RoboCup 2014: Robot World Cup XVIII

Bipedal walking is one of the most essential skills required to play soccer with humanoid robots. Superior walking speed and stability often gives teams the winning edge when their robots are the first at the ball, maintain ball control, and drive the ball towards the opponent goal with sure feet. In this contribution, we present an implementation of our Capture Step Framework on a real soccer robot, and show robust omnidirectional walking. The robot not only manages to locomote on an even surface, but can also cope with various disturbances, such as pushes, collisions, and stepping on the feet of an opponent. The actuation is compliant and the robot walks with stretched knees.

The definition of high quality datasets for benchmarking single components and entire systems in intelligent robots is a fundamental task for developing, testing and comparing different technical solutions. In this paper, we describe the methodology adopted for the acquisition and the creation of a spoken corpus for domestic and service robots. The corpus has been inspired by and acquired in the RoboCup@Home setting, with the involvement of RoboCup@Home participants. The annotated data set is publicly available for developing, testing and comparing speech understanding functionalities of domestic and service robots, not only for teams involved in RoboCup@Home or in other competitions, but also for research groups active in the field. We regard the construction of the dataset as a first step towards a full benchmarking methodology for spoken language interaction in service robotics.

The UT Austin Villa team, from the University of Texas at Austin, won the 2014 RoboCup 3D Simulation League, finishing with an undefeated record. During the course of the competition the team scored 52 goals and conceded none. Additionally the team won the RoboCup 3D Simulation League technical challenge by accumulating the most points over a series of three league challenges: drop-in player, running, and free challenge. This paper describes the changes and improvements made to the team between 2013 and 2014 that allowed it to win both the main competition and the technical challenge.

The Small Size League is one of the important events in RoboCup Soccer. ZJUNlict got the first place in Robocup 2014. In this paper, we introduce the improvement we have made in the past year. We describe the overview of the mechanical design, show the design of the protector for Infrared emission tube as well as the shield for wheels. Simulation is given to show how our design works. Then the lower level firmware architecture is illustrated. The dynamics analysis of the robot is presented to help improving the robots’ performance and reducing motion deviation in y axis. Finally we present how we organize defense employing Close-Marking defense along with Zone defense imitating human player.

In this paper we discuss improvements in mechanical, electrical and software design, which we did to become RoboCup 2014 world champion. Regarding hardware and control our progress includes first steps towards improved passing accuracy via velocity feedback control on the shooting lever. In terms of intelligent gameplay we have worked on creating possibilities for in-game optimization of strategic decisions. Via qr-code detection we can pass coaching instructions to our robots and with a basic machine learning algorithm success and failure after free-kicks is taken into account. In the final part of this paper we briefly discuss progress we have made in designing a four-wheeled soccer robot with a suspension system.

Winning the Robocup SPL World Championship is not accomplished in just one year, nor is it just a matter of writing effective software. Our success can also be attributed to the accumulation of experience since 1999, strong institutional support, and dedicated collaborative teamwork. This paper summarises the key contributing innovations from the time the software was rewritten in 2010, and provides some insight into team organisation. In this paper it is not possible to cover all aspects and intricacies of the complex systems comprising the

rUNSWift

software. We have therefore included an extensive list of references to our technical reports that provide detailed accounts of the research, algorithms and results over the last 5 years. All the reports are available on one website for easy access and make reference to many external publications, including those from other teams in our league.

In this paper, we describe the system design of the robots developed by our Team CIT Brains for the RoboCup soccer humanoid KidSize league. We have been participating in the Humanoid League for eight years. Two years ago, we redesigned the system to put a large weight on maintainability and usability. In RoboCup 2014, we won the first prizes of 4on4 soccer and technical challenge. Consequently, we were awarded the Louis Vuitton Humanoid Cup. The system we developed has high mobility, well-designed control system, position estimation by a monocular camera, user-friendly interface and a simulator. The robot can walk speedily and robustly. It also has a feedback system with a gyro sensor to prevent falls. It detects positions of landmarks by color-based image processing. A particle filter is employed to localize the robot in the soccer field fusing the motion model and landmark observation.

The RoboCup Soccer Humanoid AdultSize League is a very challenging league that requires fast and reliable locomotion, a robust vision and localization system, and long term planning to score in a limited attack window. As a result, many adult-size robots are now highly optimized for the soccer task, with very lightweight construction that enables nimble locomotion with low power actuators. However, such a soccer optimized humanoid design excludes the possibility of using the humanoid robot for more general tasks that include, for instance, dextrous manipulation and navigating over rugged terrain. In this paper, we describe how we utilize the general purpose and full sized humanoid robot, THOR-OP (Tactical Hazardous Operations Robot - Open Platform), which was originally developed to compete at the DARPA Robotics Challenge (DRC) for the 2014 RoboCup AdultSize league robotic soccer competition.

This paper describes MRESim, a multi-robot exploration simulator which aims to provide a middle ground between the RoboCup Agent and Virtual Robot competitions. A detailed description of this new infrastructure is provided, followed by examples and case studies of successful research outcomes arising from the use of MRESim. Our work on MRESim won the 2014 Infrastructure competition of the RoboCup Rescue Simulation League.

Participating in the RoboCup Rescue Real Robot League competition for approximately 5 years, the members of Team Hector Darmstadt have always focused on robot autonomy for Urban Search and Rescue (USAR). In 2014, the team won the RoboCup RRL competition. This marked the first time a team with a strong focus on autonomy won the championship. This paper describes both the underlying research and open source developoments that made this success possible as well as ongoing work focussed on increasing rescue robot performance.

In this paper, we present the details of our team WrightEagle@Home’s approaches. Our

KeJia

robot won the RoboCup@Home competition 2014 and accomplished two tests which have never been fully solved before. Our work covers research issues ranging from hardware, perception and high-level cognitive functions. All these techniques and the whole robot system have been exhaustively tested in the competition and have shown good robustness.

In this paper we summarise the approach of the smARTLab@Work team that has won the 2014 RoboCup@Work competition. smARTLab (swarms, multi-agent, and robot technologies and learning lab) is a robotics research lab that is part of the Agent ART group at the computer science department of the University of Liverpool. This team has won the competition for the second year in a row. The team previously competed as swarmlab@Work and changed name after the move of professor Tuyls and his research group from Maastricht University to the University of Liverpool, UK. The various techniques that have been combined to win the competition come from different computer science domains, including machine learning, (simultaneous) localisation and mapping, navigation, object recognition and object manipulation. While the RoboCup@Work league is not a standard platform league, all participants use a (customised) KUKA youBot. The stock youBot is a ground based platform, capable of omnidirectional movement and equipped with a five degree of freedom arm featuring a parallel gripper. We present our adaptations to the robot, in which the replacement of the gripper was the most important upgrade comparing to the version of the robot that was used last year.

The RoboCup Logistics League is one of the youngest application- and industry-oriented leagues. Even so, the complexity and level of difficulty has increased over the years. We describe decisive technical and organizational aspects of our hardware and software systems and (human) team structure that made winning the RoboCup and German Open competitions possible in 2014.

Mobile manipulators are viewed as an essential component for making the factory of the future become a reality. RoboCup@Work is a competition designed by a group of researchers from the RoboCup community and focuses on the use of mobile manipulators and their integration with automation equipment for performing industrially-relevant tasks. The paper describes the design and implementation of the competition and the experiences made so far.

The selection of an appropriate competition format is critical for both the success and credibility of any competition, both real and simulated. In this paper, the automated parallelism offered by the RoboCupSoccer 2D simulation league is leveraged to conduct a 28,000 game round-robin between the top 8 teams from RoboCup 2012 and 2013. A proposed new competition format is found to reduce variation from the resultant statistically significant team performance rankings by 75 % and 67 %, when compared to the actual competition results from RoboCup 2012 and 2013 respectively. These results are statistically validated by generating 10,000 random tournaments for each of the three considered formats and comparing the respective distributions of ranking discrepancy.

Personal robots will contribute mobile manipulation capabilities to our future smart homes. In this paper, we propose a low-cost object localization system that uses static devices with Bluetooth capabilities, which are distributed in an environment, to detect and localize active Bluetooth beacons and mobile devices. This system can be used by a robot to coarsely localize objects in retrieval tasks. We attach small Bluetooth low energy tags to objects and require at least four static Bluetooth receivers. While commodity Bluetooth devices could be used, we have built low-cost receivers from Raspberry Pi computers. The location of a tag is estimated by lateration of its received signal strengths. In experiments, we evaluate accuracy and timing of our approach, and report on the successful demonstration at the RoboCup German Open 2014 competition in Magdeburg.

The recognition of objects is a relevant ability in service robotics, especially in manipulation tasks. There are many different approaches to object recognition, but they have not been properly analyzed and compared by considering real conditions of manipulation tasks in domestic setups. The main goal of this paper is to analyze some popular object recognition methods and to compare their performance in realistic manipulation setups. Object recognition methods based on SIFT, SURF, VFH, OUR-CVFH and color histogram descriptors are considered in this study. The results of this comparison can be of interest for researchers working in the development of similar systems (e.g. RoboCup @home teams).

RoboCup is particularly well-known for its soccer leagues, but there are an increasing number of application leagues. The newest one is the Logistics League where groups of robots take on the task of in-factory production logistics. It has two unique aspects: a game environment which itself acts as an agent and a focus on planning and scheduling in robotics. We propose a simulation based on Gazebo that takes these into account. It uses the exact same referee box to simulate the environment reactions similar to the real game and it supports multiple levels of abstraction that allow to focus on the planning with a high level of abstraction, or to run the full system on simulated sensor data on a lower level for rapid integration testing. We envision that this simulation could be a basis for a simulation sub-league for the LLSF to attract a wider range of participants and ease entering the robot competition.

In this paper, we present an automatic approach for the kinematic calibration of the humanoid robot NAO. The kinematic calibration has a deep impact on the performance of a robot playing soccer, which is walking and kicking, and therefore it is a crucial step prior to a match. So far, the existing calibration methods are time-consuming and error-prone, since they rely on the assistance of humans. The automatic calibration procedure instead consists of a self-acting measurement phase, in which two checkerboards, that are attached to the robot’s feet, are visually observed by a camera under several different kinematic configurations, and a final optimization phase, in which the calibration is formulated as a non-linear least squares problem, that is finally solved utilizing the

Levenberg-Marquardt

algorithm.

We propose two real-time sound recognition approaches that are able to distinguish a predefined whistle sound on a NAO robot in various noisy environments. The approaches use one, two, and four microphone channels of a NAO robot. The first approach is based on a frequency/band-pass filter whereas the second approach is based on logistic regression. We conducted experiments in six different settings varying the noise level of both the surrounding environment and the robot itself. The results show that the robot will be able to identify the whistle reliability even in very noisy environments.

Intelligent robots require a semantic map of the surroundings for applications such as navigation and object localization. In order to generate the semantic map, previous works mainly focus on the semantic segmentations on the single RGB-D images and fuse the results by a simple majority vote. However, single image based semantic segmentation algorithms are prone to producing inconsistent segments. Little attentions are paid to the consistency over the semantic map. We present a spatio-temporally consistent semantic mapping approach which can generate the temporal consistent segmentations and enforce the spatial consistency by Dense CRF model. We compare our temporal consistent segment algorithm with the state-of-art approach and generate our semantic map on the NYU v2 dataset.

Creating, maintaining, and deducing accurate world knowledge in a dynamic, complex, adversarial, and stochastic environment such as the RoboCup environment is a demanding task. Knowledge should be represented in real-time (i.e., within ms) and deduction from knowledge should be inferred within the same time constraints. We propose an extended assertional formalism for an expressive

$$\mathcal {SROIQ}(\mathcal {D})$$

Description Logic to represent asserted entities in a lattice structure. This structure can represent temporal-like information. Since the computational complexity of the classes of description logic increases with its expressivity, the problem demands either a restriction in the expressivity or an empirical upper bound on the maximum number of axioms in the knowledge base. We assume that the terminological/relational knowledge changes significantly slower than the assertional knowledge. Henceforth, using a fixed terminological and relational formalisms and the proposed lattice structure, we empirically bound the size of the knowledge bases to find the best trade-off in order to achieve deduction capabilities of an existing description logic reasoner in real-time. The queries deduce instances using the equivalent class expressions defined in the terminology. We have conducted all our experiments in the RoboCup 3D Soccer Simulation League environment and provide justifications of the usefulness of the proposed assertional extension. We show the feasibility of our new approach under real-time constraints and conclude that a modified FaCT++ reasoner empirically outperforms other reasoners within the given class of complexity.

Control of a servomotor is a challenging real-time problem. The embedded microcontroller is responsible for fast and precise actuation of the motor shaft, and must handle communication with a master controller as well. If additional tasks such as temperature monitoring are desirable, they must take place often enough to be useful, but not so frequently that they interfere with the operation of the servo. Since microcontrollers have limited multi-tasking capabilities, it becomes difficult to perform all of these tasks at once. It was our goal to create servo firmware with high communication speeds for humanoid robots, and our solution is generalizable to non-humanoid motor control as well. In this paper, we present an event-driven operating system for the Robotis AX-12 servomotor. By using interrupts to drive functionality that would otherwise require polling, our operating system meets the real-time constraints associated with controlling a servomotor.

The RoboCup Rescue project is based on the situations that occurred during the Great Hanshin Earthquake in 1995. Various types of disasters occurred after the Hanshin-Awaji Earthquake, and the experiences from managing such disasters has shown that evacuation with prompt and appropriate information at the initial period of the disaster is as important as the rescue operations during the disasters. In this paper, we discuss validation and verification of the agent based systems that can simulate the behaviors of individuals and collective evacuation behavior during emergency situations. Collective behavior is difficult to verify by executing evacuation drills in the real world. The effectiveness of evacuation simulations is shown and a plan of experiments at RoboCup venues is proposed to challenge the validation and verification of social simulation and to prove the usefulness of as real-world applications.

RoboGrams is a lightweight and efficient message passing architecture that we designed for the RoboCup domain and that has been successfully used by the Northern Bites SPL team. This unique architecture provides a framework for separating code into strongly decoupled modules, which are combined into configurable dataflow graphs. We present several different architecture types and preexisting message passing implementations, but among all of these, we contend that RoboGrams’ features make it particularly well suited for use in RoboCup. As a success story, we describe the Northern Bites’ use of RoboGrams and the benefits it has provided to a single team, but we also suggest that it could help SPL teams collaborate in the future.

The work presented in this paper is motivated by the goal of dependable autonomous navigation of mobile robots. This goal is a fundamental requirement for having autonomous robots in spaces such as domestic spaces and public establishments, left unattended by technical staff. In this paper we tackle this problem by taking an optimization approach: on one hand, we use a Fast Marching Approach for path planning, resulting in optimal paths in the absence of unmapped obstacles, and on the other hand we use a Dynamic Window Approach for guidance. To the best of our knowledge, the combination of these two methods is novel. We evaluate the approach on a real mobile robot, capable of moving at high speed. The evaluation makes use of an external ground truth system. We report controlled experiments that we performed, including the presence of people moving randomly nearby the robot. In our long term experiments we report a total distance of 18 km traveled during 11 h of movement time.

Robots capable of assisting elderly people in their homes will become indispensable, since the world population is aging at an alarming rate. A crucial requirement for these robotic caregivers will be the ability to safely interact with humans, such as firmly grasping a human arm without applying excessive force. Minding this concern, we developed a reactive grasp that, using tactile sensors, monitors the pressure it exerts during manipulation. Our approach, inspired by human manipulation, employs an architecture based on different grasping phases that represent particular stages in a manipulation task. Within these phases, we implemented and composed simple components to interpret and react to the information obtained by the tactile sensors. Empirical results, using a Care-O-bot 3

®

with a Schunk Dexterous Hand (SDH-2), show that considering tactile information can reduce the force exerted on the objects significantly.

Coordination is an important aspect of multi-agent teamwork. In the context of robot soccer in the RoboCup Standard Platform League, our focus is on the

coach

as an external observer of the team, aiming to provide his teammates with effective tactical advice during matches. The coach problem can be approached from different angles: in order to adapt the behaviour of his teammates, he should at first be able to perform

plan recognition

on their observable actions. Furthermore, in providing them with appropriate advice, he should still adhere to the norms and regulations of the match to prevent penalties for his team. Also, when teammates’ profiles and attributes are unknown or the system is only partially observable, coordination should be more ‘ad hoc’ to ensure robustness of the Multi-Agent System (MAS). In this work, we present a formalization of the problem of designing a coach in robot soccer, employing a temporal deontic logical framework. The framework is based on

agent organizations

[

10

], in which social coordination and norms play an important part.

Many RoboCuppers share a dream: a team of fully autonomous humanoid robot soccer players that is capable of playing soccer games against human players by 2050. The demonstration of a human versus robot soccer game at RoboCup 2007 was an exciting display of the progress that has been in RoboCup community since 1997 in terms of the robot’s autonomy, sensing ability, and physical features. This game also uncovered several new issues. For example, do human soccer players play games against robots in the same way that they would if they were playing against human players?

In this study, we investigate the features of RoboCup soccer leagues and examine whether these features are necessary and sufficient to realize the dream of RoboCuppers. We compare the current RoboCup soccer leagues to human soccer leagues and discuss metrics that indicate the similarity between robot soccer games and human soccer games. In other words, we discuss our progress toward realizing the dream. In addition, we propose amendments of laws for human and robot soccer that will deal with human-related issues and facilitate research toward achieving the dream.

We describe our previous and current efforts towards achieving an unusual personal RoboCup goal: to train a full team of robots directly through demonstration, on the field of play at the RoboCup venue, how to collaboratively play soccer, and then use this trained team in the competition itself. Using our method, HiTAB, we can train teams of collaborative agents via demonstration to perform nontrivial joint behaviors in the form of hierarchical finite-state automata. We discuss HiTAB, our previous efforts in using it in RoboCup 2011 and 2012, recent experimental work, and our current efforts for 2014, then suggest a new RoboCup Technical Challenge problem in learning from demonstration.

Relative localisation of mobile robots can provide useful information to applications, from formation control, to joint exploration and inspection. One way to obtain relative localisation is to measure distances between the multiple robots. In this scope, distance estimates based on RF ranging data can be beneficial for small/inexpensive communicating robots that have no other means of measuring distances, or as disambiguation of multiple hypothesis in high accuracy localisation systems. In this work, we present a technique of estimating the relative positions of simple mobile robots in a small team using the distance information that can be captured by a wireless transceiver, only. Simulation results with a team of five mobile robots show that we can estimate their relative positions with an average accuracy of 1.3 m without any fixed reference and using RF information, only. The main contribution of our work is that we can provide consistent reliability information as the covariance of the obtained positions.

This paper shows the results of applying machine learning techniques to the problem of predicting soccer plays in the Small Size League of RoboCup. We have modeled the task as a multi-class classification problem by learning the plays of the STOx’s team. For this, we have created a database of observations for this team’s plays and obtained key features that describe the game state during a match. We have shown experimentally, that these features allow two learning classifiers to obtain high prediction accuracies and that most miss-classified observations are found early on the plays.

Plenty of work based on the Rapidly-exploring Random Trees (RRT) algorithm for path planning in real time has been developed recently. This is the most used algorithm by the top research teams in the Small Size League of RoboCup. Nevertheless, we have concluded that other simpler alternatives show better results under these highly dynamic environments. In this work, we propose a new path planning algorithm that meets all the robotic soccer challenges requirements, which has already been implemented in the STOx’s team for the RoboCup competition in 2013. We have evaluated the algorithm’s performance using metrics such as the smoothness of the paths, the traveled distance and the processing time and compared it with the RRT algorithm’s. The results showed improved performance over RRT when combined measures are used.

Preparing for robot soccer competitions by empirically evaluating different possible game strategies has been rather limited in leagues using real robots. Such limitation comes from factors related to the difficulty of extensively experimenting with games with real robots, such as their inevitable wear and tear and their usual limited number. RoboCup real robot teams have therefore developed simulation environments to enable experimentation. However, in order to run complete games in such simulation environments, an automated referee is needed. In this paper, we present AutoRef, as a contribution towards a complete automated referee for the RoboCup Small-Size League (SSL). We have developed and used AutoRef in an SSL simulation to run full games to evaluate different strategies, as we illustrate and show results. AutoRef is designed as a finite-state machine that transitions between the states of the game being either on or required to stop. AutoRef purposefully only uses the same visual and game information provided in SSL games with physical robots, which it uses to compute the features needed by the rules and to make decisions to transition between its states. Due to this real input to AutoRef, we have partially applied it to games of the physical robots. As AutoRef does not include all the rules of the real SSL games, we currently view it as an aid to human referees of SSL games, and discuss the challenges in automating several specific SSL game rules. AutoRef could be extended to other RoboCup real soccer leagues if a combined view of the game field, ball, and players is available.

Usually the legs of humanoids capable of omnidirectional walking are not underactuated. In other words each one of the six degrees of freedom of the torso can be commanded independently from the leg joint angles. However the NAO humanoid robot has a coupled joint at the hips, which makes 11 degrees of freedom instead of 12 for the locomotor apparatus. As a consequence the trunk of the robot has only 5 independent degrees of freedom when the positions of both feet are fixed, and each leg cannot be commanded independently to execute walking steps. Up to now only bypass solutions have been proposed, where the coupled joint angles are not calculated exactly. This paper describes an analytical solution to determine the exact angle to be applied to the coupled joint. The method uses the positions of both foot ankles in the trunk reference frame and the angle between footprints as inputs, and calculates the yaw angle of the trunk. The solution was demonstrated in a dynamics simulator using the NAO model.

Passive walkers are robots that can produce a stable cyclical movement similar to walking on mildly inclined surfaces. In recent years, various investigations have been conducted on this subject. Since this is a new field of research, the effect of structural parameters on the movement of these walkers and the development of more human-like movement models can be further investigated. This paper compares three popular models for passive dynamic walkers: Garcia’s mode, Wisse’s model, and our own extended Wisse’s model with arm. Our research shows that the extended models lead to more energy efficient and stable walking.

Effective object motion estimation is significant to improve the performance of soccer robots in RoboCup Middle Size League. In this paper, a hybrid vision system is constructed by combining omnidirectional vision and stereo vision for ball recognition and motion estimation in three-dimensional (3D) space. When the ball is located on the ground field, a novel algorithm based on RANSAC and Kalman filter is proposed to estimate the ball velocity using the omnidirectional vision. When the ball is kicked up into the air, an iterative and coarse-to-fine method is proposed to fit the moving trace of the ball with paraboic curve and predict the touchdown-point in 3D space using the stereo vision. Experimental results show that the robot can effectively estimate ball motion in 3D space using the hybrid vision system and the proposed algorithms, furthermore, the advantages of the 360

$$^\circ $$

field of view of the omnidirectional vision and the high object localization accuracy of the stereo vision in 3D space can be combined.

This paper describes three human-inspired approaches to balancing in highly dynamic environments. In this particular work, we focus on balancing on a bongo board - a common device used for human balance and coordination training - as an example of a highly dynamic environment. The three approaches were developed to overcome limitations in robot hardware. Starting with an approach based around a simple PD controller for the centre of gravity, we then move to a hybrid control mechanism that uses a predictive control scheme to overcome limitation in sensor sensitivity, noise, latency, and jitter. Our third control approach attempts to maintain a dynamically stable limit cycle rather than a static equilibrium point, in order to overcome limitations in the speed of the actuators. The humanoid robot Jimmy is now able to balance for several seconds and can compensate for external disturbances (e.g., the bongo board hitting the table). A video of the robot Jimmy balancing on the bongo board can be found at

http://youtu.be/ia2ZYqqF-lw

.

In robot soccer, self-localization of robots may fail because of perception failure, falling down or by being pushed by another robot. In this study, our goal is to improve self localization of robots using the teammate robots’ perceptions. Robots which have perceived more landmarks and have moved less can share their localization and observations with other robots to improve localization accuracy. Currently, in the RoboCup Standard Platform League it is not feasible to identify the jersey number of robots using vision. Therefore, we merge perceptions of all robots depending on their reliability, and then identify them in a probabilistic manner to increase localization performance and provide a common world model that can be used for planning. We show that our approach has a significant advantage with respect to single robot localization on estimating the poses of the robots.

The game of soccer is one of the main focuses of the RoboCup competitions, being a fun and entertaining research environment for the development of autonomous multi-agent cooperative systems. For an autonomous robot to be able to play soccer, first it has to perceive the surrounding world and extract only the relevant information in the game context. Therefore, the vision system of a robotic soccer player is probably the most important sensorial element, on which the acting of the robot is fully based. In this paper we present a new modular time-constrained vision library, named UAVision, that allows the use of video sensors up to a frame rate of 50 fps in full resolution and provides accurate results in terms of detection of the objects of interest for a robot playing soccer.

The paper presents a multi-robot cooperative framework to estimate the 3D position of dynamic targets, based on bearing-only vision measurements. The uncertainty of the observation provided by each robot equipped with a bearing-only vision system is effectively addressed for cooperative triangulation purposes by weighing the contribution of each monocular bearing ray in a probabilistic manner. The envisioned framework is evaluated in an outdoor scenario with a team of heterogeneous robots composed of an Unmanned Ground and Aerial Vehicle.

This paper addresses an estimation problem of the ball’s state of spin in RoboCup Small Size League (SSL). A spinning ball varies its speed after the ball bounces off the floor. This paper proposes an image-based estimation method of the ball’s state of spin, in particular, by using inertia feature of co-occurrence matrix of the image sequence. The effectiveness of our proposed method is shown by some experiments.

Robot localization and mapping algorithms commonly represent the world as a static map. In reality, human environments consist of many movable objects like doors, chairs and tables. Recognizing that such environment often have a large number of instances of a small number of types of objects, we propose an alternative approach,

Model-Instance Object Mapping

that reasons about the models of objects distinctly from their different instances. Observations classified as short-term features by Episodic non-Markov Localization are clustered to detect object instances. For each object instance, an occupancy grid is constructed, and compared to every other object instance to build a directed similarity graph. Common object models are discovered as strongly connected components of the graph, and their models as well as distribution of instances saved as the final Model-Instance Object Map. By keeping track of the poses of observed instances of object models, Model-Instance Object Maps learn the most probable locations for commonly observed object models. We present results of Model-Instance Object Mapping over the course of a month in our indoor office environment, and highlight the common object models thus learnt in an unsupervised manner.

Detection of objects in the air is a difficult problem to tackle given the dynamics and speed of a flying object. The problem is even more difficult when considering a non-controlled environment where the predominance of a given color is not guaranteed, and/or when the vision system is located on a moving platform. As an example, most of the Middle Size League teams in RoboCup competition detect the objects in the environment using an omni directional camera that only detects the ball when in the ground, and losing any precise information of the ball position when in the air. In this paper we present a first approach towards the detection of a ball flying using a Kinect camera as sensor. The approach only uses 3D data and does not consider, at this time, any additional intensity information. The objective at this stage is to evaluate how useful is the use of 3D information in the Middle Size League context. A simple algorithm to detect a flying ball and evaluate its trajectory was implemented and preliminary results are presented.

In the context of the humanoid robotics soccer, ball dribbling is a complex and challenging behavior that requires a proper interaction of the robot with the ball and the floor. We propose a methodology for modeling this behavior by splitting it in two sub problems: alignment and ball pushing. Alignment is achieved using a fuzzy controller in conjunction with an automatic foot selector. Ball-pushing is achieved using a reinforcement-learning based controller, which learns how to keep the robot near the ball, while controlling its speed when approaching and pushing the ball. Four different models for the reinforcement learning of the ball-pushing behavior are proposed and compared. The entire dribbling engine is tested using a 3D simulator and real NAO robots. Performance indices for evaluating the dribbling speed and ball-control are defined and measured. The obtained results validate the usefulness of the proposed methodology, showing asymptotic convergence in around fifty training episodes, and similar performance between simulated and real robots.

In this paper we propose a modification to the well known Delaunay Triangulation based positioning in the attacking situation positioning of the agents in 2D Soccer Simulation environment. Due to advanced defensive skills such as marking skill, the attacker agents should have a dynamic positioning with respect to the rival team defenders. The proposed method employs the vertices of the Voronoi Diagram of the defending team agents as potential positions for the attacker team agents, since these positions are dynamic and change with the movement of the defending team agents and are always safe positions regarding the distance to the defending team agents, and also have a good coverage of the field. So the attacking agents can increase the chance of receiving pass by the ball owner agent and the scoring chance by taking positions on these vertices. This proposed method then applies Maximum Weighted Bipartite Matching to match these vertices to the attacking agents. This algorithm can be applied by each agent individually, but in order to reduce the possible decision conflicts in this matching which is the result of the limitation in the incoming information of the field from the agents’ sensors, this algorithm can be performed by one agent and then this agent should inform the other attacking team agents of the result by communication skills like “say ability” in 2D Soccer Simulation (SS) environment. This method shows better performance in offensive situation than the conventional Delaunay Triangulation based positioning. It is tested in 2D SS environment as a highly dynamic multi-agent environment but its application is not restricted to the 2D SS League.

Even with improvements in machine learning enabling robots to quickly optimize and perfect their skills, developing a seed skill from which to begin an optimization remains a necessary challenge for large action spaces. This paper proposes a method for creating and using such a seed by (i) observing the effects of the actions of another robot, (ii) further optimizing the skill starting from this seed, and (iii) embedding the optimized skill in a full behavior. Called KSOBI, this method is fully implemented and tested in the complex RoboCup 3D simulation domain. To the best of our knowledge, the resulting skill kicks the ball farther in this simulator than has been previously documented.

In biped locomotion, the energy minimization problem is a challenging topic. This problem cannot be solved analytically since modeling the whole robot dynamics is intractable. Using the inverted pendulum model, researchers have defined the Zero Moment Point (ZMP) target trajectory and derived the corresponding Center of Mass (CoM) motion trajectory, which enables a robot to walk stably. A changing vertical CoM position has proved to be crucial factor in reducing mechanical energy costs and generating an energy efficient walk [

1

]. The use of Covariance Matrix Adaptation Evolution Strategy (CMA-ES) on a Fourier basis representation, which models the vertical CoM trajectory, is investigated in this paper to achieve energy efficient walk with specific step length and period. The results show that different step lengths and step periods lead to different learned energy efficient vertical CoM trajectories. For the first time, a generalization approach is used to generalize the learned results, by using a programmable Central Pattern Generator (CPG) on the learned results. Online modulation of the trajectory is performed while the robot changes its walking speed using the CPG dynamics. This approach is implemented and evaluated on the simulated and real NAO robot.

Soccer simulation league is one of the founding leagues of RoboCup. In this paper we discuss the past, present and planned future achievements and changes. Also we summarize the connections and inter-league achievements of this league and provide an overview of the community contributions that made this league successful.

The Small Size Robot League (SSL) was among the founding RoboCup leagues in the 1997 competition held during IJCAI’97 in Nagoya, Japan. Since then, the league has experienced various advances in terms of robot design, number of robots, field size, software algorithms and other infrastructure used during the games, among these the recent standardization of the vision system shared by all teams. The SSL league has been one of the fastest paced leagues in RoboCup where teamwork, coordination, high-level strategies and artificial intelligence have played a critical role in the league development. As robots speeds have greatly increased in the past years, the league has witnessed the development of advanced control and cooperative algorithms. In parallel, shared open software, in particular the shared vision system has made it easier for new teams to join the league. In this paper we discuss the past, present and future of the Small Size League in its path towards the goal of achieving robot vs. human soccer in 2050.

The RoboCup Middle-Size League (MSL) is one of the founding leagues of the annual RoboCup competition. Ever since its birth it has been a league where development of hard- and software happens simultaneously in a real-world decentralized multi-robot soccer setting. Over the years the MSL achieved scientific results in robust design of mechatronic systems, sensor-fusion, tracking, world modelling and distributed multi-agent coordination. Because of recent rule changes which actively stimulate passing, matches in RoboCup MSL have become increasingly appealing to a general audience. Approximately five thousand spectators were present during last years final match. In this paper we present our plan to build on this momentum to further boost scientific progress and to attract new teams to the league. We also give a historical overview and discuss the current state of the MSL competition in terms of strengths, weaknesses, opportunities and threats.

The Standard Platform League is unique among RoboCup soccer leagues for its focus on software. Since all teams compete using the same hardware (a standard robotic platform), success is predicated on software quality, and the shared hardware makes quality judgments simpler and more objective. Growing out a league based on the Sony AIBO quadruped robots, the league has constantly evolved while moving ever closer to playing by human rules, and currently features the Aldebaran NAO humanoid robots. The hallmark of the league has been a focus on individual agents’ skills, such as perception, localization, and motion, at the expense of more team-oriented skills, such as positioning and passing. The league has begun to address this deficiency with the creation of the Drop-in Challenge, where robots from multiple teams will work together. This new focus should force teams to work on multi-agent coordination in more abstract and general terms and promises to create fruitful new lines of research.

This paper describes the major achievements in the history of the RoboCup Humanoid League from its start in 2002 to 2014. We provide a perspective for the future of the league with a strong push towards larger robots and FIFA-like playing fields. We also discuss some risks associated with these intended changes.

The RoboCup rescue simulation competitions have been held since 2001. The experience gained during these competitions has supported the development of multi-agent and robotics based solution for disaster mitigation. The league consists of three distinct competitions. These competitions are the agent competition, the virtual robots competition, and the infrastructure competition. The main goal of the infrastructure competition is to increase every year the challenge and to drive the innovation of the league, while the agent and virtual robot competition are focused on developing intelligent agents and robot control systems that can cope with those challenges. This paper provides an overview on the current state-of-the-art in the league and developments and innovations planned for the future.

The RoboCup Rescue Robot League (RRL) aims to foster the development of rescue robots that can be used after disasters such as earthquakes. These robots help to discover victims in the collapsed structure without endanger the rescue personnel. The RRL has been held since 2000. The experience gained during these competitions has increased the level of maturity of the field, which allowed to deploy robots after real disasters, e.g. at the Fukushima Daiichi nuclear disaster. This article provides an overview on the competition and its history. It also highlights the current state of the art, the current challenges and the way ahead.

RoboCup@Home is an application-oriented league within the annual RoboCup events. It focuses on domestic service robots and mobile manipulators interacting with human users. Participating robots need to solve tasks ranging from following and guiding human users to delivering objects, e.g., in a supermarket.

In this paper, we present the @Home league and how it evolved over the last seven years since its existence. We place particular emphasis on how we evaluate the teams’ performances over the years and how we use the obtained statistics to drive the development of the league. This process is shown in detail on two examples—following human guides, and finding and manipulating objects. Finally, we will outline possible future directions and developments.

In this article we present a compact and modular device designed to allow a conventional stiff servo actuator to be easily upgraded into a series elastic actuator (SEA). This is a low cost, open source and open hardware solution including mechanical CAD drawings, circuit schematics, board designs and firmware code. We present a complete overview of the project as well as a case study where we show the device being employed as an upgrade to add compliance to the knee joints of an existing humanoid robot design.

The Setplay Framework (available from SourceForge as free software) is composed of a C++ library (Project name:

fcportugalsetplays

), a fully functional RoboCup Simulation 2D demonstration team (

fcportugalsetplaysagent2d

), and a complete graphical tool (

SPlanner

), that can be used to design and plan the collaborative behavior between the soccer player agents. In order to demonstrate the usage of the Setplay library, a complete 2D simulation team, based on Agent2D, was developed. This example team uses the framework to execute previously planned collaborative behavior. This framework can be used both within simulated environments, such as the Robocup Soccer Simulation leagues, and with real soccer playing robots. This paper presents the free software Setplay Framework, and provides the necessary information for any team to use the framework with the goal of providing collaborative behavior to a team of soccer playing robots.