Petri-net-based implementations for FIRA weightlifting and sprint games with a humanoid robot

https://doi.org/10.1016/j.robot.2013.08.012Get rights and content

Abstract

In this paper, the Petri net-based wireless sensor node architecture (PN-WSNA) is used to control a humanoid robot to play weightlifting and sprint games in the FIRA HuroCup league. With the PN-WSNA approach, the control scenario and decision-making for playing weightlifting and sprint games can be modeled as a PN-WSNA model. The PN-WSNA inference engine is further used to interpret and execute the PN-WSNA model according to the sensor information from visual perception. Therefore, the implementation of playing weightlifting and sprint games is achieved in terms of the PN-WSNA model instead of native code. To verify the PN-WSNA-based implementation approach, an autonomous humanoid robot equipped with a camera and a single-board computer is used for experiments, where the camera is responsible for grabbing image frames; the single-board computer is responsible for visual localization; and the PN-WSNA models the execution and locomotion command generation. Finally, several PN-WSNA models for playing weightlifting and sprint games are proposed and the experimental results are demonstrated and discussed to validate the feasibility of applying the proposed PN-WSNA-based implementation approach.

Introduction

Participating in autonomous robot competitions is an important way to acquire education in robotics. Participants may learn not only hands-on skills and technologies but also novel intelligent control approaches for robotics. The Federation of International Robot-soccer Association (FIRA) HuroCup League  [1] is an important organization that defines a number of challenging sport-related skills for autonomous humanoid robots. Current FIRA HuroCup matches are sprinting, penalty kicking, obstacle racing, lifting and carrying, weightlifting, marathon, wall-climbing and basketball. With the above-mentioned matches, humanoid robots have to use cameras to detect image features, such as lines, objects, markers, obstacles, etc., and then make decisions. As a consequence, the autonomous humanoid must be capable of image processing and recognition, localizing image patterns  [2], decision making for navigation  [3] and locomotion  [4] to perform the tasks specified in the FIRA HuroCup league.

Research into humanoid robots for performing autonomous tasks in competitions has been getting more popular in recent years. Vadakkepat et al.  [5] presented humanoid robot research into processing architecture, gait generation and vision systems in their laboratory. Those humanoid robots have been successfully participating in various robotic soccer competitions. Haddadin et al.  [6] proposed the idea of a kicking motion with elasticity for humanoid robots to meet the safety and performance concerns in human–robot soccer games.

Zagal et al.  [7] presented the techniques of self-modeling approaches for humanoid soccer robots. Calderon et al.  [8] identified the generation of human-like soccer primitives from human data for operating humanoid robots. Cherubini et al.  [9] proposed policy gradient learning for a humanoid soccer robot. Finally, Friedmann et al.  [10] presented adequate motion simulation and collision detection for playing soccer games with humanoid robots.

On the other hand, the implementation of humanoid robot control system is important in executing tasks. Programming with a high-level language, such as the C language, is a conventional way to realize autonomous navigation and task execution for robots. Although programming with a native high-level language is convenient for autonomous robot developers, such kinds of high-level programming code are hard to maintain. As a consequence, several popular commercial solutions for robot programming have been proposed, such as LabVIEW from National Instruments  [11] and Simulink from MathWorks  [12]. With these solutions, the developers do not need to deal with complicated programming efforts from coding huge high-level programs. Instead, they can realize their system in terms of model-based implementation approaches that are capable of proposing fast and reliable robotic solutions.

The Petri net-based wireless sensor node architecture (PN-WSNA) [13] is a model-based implementation approach that can be used to develop autonomous sensing and decision functions for intelligent control systems. The PN-WSNA is a high level Petri net (PN), and PNs  [14] are usually used to model discrete event dynamic systems (DEDSs) with concurrent and asynchronous characteristics. Practically, the behaviors of autonomous robots can be characterized as DEDSs, where the decision is made according to a change of sensor status.

Therefore, a number of Petri-net-based approaches have been proposed to model and realize autonomous robot control systems. For example, Costelha et al.  [15] used Petri nets as solutions for the modeling, analysis and execution of robotic tasks. Zhang et al.  [16] presented an agent oriented hierarchical Petri net to deal with environment perception, information fusion, path-planning and autonomous driving for autonomous robots. You et al.  [17] presented a household mobile robot, and a Petri net was used to develop event-driven dual-driving-wheel synchronization control models. Huang  [18] proposed an action selection strategy for soccer robots based on a Petri net.

Usually, a conventional PN cannot deal with sensor and actuation interfaces. As a consequence, they are hardly used for task execution. However, the PN-WSNA provides additional interfaces and functions for sensor data collection and inference, intra-communication and device actuation. Hence, the PN-WSNA can be applied directly to deal with real control problems.

To verify the proposed PN-WSNA-based modeling approach, PN-WSNA models for executing the tasks in the HuroCup league for weightlifting  [19] and sprint games are discussed in this paper with a small size humanoid robot, named HuroEvolution-JR. The rest of this paper is organized as follows: Section  2 introduces HuroEvolution-JR and PN-WSNA; Section  3 describes the PN-WSNA models and experiments for the weightlifting games; Section  4 elaborates the PN-WSNA models and experiments for the sprint games; finally, Section  5 summarizes the conclusions and future work.

Section snippets

Mechanical design of HuroEvolution-JR

The HuroEvoluation-JR is a humanoid robot, especially developed for the FIRA kid-size competition. The HuroEvoluation-JR is configured with 22 degrees of freedom (DOF), where 12 DOFs are arranged in two legs; 8 DOFs are arranged in two arms; and 2 DOFs realize the pan-and-tilt motions on the neck joint for driving the head camera. Commercial RC servos  [20] are used for driving the joint motions of the HuroEvoluation-JR. The height and weight of HuroEvoluation-JR are 45 cm and 2.8 kg,

PN-WSNA weightlifting games

The PN-WSNA-based weightlifting game is an extended work of  [19], and this paper decreases the model complexity in terms of reducing the number of transmitter and receiver places. In addition, the two-dimensional (2-D) concurrent model presentation in  [19] was changed to a cascade model presentation. Therefore, the models proposed in this paper are much easier to understand. Moreover, a newly designed humanoid robot was used for the experiments.

As described in  [19], the PN-WSNA model in the

PN-WSNA sprint games

The robot sprint match is an event for humanoid robots. The goal of this match is for the robots to move as quickly as possible from the start line to the end line. The sprint game can be divided into two stages, including forward (stage I) and backward (stage II) motions. In the first stage, the robot is set at start line and had to move forward to approach an end line in its own lane. Hence, a beacon marker (with a filled red circle) is used to guide the robot to walk within its lane width.

Conclusions and future work

This paper presents a novel model-based implementation approach to navigate an autonomous humanoid robot in executing the weightlifting and sprint games in the FIRA HuroCup league. The visual perception module provides the sensor data for the corresponding sensor places. The PN-WSNA models can be created with the PN-WSNA IDE, and these models are further inferred and executed to control the HuroEvolution-JR to finish the games. The PN-WSNA models were discussed based on different sensor

Chung-Hsien Kuo received a Ph.D. degree in mechanical engineering from National Taiwan University, Taipei, Taiwan, in 1999.

Since 2012, he has been a Professor with the Department of Electrical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan. His current interests include medical robotics, smart sensor systems, and rehabilitation engineering.

References (23)

  • J. Rodriguez-Resendiz et al.

    Adjustable speed drive project for teaching a servo systems course laboratory

    IEEE Transactions on Education

    (2011)
  • Cited by (3)

    Chung-Hsien Kuo received a Ph.D. degree in mechanical engineering from National Taiwan University, Taipei, Taiwan, in 1999.

    Since 2012, he has been a Professor with the Department of Electrical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan. His current interests include medical robotics, smart sensor systems, and rehabilitation engineering.

    Yu-Cheng Kuo received a master degree in electrical engineering from National Taiwan University of Science and Technology, Taipei, Taiwan, in 2012. He is currently working toward a Ph.D. degree at the Department of Electrical Engineering, National Taiwan University of Science and Technology.

    His current interests include sensor fusion, system engineering, and intelligent robots.

    Ting-Shuo Chen received a Ph.D. degree in electrical engineering from National Taiwan University of Science and Technology, Taipei, Taiwan, in 2013.

    His current interests include wireless sensor networks, smart sensor systems, and rehabilitation engineering.

    Yu-Ping Shen received a B.S. degree in electrical engineering from National Taiwan University of Science and Technology, Taipei, Taiwan, in 2013. He is currently working toward a masters degree at the Department of Electrical Engineering, National Taiwan University of Science and Technology.

    His current interests include autonomous robots, humanoid robots, and machine vision.

    Chia-Che Cheng received a B.S. degree in electrical engineering from National Taiwan University of Science and Technology, Taipei, Taiwan, in 2013.

    His interests include autonomous robots, soccer robots, humanoid robots, and machine vision.

    View full text