Available online at www.sciencedirect.com Procedia Engineering 34 (2012 ) 646 651 9 th Conference of the International Sports Engineering Association (ISEA) Improved rehabilitation and training techniques through the use of motion simulation Core strength conditioning for elite rowers Caleb Sawade a, Stephen Turnock a, Alexander Forrester b, Martin Toward c a Fluid Structure Interactions Group, University of Southampton b Aeronautics, Astronautics and Computational Engineering Unit, University of Southampton c Institue of Sound and Vibration, University of Southampton University of Southampton, University Road, Southampton, SO17 1BJ, UK Accepted 05 March 2012 Abstract The most common injuries experienced by sweep rowers relate to the lower back; rowers with greater core strength have been shown to be less prone to these injuries. Stationary rowing trainers (Ergometers) currently used for training and rehabilitation generally have a fixed support base, contrary to a real boat. It is hypothesized that training with an unstable base may help develop core strength in athletes. This paper describes early development of a rowing simulator with a realistic unstable platform. Using SolidWorks, a conceptual motion simulator was designed, which was then linked to MATLAB/Simulink to produce a controllable simulation of rowing. A control algorithm was developed to compute the forces and virtual rowing parameters required. The algorithm uses a PID controller to maintain a coordinate position of the virtual boat s centre of gravity. In particular, the issue of creating an unstable feeling about the roll axis was developed. Comparison between the simulated output and real rowing data is required for verification of effective control. The resultant control system could be linked to a real-time motion platform, whereby EMG and or other biometric measurements could be used to determine the effectiveness of core muscle activation between the simulator, Ergometer and on water rowing. 2012 Published by Elsevier Ltd. Open access under CC BY-NC-ND license. Keywords: Rowing; simulation; rehabilitation; MATLAB/Simulink; motion control 1. Introduction As elite athletes undergo physiotherapy to overcome injury, the joint or muscle which has been damaged is typically focused on, in order to strengthen and increase flexibility in the area which has 1877-7058 2012 Published by Elsevier Ltd. doi:10.1016/j.proeng.2012.04.110 Open access under CC BY-NC-ND license.
Caleb Sawade et al. / Procedia Engineering 34 ( 2012 ) 640 651 647 experienced trauma. Modern day rehabilitation techniques rarely include the process of incorporating the athletes natural sporting environment. Naturally this environment affects the performance and technique used by the athlete to achieve his or her optimal physical ability. Therefore, by neglecting these environment parameters, rehabilitation may only improve the particular joint or muscle injured, while there may be various other factors that caused the injury in the first instance. Such factors often relate to the athlete trying to overcompensate a particular movement due to the changing environment conditions they are in. With overcompensation, more force or greater flexibility is required for the athlete to achieve normal sporting performance. By simulating real environment conditions, and incorporating the resultant external forces within the rehabilitation process, the athlete can not only train and strengthen the damaged joint or muscle, but train other joint and muscle groups to support the injured body part during the sporting activity. Elite rowers predominantly train on Ergo s [1]. As upper body twisting encountered in sweep style rowing is not reproduced, Ergo s can induce unwanted forces on the rower s lower back. Additionally, as traditional Ergo s (some include slides for movement in the direction of travel) are fixed in the rotational and vertical axes, no core muscle strengthening or training is provided for boat stability. This lack of core training can cause overextension of particular core muscle groups while on water to stabilize the boat [2] and deliver maximum power during the stroke. This overextension can result in muscle damage and injury [3]. It is hypothesized that by introducing a motion platform for the rower to train on, core strength and flexibility will be increased during training, which could result in improved stability and a stronger kinetic chain to deliver effective force production while on water. This article will discuss and present a computer simulation of a motion platform for rehabilitation purposes for rowers, through the use of physical simulation. The platform shall mimic the motion of an elite rowing boat for condition training and rehabilitation of injured rowers. MATLAB/Simulink was used in conjunction with other software programs to create a computer simulation and a control law model for implementation onto the physical motion platform. At this stage of research, only the simulation has been carried out, with the physical rowing simulator to be built at a later stage. 2. Fundamentals of Motion To reproduce the motions of a boat for the purpose of rowing training, the relevant degrees of freedom (DOF) must be specified. That is, a set of independent translations and/or rotations that specify the displaced or transformed position and orientation of the boat. A fully defined system in all possible directions and rotations has six DOF [4]. Developing a motion platform with a full six-dof may not be necessary for all environments e.g. an axis might be excluded if the motion in that axis is small or where it does not significantly contribute to the observed effect (i.e. core stability in this study). Furthermore, by minimising the number of DOF, the mechanical and control systems can become simplified. In this study, simulating core muscle activity in sweep rowers, heave, roll and pitch motions were considered of most importance due to their relatively large magnitude and strong effect on core muscle activity [2]. Reproduction of motions in the yaw and sway directions were considered of lesser importance due to their comparatively small magnitude, while motion in the surge direction though large in a boat was thought to have only a small effect on core muscle activity. 3. Conceptual Rowing Simulator Control Design For the purpose of computer simulation and control investigation, a conceptually designed rowing simulator for injury recovery will be used as the virtual motion platform. The rowing simulator design
648 Caleb Sawade et al. / Procedia Engineering 34 ( 2012 ) 646 651 incorporates actuators attached from the base to the rowing simulator chassis. The actuators movements intend to give the desired three DOF. At this stage of research, only simulation is produced, with the intent of building the full physical simulator for hypothesis verification at a later stage. In order to control the motion platform in real time, computer simulation will be used to calculate the relevant forces of each actuator, and to produce the desired motion of the platform. MATLAB/Simulink was used to create this computer simulation. The simulation comprises of three main components; the motion platform input or desired position matrix, the controller and the plant. The plant, refers to the motion platform as hardware, or a virtual representation of the intended hardware. To create that representation Solidworks and Simulink s SimMechanics was used to build the 3D design mentioned in section 4, into virtual objects. These objects have defined masses and inertias. 3.1 Motion platform input For accurate modelling of a physical system, real world parameters (such as weight, dimensions, etc.) and physical constraints must be accounted for. The motion platform input must determine the coordinate position and rotation matrix for the virtual rowing hull with respect to the absolute (or reference) virtual co-ordinate system. This is derived from physical virtual constraints, and the inputs of both the actual and virtual rowers. Once determined, the position and rotation matrices are used to calculate the actuator positions, and hence the motion platforms co-ordinate translation and rotation with respect to the reference co-ordinate system in the real world. 3.1.1 Virtual Hull Model The desired position of the simulator is calculated to mimic the intended virtual rowing hull. Focus will be placed on roll motion, as both heave and pitch will be calculated using generic fluid structure interaction methods. This is because only the roll effects weather the hull is stable, neutral or unstable, as outlined by Rawson, et al. [7]. Basic ship theory is applied and an overview for the context of a rowing hull is explained (for further details of ship theory the reader should consult [7]). It is assumed for this article that the cross-sectional geometry of the hull is semi-circular. This implies that the righting lever (known as GZ) of the hull is proportional to the roll angle, until GZ reaches a maximum, before the hull tips over. This simplifies the model, however, rowing teams could input the GZ curve of their particular boat into the simulator for increased realism. This would also enable testing of various hull geometries for different crew skill level. Figure 2 shows the three equilibrium states of the rowing hull. The left diagram illustrates the neutral equilibrium state, whereby the centre of gravity (CG) coincides with the metacentre (M) of the hull. If this occurs, no righting moment is generated. The centre diagram shows that the CG is above M and hence a negative GZ is generated. Finally, in the third diagram, the CG is below M, creating a positive righting lever, making the boat stable. Fig. 1. Three states of hull equilibrium [8]
Caleb Sawade et al. / Procedia Engineering 34 ( 2012 ) 640 651 649 In the case of rowing, M is approximately level with the waterline for near semi-circular boats. For the rowers to achieve an efficient stroke, they sit above this point [8]. Therefore the CG is above the M and hence the hull is considered to be in an unstable equilibrium. The negative moment generated is described by the following equation;, (1) where GM is the distance from the CG to the metacentre of the hull and is the roll angle. As CG is above the metacentre, GM is negative [8]. 3.1.2. Virtual Rower Model As the rower-hull system is not a rigid body, the CG is free to move about the M and centre of buoyancy. The rower is able to shift their weight to oppose the negative hull righting moment, making the system stable. This is done by using both core strength and leg strength to apply a force which changes the centre of mass of the system. In this article, both of these forces are represented as a single core strength force to simplify the rower. Two virtual muscles were created which provide an actuating force to the rower with respect to the seat. The actuating force is controlled by a proportional-integralderivative (PID) controller which aims to mimic the response of the rower. As the boat roll s, the virtual rower applies force to each side of the body s virtual core to shift the CG of the rower. The change in force applied to the seat by the rower results in a positive righting moment on the hull. This moment is greater than the hull s negative moment, due to the larger mass of the rower. 3.1.3. Virtual Oar Model As can be seen in Figure 1, additional forces on the hull are due to the moments applied to the oarlocks. These moments are a result of the rower moving the oar up and down. Additionally, a hydrodynamic force is applied to the oar blade when in contact with the water during the drive phase of the stroke. Each rower must adjust their oars to minimise the difference in moments between the left and right side of the hull. For initial simulation, the vertical force on the oarlock will be a function of handle force and oar motion during the stroke [5]. 3.2. The controller The controller uses a PID algorithm. The simplest way to control the desired trajectory is to apply forces to the plant proportional to the position error [6]. Using a PID feedback is a common solution for linear control. The PID controller determines the error between the desired position of the actuators and the actual position of the actuators. The control law for each actuator has the form: (2) 4. Simulation Results Three tests were conducted to determine the effectiveness of the simulation. The first was used to observe the virtual hulls unstable equilibrium. The virtual rower was fixed to the simulator seat, meaning no core movement was permitted. Figure 2 shows how by applying a small disturbance force of 5 Nm, the virtual hull will roll over within 3 seconds. This will vary from rower to rower, as GM and the mass of the rower will vary. The second test shows how the rower uses core muscle force to stabilize the hull. A 50 Nm moment was applied to the virtual hull for 2 seconds. As the rower feels a change in force
650 Caleb Sawade et al. / Procedia Engineering 34 ( 2012 ) 646 651 and hull roll angle, it activates the core to shift the CG to create a positive righting moment. This can be seen in Figure 3. Fig. 2. L-R Test 1 - Virtual rowerr sitting on unstable rowing motion platform at time = 0, 1 and 2 seconds respectively. Graph of the negative righting moment generated due to the position of the CG above the M of thee virtual hull, and virtual hull roll angle Fig. 3. L-R Test 2 - Virtual rowerr using core muscle strength to stabilize the motion platform at time = 0, 0.25, 1, 2, 2.5, 3 and 4 seconds respectively. Graph of disturbance force acting on the virtual hull and the hull roll angle from timee = 0 to 4 seconds Finally, additional forces were introduced into the system. Delta Fz is the resultant force due to uneven rowing from the left and right oars as described previously. Because this simulation is concentrated on sweep rowing, another virtual rower must be present within the system for moment balance. Sine waves were used to mimic a basic wave pattern causing the virtual hull to move. All three DOF were given a command signal. Figure 4 shows how the uneven disturbance force causes the rowerr to actively stabilize the rowing hull with their core. Fig. 4. Comparison between Delta Fz (Oar Force total) and one side of the rowers core muscle actuation. Itt is clear that the rower is moving their core to stabilize the force applied by delta Fz.
Caleb Sawade et al. / Procedia Engineering 34 ( 2012 ) 640 651 651 5. Conclusion Through the use of Solidworks, MATLAB/Simulink and SimScape, computer simulation of a three DOF motion platform was created. This platform was designed to mimic the motions of the relevant DOF of an elite rowing boat. In particular, the feeling of an unstable rowing hull has been presented. Simulation results indicate that a rower must use their core to stabilise the rowing hull. If no core stabilisation is applied, small disturbance forces will cause the hull to roll over. These results suggest that by using a motion platform for rowing rehabilitation and training, core muscle activity will be higher than compared to other on-land fixed frame-rowing devices. This early stage of research could lead to an increase of core strength which may help reduce the likelihood of further injury. It may also provide better core training for improved power delivery during the stroke, as well as control of trunk rotation when reaching for the catch. To improve control accuracy, development of the desired position model will be taken further. In particular, the virtual oar model should be tested with real rowing data with both horizontal and vertical oar angle parameters. Additionally, a more thorough investigation into the oar-hydrodynamic forces and the corresponding moments applied to the hull will be undertaken. Results verification is required by comparing on-water hull roll, with simulated hull roll, given a fully defined rower input. A direct comparison between core muscle activity observed during on water, simulator and ergometer rowing will also take place. The results of such an investigation will show if training and rehabilitating rowers on such a motion platform is in fact more beneficial, compared to training on a fixed ergometer. Acknowledgements This work was conducted under the auspices of the Faculty of Engineering and the Environment at the University of Southampton and with support from the Engineering and Physical Sciences Research Council (UK), UK Sport and McLaren Applied Technologies. References [1] J. Mäestu, et al., "Monitoring of Performance and Training in Rowing," Sports Medicine, vol. 35, pp. 597-617, 2005. [2] W. B. Kibler, et al., "The Role of Core Stability in Athletic Function," Sports Medicine, vol. 36, pp. 189-198, 2006. [3] J. S. Rumball, et al., "Rowing Injuries," Sports Medicine, vol. 35, pp. 537-555, 2005. [4] H. Lonescu, "6 Degrees of Freedom," vol. 1032x683, ed, 2010. [5] L. Formaggia, et al., "A model for the dynamics of rowing boats," International Journal for Numerical Methods in Fluids, vol. 61, pp. 119-143, 2009. [6] T.MathWorks. (2010, 1/10/2010). Modeling the Stewart Platform. Available: http://www.mathworks.com/help/toolbox/physmod/mech/ug/f15-36487.html#f15-37547. [7] K.J Rawson, et al., "Basic Ship Theory Vol 1," New York, 1968. Longman Group Limited. [8] C. Pulman, "The Physics of Rowing," University of Cambridge.