The Effects of Ship’s Roll Motion on the Center of Mass and Margin of Stability During Walking: A Simulation Study

Walking strategies in an unstable environment like a ship differ from walking on stable ground. Extreme ship motions may endanger the safety of the crews. Notably, a loss of balance on board can lead to an injury or an accident of falling off a ship. Keeping one’s balance on board a ship is strongly influenced by the ship’s motion. Therefore, the objective of this study is to determine how walking on a ship differs from walking in a stable environment and explore the effects of the ship’s roll motion on balance control and stability while walking in sea environments. We hypothesized that step time variability, center of mass (COM), and margin of stability (MOS) would significantly differ between stable and unstable walking conditions. We also hypothesized that there would be an effect of rolling cycles and angles on increasing step time variability, COM excursion, and MOS variability. We recruited 30 healthy individuals between 21 and 39 years old for this study. Participants walked for two minutes at their self-selected speeds during the study with and without rolling on a computer-assisted rehabilitation environment (CAREN) system. The CAREN system was used to simulate the parametric roll motion of ships up to 20 degrees. This study quantified step time variability, peak COM excursion, and MOS variability in different rolling conditions. We found a significant difference in step time variability (p < 0.001), lateral peak COM excursion (p < 0.001), and MOS variability (p < 0.001) between waking on land and walking at sea.

considerably different from walking on land. A ship's motion 23 plays an important role in affecting walking ability, thereby 24 The associate editor coordinating the review of this manuscript and approving it for publication was Jingang Jiang . directly limiting the human gait [1], [2]. Several studies have 25 examined walking on a ship [3], [4], [5], [6], but there is still 26 a significant lack of research analyzing gait characteristics 27 in unstable moving environments. Thus, we are investigating 28 how walking on a ship differs from walking on land. 29 Extreme fluctuations of the ship at sea may threaten not 30 only the ship itself but also the safety of the crew. Notably, 31 a loss of balance on board can lead to a severe injury as the 32 ship is made of steel. Balancing the body at sea is strongly 33 influenced by the motion of ships, such as rolling and pitch-34 ing. To reduce the rate of man overboard accidents, safety 35    To our best knowledge, this paper is the first attempt to and we can propose a specific working safety regulation 90 for seafarers as well as this can be used to prevent crew's 91 injuries or accidents on the ship at sea by assessing gait 92 instability.

95
Thirty healthy young adults (20 males and 10 females) par-96 ticipated in this study. The characteristics of the participants 97 are shown in Table 1. Participants were excluded if they had 98 1) major lower extremity injury or surgery; 2) known cardio-99 vascular conditions that make it unsafe for them to exercise; 100 3) a history of dizziness due to vestibular disorders such as 101 Meniere's disease and vertigo; 4) any difficulty in walking 102 in unstable moving environments. All subjects signed an 103 informed consent form before data collection. This study was 104 approved by the Institutional Review Board at the University 105 of Nebraska Medical Center (IRB 141-21-EP).

107
A 3-dimensional motion capture system (Vicon Motion Sys-108 tem Ltd., Oxford, UK) with ten cameras was used to record 109 marker trajectories at 100 Hz. A total of 37 reflective markers 110 were placed on anatomical landmarks according to the Plug-111 in Gait full-body model [16], including 4 markers on the head, 112 5 on the torso, 12 on the upper limb, 4 on the pelvis, and 12 on 113 the lower limb. 114 We placed 7 wireless inertial measurement unit (IMU) sen-115 sors (Xsens, Enschede, Netherlands) to obtain 3-axial accel-116 erations from the pelvis and each foot/shank/thigh segment. 117 Fig. 1 shows the placement of reflective makers and IMU 118 sensors attached to each subject's body. The CAREN system 119 (Motek, Amsterdam, Netherlands) was also used to simulate 120 the roll motion of a ship for up to 20 degrees of rolling.

122
All participants walked on a split-belt treadmill for 2 minutes 123 at a self-selected comfortable pace. Each participant com-124 pleted nine 2-minute walking trials in the CAREN system for 125 each of the following conditions: no rolling (NR), 5-, 10-, 15-, 126 and 20-degrees of rolling with slow (12s) and fast (6s) rolling 127 cycles (i.e., each rolling condition was abbreviated as SR5, 128 SR10, SR15, SR20, FR5, FR10, FR15, and FR20).     [29], [30], but we focused on 172 the changes in the COM excursion with a ship's roll motion 173 simulations. where BOS is the lateral boundary of the base of support (the 181 lateral malleolus marker on each ankle at heel strike), and 182 XcoM is calculated as: where COM is the lateral position of COM at heel strike, 185 vCOM is the COM velocity that is computed as the derivative 186 of the COM position at heel strike, and w 0 is defined as: where g is the gravitational constant (9.81 m/s 2 ) and l is 189 the pendulum length, which is defined as the mean distance 190 between the ankle marker and the COM in this study.     Table 2. There were statistically 224 significant differences between rolling conditions (Table 3,  Table S1, p < 0.05).

247
Step time variabilities in SR15, FR15, SR20, and FR20 248 conditions (Fig. 3a, supplemental  Table S4, p = 0.682) and no significant interaction effect 253 between the rolling cycle and the rolling angle (Table 5, (Fig. 4a). We also found that there were significant 267 differences in step time variability in most of the rolling 268 angles except for between 5 and 10 degrees and between 269 10 and 15 degrees (Fig. 4a).   Table S6). Based on the results of the effect of the rolling 281 angle only, we also observed that there were significant dif-282 ferences in peak COM excursion at different rolling angles 283 (Fig. 4b). A significant interaction effect between the rolling 284 cycle and the rolling angle was found in the lateral peak COM 285 excursion (Table 5).

287
The MOS variabilities in a lateral direction for all rolling 288 conditions (Fig. 3c, supplemental Table S8, p < 0.001) were 289 significantly increased compared to the NR condition. There 290 was a significant main effect of the rolling cycle (Table 5, p < 291 0.001) in the MOS variability. Post-hoc analysis showed that 292 the MOS variability during the fast-rolling cycle was higher 293 than during the slow-rolling cycle in most rolling angles other 294 than in 5 degrees of rolling motion (supplemental Table S10). 295 A significant main effect of the rolling angle on the MOS 296 variability was found with all rolling conditions (Table 5, 297  supplemental Table S9, p < 0.001). There was a significant 298 interaction effect between the rolling cycle and the rolling 299 angle in the MOS variability. Additionally, we noted that the 300 MOS variability at the different rolling angles, irrespective 301 of rolling cycles, was significantly different from the NR 302 condition (Fig. 4c).

304
This study aimed to investigate how walking in an unstable the simulated sea conditions were significantly greater than 319 with no rolling motions. We partially confirmed our second 320 hypothesis since rolling angles affected increasing step time 321 variability, peak COM excursion, and MOS variability while 322 rolling cycles influenced only the MOS variability. Moreover, 323 based on the MANOVA results, we also found that the rolling 324 cycle, rolling angle, and their interaction were significant 325 for all three dependent variables, which shows that there is 326 interdependency among the dependent variables: step time 327 variability, peak COM excursion, and MOS variability. 328 We found that the ship's rolling motion increased the step 329 time variability (Fig. 3a), and this is thought to have rapidly 330 changed the walking steps to balance in an unstable envi-331 ronment. We also found that the step time variability was 332 significantly increased at a rolling angle of 15 degrees or 333 higher compared to no rolling condition (Fig. 3a, Fig. 4a). 334 In addition, many participants responded that the balance 335 difficulty increased rapidly at 20 degrees in their self-reported 336 VOLUME 10, 2022 questionnaires ( Table 2). Based on these two results, this 337 study could propose a crew's work safety rule that limits or 338 requires attention to deck work in at least 15 degrees or higher 339 rolling environments.

340
Peak COM excursion was increased substantially during 341 walking in rolling conditions (Fig. 3b, Fig. 4b). We found 342 that the higher degree of rolling increased the peak COM 343 excursion in the lateral direction, which means the COM 344 moved more laterally to balance in higher rolling motions. 345 Furthermore, there was a significant increase in MOS vari-346 ability under rolling conditions (Fig. 3c, Fig. 4c ingly. Additionally, a previous study found that the lateral 387 MOS is affected by age and BMI [32]. Men and women may 388 differ in their levels of dynamic stability during walking [33], 389 which could be an interesting topic to investigate since the 390 difference in balance control ability between men and women 391 could affect walking differently in moving environments.

392
Thus, these human factors, such as age, sex, height, and BMI, 393 should be taken into account in future studies. Lastly, the 394 rolling angle was limited to 20 degrees in the experimental 395 setting. In fact, rolling of more than 20 degrees occurs in bad 396 weather at sea, which significantly hinders the crew's safety 397 by making it difficult to control the balance. In this study, 398 it was inevitable to set the rolling angle up to 20 degrees 399 due to technical problems with the CAREN system, which 400 supports up to 20 degrees. However, this study found walking 401 characteristics that could sufficiently endanger safety even 402 at 15 or 20 degrees. Additional studies would need to vali-403 date our experimental results in the actual ship environment 404 at sea.

406
In conclusion, this study examined the effects of a ship's 407 rolling motion on the changes in human walking characteris-408 tics such as step time variability, COM excursion, and MOS 409 variability in the sea environment. Study results indicate that 410 different rolling angles have an impact on increasing step 411 time variability, peak COM excursion, and MOS variability, 412 but the rolling cycles influence MOS variability only. Peak 413 COM excursion and MOS variability can effectively assess 414 dynamic stability during walking on a ship at sea. Thus, this 415 study could propose a crew's work safety rule that limits or 416 requires attention to deck work on a ship and help prevent 417 injuries on the ship at sea by assessing gait instability. Further 418 studies are needed to confirm our results in a real ship at sea 419 and to investigate the possibility of the use of our measures 420 to prevent falling overboard. He has inves-577 tigated a number of fundamental theoretical and 578 applied problems preventing complete utilization 579 of high-speed wireless data network technology. 580 His current research interests include wireless sen-581 sor networks, development of mobile applications, design and analysis of 582 low-power communication protocols, mobility monitoring using wireless 583 sensors, and gait analysis using machine learning approaches. 584 585 VOLUME 10, 2022