Available online at www.sciencedirect.com ScienceDirect Procedia Engineering 112 (2015 ) 517 521 7th Asia-Pacific Congress on Sports Technology, APCST 2015 Improvement of crawl stroke for the swimming humanoid robot to establish an experimental platform for swimming research Motomu Nakashima* and Yuto Tsunoda Graduate school of information science and engineering, Tokyo Institute of Technology, 2-12-1 Ookayama, Meguro, Tokyo 152-8552, Japan Abstract The objective of this study was to improve the stability of crawl swimming for the swimming humanoid robot. The swimming humanoid robot had 24 servo motors for the whole body and could perform basic human swimming motion. This robot can swim freely by the crawl stroke. However, the robot sometimes changed the propulsive direction while free swimming. The change in the propulsive direction was caused by instability of body rolling. Therefore, swimming motion which could change propulsive direction and roll angle amplitude was examined by simulation. Effectiveness of the devised motion was confirmed by a freeswimming experiment. 2015 The The Authors. Published Published by Elsevier by Elsevier Ltd. This is Ltd. an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the School of Aerospace, Mechanical and Manufacturing Engineering, RMIT University. Peer-review under responsibility of the the School of Aerospace, Mechanical and Manufacturing Engineering, RMIT University Keywords: Swimming; Crawl Stroke; Stability; Humanoid Robot 1. Introduction In spite of such a long history of competitive swimming, its mechanics still have not been fully clarified yet since it is an extremely complicated phenomenon in which a complex human body moves unsteadily with many degreesof-freedom in the three-dimensional water flow. For example, the hand path in the crawl stroke depicts a distorted ellipse when viewed from the side in the absolute space [1], showing that the hand does not push the water straight at a constant depth. Furthermore, the kinematics of the arm and hand during the underwater stroke is highly unsteady [2]. From this viewpoint, many attempts were made recently to quantify the unsteady fluid forces acting on * Corresponding author. Tel.: +81-3-5734-2586; fax: +81-3-5734-2586. E-mail address: motomu@mei.titech.ac.jp 1877-7058 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the the School of Aerospace, Mechanical and Manufacturing Engineering, RMIT University doi:10.1016/j.proeng.2015.07.235
518 Motomu Nakashima and Yuto Tsunoda / Procedia Engineering 112 ( 2015 ) 517 521 a swimmer while swimming. The first approach was an experiment involving a human subject [3 5]. However, this method had problems with insufficient repeatability, physical fatigue of the subject, and difficulty in installing sensors on the subject. For this reason, some researchers have conducted experiments using physical models such as a robot instead of human subjects. A lot of measuring experiments using physical models have been conducted to date [6 9], but there was no full-body experimental platform which can consider interactions between the many segments involved in normal swimming motions. Therefore, analysis using physical models had been performed on an isolated segment and misleading conclusions could have been developed. To solve such problems, a full-body swimming humanoid robot for research about human swimming was developed by Chung and Nakashima [10]. The robot was named SWUMANOID. SWUMANOID had a detailed human body shape using 3D scanning and printing equipment since it was developed for the experimental model substituting a human subject. Not only the appearance but also the methodology to realize various swimming strokes was considered. In order to reproduce the complicated swimming motions with high fidelity, 20 waterproof actuators were installed. The free swimming of the crawl stroke has been already realized in the previous study [11]. However, there were two problems with the realized free swimming of the crawl stroke. One was the instability of the rolling motion. The roll angle amplitude of the trunk became unstable. When the rolling motion became small, the recovery stroke motion of the robot sometimes failed, resulting in the significant decrease in the swimming velocity. Another problem was the direction control. Due to the instability of the rolling motion as well as the manufacturing error, the robot did not go straight perfectly. To solve these problems, a method to vary the roll angle amplitude as well as the propulsive direction was examined by simulation in the present study. The validity of the proposed method was examined by a free swimming experiment. 2. Overview of SWUMANOID SWUMANOID is a humanoid robot whose size is half of an actual competitive swimmer. It has 24 servo motors (DYNAMIXEL series, Robotis inc) inside the body. The location of joints and motors are shown in Fig. 1(a). Scapular joints are installed in order to conduct the recovery stroke smoothly. O-rings and X-rings (Quad rings) are used for waterproofing the joint parts. The specifications of SWUMANOID are shown in Table 1. The flow of swimming motion generation is shown in Fig. 1(b). First, the swimming motion of an actual swimmer was analyzed and input into the swimming human simulation model SWUM as step motions. SWUM is a simulation model, in which the human body is divided into 21 segments, and rigid body dynamics as well as unsteady fluid forces are taken into account [12]. By inputting the joint motion, it can calculate the absolute movement of the body as well as the joint torque. The simulation model for SWUMANOID was created on SWUM. The resultant motion and the joint torque were confirmed by the SWUM simulation. Then the joint angles of the (a) (b) Fig. 1. Overview of SWUMANOID. (a) location of joints and motors; (b) flow of swimming motion generation.
Motomu Nakashima and Yuto Tsunoda / Procedia Engineering 112 ( 2015 ) 517 521 519 Table 1 Specifications of SWUMANOID Item Specifications Dimensions Height: 925 mm Width:270 mm Depth:100 mm Weight inc. battery 7.4 kg Degree of freedom Total; 24 DOF Arm: 6 DOF 2 Arms Waist; 2 DOF Leg: 5 DOF 2 Legs Battery Li-Po 14.8 V 1550 mha 2 robot were computed by the kinematic analysis. Finally, the resultant joint motion of the robot was checked by using a 3D CAD software. 3. Examination of method to vary roll angle amplitude and propulsive direction by simulation 3.1. Roll angle amplitude The method to vary the roll angle amplitude was examined by the SWUM simulation. The motion to increase the roll angle amplitude was devised as shown in Fig. 2(a). In the devised motion, the horizontal flexion angle of the shoulder only in the beginning of the underwater stroke was increased. The simulation result of the roll angle is shown in Fig. 2(b). In this simulation, the normal crawl stroke was conducted until the fifth cycle. Then the motion was changed into the devised one. The changed angle in the horizontal flexion in this case was 60 degrees. As a result, it was found that roll angle amplitude just after the change increased to 89.3 degrees (44.3 degrees in the normal stroke), and that it still increased to 54.4 degrees after it stabilized. Therefore, it can be confirmed that the devised motion had an effect to increase the roll angle amplitude. a b Fig. 2. Devised motion to increase the roll angle amplitude. (a) devised motion in which the horizontal flexion angle of the shoulder only in the beginning of the underwater stroke was increased for 60 degrees; (b) simulation result in which the normal crawl stroke was conducted until the fifth cycle, and then the motion was changed into the devised one. 3.2. Propulsive direction The devised motion was also used to vary the propulsive direction. The simulation result of x-y (plane of the water surface) locus of the robot is shown in Fig. 3. In this simulation, the robot performed the normal crawl stroke for the first five cycles. Then the stroke motion of the left arm was changed into the devised one from the sixth cycle. The changed angle was 40 degrees in this case. From the figure, it was found that the propulsive direction was changed to the right after the angle change. Therefore it was confirmed that the devised motion had an effect to change the propulsive direction.
520 Motomu Nakashima and Yuto Tsunoda / Procedia Engineering 112 ( 2015 ) 517 521 Fig. 3. Simulation result of x-y locus of the robot. The robot performed the normal crawl stroke for the first five cycles. Then the stroke motion of the left arm was changed into the devised one from the sixth cycle. The changed angle was 40 degrees in this case. 4. Free swimming experiment In order to examine the effect of the method to vary the roll angle amplitude and propulsive direction shown in the previous section, a free swimming experiment using SWUMANOID was conducted. The experimental setup is shown in Fig. 4. The experiment was conducted in a 25m swimming pool. The motion of the robot was filmed by a video camera on the ground. Four modes for the stroke motion were prepared as below: (1) normal stroke, (2) devised motion only for the right arm, (3) devised motion only for the left arm and (4) devised motion for both arms. The changed angle in the devised motion was 25 degrees. These four modes were switched by an operator via wireless module. The experimental result of x-y locus is shown in Fig. 4. The blue line with dots represents the locus of the robot. Each blue dot represents the position of the robot at the beginning of each cycle. The green and purple circle areas represent the periods in which the devised motions only for the right and left arms were performed, respectively. The red dots represent the cycle in which the devised motion for both arms was performed. The normal stroke motion was performed in other areas. From this figure, it was confirmed that the robot could change the propulsive direction to the right (purple circles) and left (green circles) during or after the changed motions. Therefore, the validity of the method to change the propulsive direction was confirmed. In addition, it was found that the blue dots were dense just before (in the figure, just the right of) the red dots. This meant that the swimming velocity decreased since the recovery strokes failed due to the decrease in the roll angle amplitudes. However, it was Fig. 4. Experimental result of x-y locus of the robot. The blue line with dots represents the locus of the robot. Each blue dot represents the position of the robot at the beginning of each cycle. The green and purple circle areas represent the periods in which the devised motions only for the right and left arms were performed, respectively. The red dots represent the cycle in which the devised motion for both arms was performed. The normal stroke motion was performed in other areas.
Motomu Nakashima and Yuto Tsunoda / Procedia Engineering 112 ( 2015 ) 517 521 521 also found that the blue dots became sparse after (in the figure, the left of) the red dots. This meant that the roll angle amplitudes were increased by the devised motion, and the recovery strokes succeeded. Therefore, it was confirmed that the devised motion was effective to increase the roll angle amplitude. 5. Conclusions In order to improve the crawl stroke for the swimming humanoid robot, a motion to increase the roll angle amplitude was devised. This devised motion also had an effect to change the propulsive direction. The validity of these two effects was confirmed numerically and experimentally. As the future task, the changed angle of the devised motion should be varied continuously. The automated control of the roll angle also should be realized by means of sensor installation. Acknowledgements This work was supported by the JSPS Grants-in-Aid for Scientific Research (No. 26282174). References [1] E.W. Maglischo, Swimming Fastest, Human Kinetics, 2003, p.97. [2] H. M. Toussaint, C. Van Den Berg, and W. J. Beek, Pumped-up propulsion during front crawl swimming, Medicine and science in sports and exercise, 34 (2002) 314-319. [3] A.P Hollander, G. De Groot, G.J van Ingen Schenau, H.M Toussaint, H. De Best, W. Peeters, A. Meulemans and A.W Schreurs, Measurement of active drag during crawl arm stroke swimming, Journal of Sports Sciences 4 (1986) 21-30. [4] S.V Kolmogorov, O.A Duplishcheva, Active drag, useful mechanical power output and hydrodynamic force coefficient in different swimming strokes at maximal velocity, Journal of Biomechanics 25 (1992) 311-318. [5] H. Takagi and R. Sanders, Measurement of propulsion by the hand during competitive swimming, The Engineering of Sport 4 (Eds. S. Ujihashi and S.J. Haake): Blackwell Publishing, 2002, pp.631-637. [6] M.A Lauder, P. Dabnichki, Estimating propulsive forces-sink or swim?, Journal of Biomechanics 38 (2005) 1984-1990. [7] N.O Sidelnik and B.W Young, Optimising the freestyle swimming stroke: the effect of finger spread, Sports Engineering 9 (2006) 129-135. [8] M. Nakashima and A. Takahashi A, Clarification of unsteady fluid forces acting on limbs in swimming using an underwater robot arm (development of an underwater robot arm and measurement of fluid forces), Journal of Fluid Science and Technology 7(2012) 100-113. [9] M. Nakashima and A. Takahashi, Clarification of unsteady fluid forces acting on limbs in swimming using an underwater robot arm (2nd report, modeling of fluid force using experimental results). Journal of Fluid Science and Technology 7(2012) 114-128. [10] C. Chung, M. Nakashima, Development of a swimming humanoid robot for research of human swimming, Journal of Aero Aqua Bio mechanisms, 3 (2013) 109-117. [11] C. Chung, M. Nakashima, Free swimming of the swimming humanoid robot for the crawl stroke, Journal of Aero Aqua Bio-mechanisms, 3 (2013) 118-126. [12] M. Nakashima, K. Satou, Y. Miura, Development of swimming human simulation model considering rigid body dynamics and unsteady fluid force for whole body, Journal of Fluid Science and Technology, 2 (2007) 56-67.