper. Based on whether these quantities vary or not, the circumnavigation control problem is divided into two categories: the static one and the dynamic one. Corresponding to these two classes, distributed control algorithms are proposed for any number of mobile robots in random 3D positions to circumnavigate a target with arbitrarily given spacings or dynamic spacings. It should be noted that arbitrary spacings or dynamic spacings are useful for a heterogeneous multi-robot system in which robots may possess different kinematics capabilities; robots with higher movement speeds, for instance, can compensate for the insufficiency of those with lower movement speeds by decreasing the corresponding spacings. The robots can only perceive the positions of their two neighbouring robots, so the proposed control algorithms are distributed and scalable. Simulations along with real-robot experiments using soccer-playing robots are conducted to validate the theoretical results.

Variety Wins: Soccer-Playing Robots and Infant Walking.

Although both infancy and artificial intelligence (AI) researchers are interested in developing systems that produce adaptive, functional behavior, the two disciplines rarely capitalize on their complementary expertise. Here, we used soccer-playing robots to test a central question about the development of infant walking. During natural activity, infants' locomotor paths are immensely varied. They walk along curved, multi-directional paths with frequent starts and stops. Is the variability observed in spontaneous infant walking a “feature” or a “bug?” In other words, is variability beneficial for functional walking performance? To address this question, we trained soccer-playing robots on walking paths generated by infants during free play and tested them in simulated games of “RoboCup.” In Tournament 1, we compared the functional performance of a simulated robot soccer team trained on infants' natural paths with teams trained on less varied, geometric paths—straight lines, circles, and squares. Across 1,000 head-to-head simulated soccer matches, the infant-trained team consistently beat all teams trained with less varied walking paths. In Tournament 2, we compared teams trained on different clusters of infant walking paths. The team trained with the most varied combination of path shape, step direction, number of steps, and number of starts and stops outperformed teams trained with less varied paths. This evidence indicates that variety is a crucial feature supporting functional walking performance. More generally, we propose that robotics provides a fruitful avenue for testing hypotheses about infant development; reciprocally, observations of infant behavior may inform research on artificial intelligence.

Decision-making system of soccer-playing robots using finite state machine based on skill hierarchy and path planning through Bezier polynomials

The decision-making mechanism of robotic systems is a fundamental component for the robots’ autonomous control. Like soccer-playing robots, a system must be created such that the strategy is maximized to both chances of scoring and blocking, thus, obtaining a successful play. With the goal of expanding the range of practical applications for robotic systems and contributing to the progress of the design of mobile robots, this study designs a decision-making system using Finite State Machine (FSM) for the action-role selection and Bezier curvature algorithm for the path planning. The presented design follows the skill hierarchy – team strategy layer, role layer and action layer. The study limits the number of players per team to two soccer-playing robots. Each role has a different action to perform as strategized by FSM. The navigation of the robots is established by following a repeatedly modified Bezier curve. The results show that the system using FSM wins the game at 60% chance compared to a system without action planning. Moreover, the Bezier curvature algorithm is effective if both the robot is in the center area of the soccer field. However, the robot spins erroneously at locations near the edges of the soccer field.

Overlapping layered learning

Layered learning is a hierarchical machine learning paradigm that enables learning of complex behaviors by incrementally learning a series of sub-behaviors. A key feature of layered learning is that higher layers directly depend on the learned lower layers. In its original formulation, lower layers were frozen prior to learning higher layers. This article considers a major extension to the paradigm that allows learning certain behaviors independently, and then later stitching them together by learning at the “seams” where their influences overlap. The UT Austin Villa 2014 RoboCup 3D simulation team, using such overlapping layered learning, learned a total of 19 layered behaviors for a simulated soccer-playing robot, organized both in series and in parallel. To the best of our knowledge this is more than three times the number of layered behaviors in any prior layered learning system. Furthermore, the complete learning process is repeated on four additional robot body types, showcasing its generality as a paradigm for efficient behavior learning. The resulting team won the RoboCup 2014 championship with an undefeated record, scoring 52 goals and conceding none. This article includes a detailed experimental analysis of the team's performance and the overlapping layered learning approach that led to its success.

An Online Scalable Approach to Unified Multirobot Cooperative Localization and Object Tracking

In this paper, we present a unified approach for multi-robot cooperative simultaneous localization and object tracking based on particle filters. Our approach is scalable with respect to the number of robots in the team. We introduce a method that reduces, from an exponential to a linear growth, the space and computation time requirements with respect to the number of robots in order to maintain a given level of accuracy in the full-state estimation. Our method requires no increase in the number of particles with respect to the number of robots. However, in our method, each particle represents a full-state hypothesis, leading to the linear dependency on the number of robots of both space and time complexity. The derivation of the algorithm implementing our approach from a standard particle filter algorithm and its complexity analysis are presented. Through an extensive set of simulation experiments on a large number of randomized datasets, we demonstrate the correctness and efficacy of our approach. Through real robot experiments on a standardized open dataset of a team of four soccer-playing robots tracking a ball, we evaluate our method's estimation accuracy with respect to the ground truth values. Through comparisons with other methods based on 1) nonlinear least squares minimization and 2) joint extended Kalman filter, we further highlight our method's advantages. Finally, we also present a robustness test for our approach by evaluating it under scenarios of communication and vision failure in teammate robots.

Sport and exercise psychology in 2050

Sport and exercise psychology by definition describes, explains, and predicts human behaviour. Yet exact predictions of human behaviour are more the exception than the rule and thus it is no wonder that sport and exercise psychologists are not able to predict very well who will be a talent or win a gold medal in 10 years. In the same vein, it is somewhat easier to describe scientific endeavours in hindsight or by analysis of the current state of affairs than to predict what a discipline will be like in, say, 2050. Other disciplines that face similar levels of complexity have, however, tested their models on (near) future events, such as the mathematical prediction of president elections, climate change, or which team will win the next soccer World Cup. Some disciplines have even tried to forecast what will be in 2050, for instance, predicting that soccer-playing robots will win an official soccer game against humankind. In this opinion statement about sport and exercise psychology that starts with the current state of affairs in our field, I predict sport and exercise psychology activities and potential successes in 2050, enabling a discussion of our mission as well as goal setting for the coming years.

The design of an intelligent soccer-playing robot

Purpose – The purpose of this paper is to design intelligent robots operating in such dynamic environments like the RoboCup Middle-Size League (MSL). In the RoboCup MSL, two teams of five autonomous robots play on an 18- × 12-m field. Equipped with sensors and on-board computers, each robot should be able to perceive the environment, make decision and control itself to play the soccer game autonomously. Design/methodology/approach – This paper presents the design of our soccer robots, participating in RoboCup MSL. The mechanical platform, electrical architecture and software framework are discussed separately. The mechanical platform is designed modularly, so easy maintainability is achieved; the electronic architecture is built on industrial standards using PC-based control technique, which results in high robustness and reliability during the intensive and fierce MSL games; the software is developed upon the open-source Robot Operating System (ROS); thus, the advantages of ROS such as modularity, portability and expansibility are inherited. Findings – Based on this paper and the open-source hardware and software, the MSL robots can be re-developed easily to participate in the RoboCup MSL. The robots can also be used in other research and education fields, especially for multi-robot systems and distributed artificial intelligence. Furthermore, the main designing ideas proposed in the paper, i.e. using a modular mechanical structure, an industrial electronic system and ROS-based software, provide a common solution for designing general intelligent robots. Originality/value – The methodology of the intelligent robot design for highly competitive and dynamic RoboCup MSL environments is proposed.

Architecting centralized coordination of soccer robots based on principle solution

Coordination strategy is a relevant topic in multi-robot systems, and robot soccer offers a suitable domain to conduct research in multi-robot coordination. Team strategy collects and uses environmental information to derive optimal team reactions, through cooperation among individual soccer robots. This paper presents a diagrammatic approach to architecting the coordination strategy of robot soccer teams by means of a principle solution. The proposed model focuses on robot soccer leagues that possess a central decision-making system, involving the dynamic selection of the roles and behaviors of the robot soccer players. The work sets out from the conceptual design phase, facilitating cross-domain development efforts, where different layers must be interconnected and coordinated to perform multiple tasks. The principle solution allows for intuitive design and the modeling of team strategies in a highly complex robot soccer environment with changing game conditions. Furthermore, such an approach enables systematic realization of collaborative behaviors among the teammates.

UT Austin Villa 2014: RoboCup 3D Simulation League Champion via Overlapping Layered Learning

Layered learning is a hierarchical machine learning paradigm that enables learning of complex behaviors by incrementally learning a series of sub-behaviors. A key feature of layered learning is that higher layers directly depend on the learned lower layers. In its original formulation, lower layers were frozen prior to learning higher layers. This paper considers an extension to the paradigm that allows learning certain behaviors independently, and then later stitching them together by learning at the "seams" where their influences overlap. The UT Austin Villa 2014 RoboCup 3D simulation team, using such overlapping layered learning, learned a total of 19 layered behaviors for a simulated soccer-playing robot, organized both in series and in parallel. To the best of our knowledge this is more than three times the number of layered behaviors in any prior layered learning system. Furthermore, the complete learning process is repeated on four different robot body types, showcasing its generality as a paradigm for efficient behavior learning. The resulting team won the RoboCup 2014 championship with an undefeated record, scoring 52 goals and conceding none. This paper includes a detailed experimental analysis of the team's performance and the overlapping layered learning approach that led to its success.

RoboCup Soccer Leagues

RoboCup was created in 1996 by a group of Japanese, American, and European artificial intelligence and robotics researchers with a formidable, visionary long-term challenge: By 2050 a team of robot soccer players will beat the human World Cup champion team. In this article, we focus on RoboCup robot soccer, and present its five current leagues, which address complementary scientific challenges through different robot and physical setups. Full details on the status of the RoboCup soccer leagues, including league history and past results, upcoming competitions, and detailed rules and specifications are available from the league homepages and wikis.

Design and Kick Analysis of a Soccer Playing Robot

For the implementation of an idea, it is essential to have a full engineering system design, simulation and analysis. MISTBOY is such a dream-bot which can perform football skills like a human being. This is an on going humanoid robot project, initially having 17 DoF, 55 cm height and 5 kg weight. In this paper, the initial step towards the making of a humanoid robot is discussed and the balancing and structural analysis through mathematical derivation and simulation are also presented here with good result. This paper is focused on forward kinematics, trajectory planning, balancing, force and torque calculation and kick analysis.

The Research Progress of Soccer Learning Mechanisms for Robot

This study makes a review of learning theories, methods, technologies and applications about robot soccer player’s behavior how to implement complex tasks. And pointed out that the limitations of its existence, as well as learning strategies of robot soccer. As a test bed of multi-agent system research of robot soccer system, many researchers conducted a study on the technology from different sides, and have achieved some success. The study indicated that researches currently on robot soccer system includes soccer robot architecture, collaboration, under the dynamic environment of a robot, sensor data fusion, complex reasoning and action learning in task, opponent modeling, and more.

Robot soccer.

Robot soccer is a test bed for a variety of robotic and Artificial Intelligence (AI) methods. Its relevance to Cognitive Science is that it confronts the designer with a task that requires the integration of almost all aspects of AI to create an agent that is capable of working in a complex, dynamic environment inhabited by other agents, some of which are cooperative and others competitive. We describe the main elements that make up a robot soccer player and how these players associate to create effective teams. We pay special attention to the architecture of the players. WIREs Cogn Sci 2010 1 824-833 For further resources related to this article, please visit the WIREs website.

It shoots - it scores! [RoboCup

The prime driver for this sporting possibility is the RoboCup international robotics competition. Founded in 1997, its official aim is that, by 2050, a team of fully autonomous humanoid robot soccer players will play against the winner of the most recent World Cup under the rules of football's governing body FIFA -and win.