Brain-computer interface (BCI) is a technology used in various fields to analyze electroencephalography (EEG) signals to recognize an individual"s intention or state and control a computer or machine. However, most of the research on BCI is on motor imagery, and research on active movement is concentrated on upper limb movement. In the case of lower limb movement, most of the research is on the static state or single movements. Therefore, in this research, we developed a deep-learning model for classifying walking behavior(1: walking, 2: upstairs, 3: downstairs) based on EEG signals in a dynamic environment to verify the possibility of classifying EEG signals in a dynamic state. We developed a model that combined a convolutional neural network (CNN) and a bidirectional long short-term memory (BiLSTM). The model obtained an average recognition performance of 82.01%, with an average accuracy of 93.77% for walking, 76.52% for upstairs, and 75.75% for downstairs. It is anticipated that various robotic devices aimed at assisting people with disabilities and the elderly could be designed in the future with multiple features, such as human-robot interaction, object manipulation, and path-planning utilizing BCI for control.

Estimating energy expenditure is essential in monitoring the intensity of physical activity and health status. Energy expenditure can be estimated based on wearable sensors such as inertial measurement unit (IMU). While a variety of methods have been developed to estimate energy expenditure during day-to-day activities, their performances have not been thoroughly evaluated under walking conditions according to various speeds and inclines. This study investigated IMU-based neural network models for energy expenditure estimation under various walking conditions and comparatively analyzed their performances in terms of sensor attachment locations and training/testing datasets. In this study, two neural network models were selected based on a previous study (Slade et al., 2019): (M1) a multilayer perceptron using sensor signals during each gait cycle, and (M2) a recurrent neural network using sensor signal sequences of a fixed window size. The results revealed the following: (i) the performance of the foot attachment model was the best among the five sensor attachment locations (0.89 W/kg for M1 and 1.14 W/kg for M2); and (ii) although the performance of M1 was superior to that of M2, M1 requires accurate gait detection for data segmentation by each stride, which hinders the usefulness of M2.

In highly mobile workplaces, wearable walking assistant robots can reduce muscle fatigue in the lower extremities of workers and increase energy efficiency. In this study, walking efficiency according to the development of an ultralight wearable hip-assist robot for industrial workers was verified. Five healthy adult males participated in this study. Their muscle fatigue and energy consumption were compared with and without the robot while walking on a flat treadmill and stairs. When walking on the treadmill while wearing the robot, muscle fatigue in the rectus femoris and gastrocnemius decreased by 90.2% and 37.7%, respectively. Oxygen uptake and energy expenditure per minute also decreased by 8.9% and 13.1%, respectively. When climbing stairs while wearing the robot, fatigue of the tibialis anterior, semitendinosus, and gastrocnemius muscles decreased by 18.2%, 33.3%, and 63.6%, respectively. Oxygen uptake and energy expenditure per minute also decreased by 3.6% and 3.7%, respectively. Although wearing a hip-assist robot could reduce muscle fatigue and use metabolic energy more efficiently, it is necessary to further increase the energy efficiency while climbing stairs. This study is intended to provide basic data to improve the performance of robots.

Climbing stairs places a greater load on lower limb joints compared to walking on level ground. Variations in anatomical structures and muscle characteristics between genders suggest potential differences in the distribution of required mechanical work among the three lower limb joints. This study aimed to identify gender disparities in the allocation of mechanical work to lower limb joints during stair climbing. A total of thirty-six adults (equally divided between men and women) participated in the study. Participants ascended stairs equipped with force plates at their comfortable speeds, while motion was captured using nine cameras. Inverse dynamics analysis was employed to calculate the mechanical work performed by each joint during four phases of stance: weight acceptance, pull-up, forward continuation, and push-up. Male participants exhibited significantly higher mechanical work than females at the hip and ankle joints (p < 0.05) from the 1st- 3rd phases and the 2nd phase, respectively. Conversely, female subjects displayed greater knee joint work during the 2nd- 3rd phases (p < 0.05). Notably, a pronounced gender difference was observed during the 2nd pull-up phase, where body mass is lifted by a single leg. These findings suggest that men and women employ distinct strategies in distributing mechanical work across lower limb joints.

Hybrid mobile robot is the system that will practically combine legged walking and skated driving in the same system. Therefore, this robot has own problems of inverse kinematics that are not considered in typical walking robots. In this paper, I fully categorized the inverse kinematics problems for hybrid mobile robot with general motion by walking and driving on an inclined plane, including switching end-effectors between foots and blades. I also solved the inverse kinematics for each case of problems. I here actively adopted the coordinate transformation derived from the inclined plane to cope with the random motion of foots and blades on the plane. I then presented several examples of the inverse kinematics problems with specific situations, and verified the validity of the analysis method from the results.

In this study, the Inertial Measurement Unit (IMU) signals and clinical evaluation scales for Parkinson"s disease were correlated. The study included 16 patients diagnosed with Parkinson"s disease. Each subject was evaluated based on Korean Mini-Mental State Examination (KMMSE), Unified Parkinson"s Disease Rating Scale (UPDRS) part 3, New Freezing of Gait Questionnaire (NFOGQ) parts 2 & 3, and Hoehn & Yahr Scale (H&Y). All subjects performed the Time Up and Go test by attaching IMU sensors to both ankles and torso. Based on the tilting angle of torso and the time of first step, the freezing and non-freezing windows were determined. Seven IMU features involving the ankle signals were calculated in the specific window. Spearman’s correlation analysis of clinical evaluation scales was performed. As a result, the freezing index and power of locomotion band (0.3-3 Hz) were recommended to determine UPDRS part 3. Also, the intensity of the locomotion band facilitated evaluation of NFOGQ part 3 regardless of freezing of gait.

There are no known studies on the changes in walking ability in patients with transfemoral amputations returning to daily activities after prosthetic gait training. The ability to walk after discharge may vary depending on an individual’s physical, psychological, and social factors. This study compared spatiotemporal variables and lower limb coordination ability at the end of training and one year after the end of training in seven unilateral transfemoral amputees and analyzed the factors affecting walking ability. The study results confirmed that there was no significant difference in spatiotemporal parameters such as walking speed and lower limb coordination ability after one year of training, and walking ability was well maintained after training. Five out of seven (71.4%) participants in this study returned to work, and there was a strong correlation between employment and gait improvement (r = 0.806, p < .05). In conclusion, activities such as social participation, employment, and exercise were very important factors in maintaining and improving an individual’s walking ability. The findings are intended to be used as basic data to provide guidelines for maintaining the health of lower limb amputees.

The understanding of impaired neural control of gait after stroke is important to evaluate mobility impairments focused on improving walking function. Previous studies have shown that the central nervous system may control gait via muscle synergies, which modularly organizes multiple muscles. However, there are insufficient studies to evaluate mobility impairments, using muscle synergy during walking in post-stroke patients. Thus, the purpose of this study was to determine if the variability of muscle synergies during gait reflects impaired motor performance. Electromyography (EMG) signals were collected from five persons with post-stroke hemiparesis and five similarly age healthy persons, as they walked on a treadmill at a comfortable speed. EMG signals were decomposed using non-negative matrix factorization and the variability of muscle synergies was calculated using a synergy stability index (SSI). We also investigated correlation between the SSI and Fugl-Meyer assessment and Berg Balance Scale, which are clinical evaluation indicators. Post-stroke patients were found to have variable muscle synergies. We also observed a positive proportional relation, between SSI and clinical motor impair evaluation indicators. These results could yield a quantitative assessment of gait after stroke, and provide a causal relationship between internal neuromuscular mechanisms and functional performance.

In this study, an insole-type ground reaction force (GRF) measurement system using a load cell was manufactured and configured as a system that can measure joint angle and GRF, when walking in conjunction with a commercialized inertial sensor. The data acquisition device was used to acquire synchronized data, between the inertial measurement unit (IMU) sensor and the load cell insole. A three-dimensional motion analysis system comprising six infrared cameras and two ground reaction forces, was used to check the accuracy of the gait measurement system, comprising an inertial sensor and a load cell insole. The motion and force data were acquired while performing five times six-meter walking test by five young adult male subjects (Age: 26.0±1.8, Height: 171.4±6.8 cm, Weight: 62.2±10.8 kg). It was measured and as a result of comparing the calculated sagittal joint angle with the vertical GRF, the sagittal lower extremity joint angle correlation coefficient (Pearson’s r) was 0.40 to 0.94, and the vertical GRF to be 0.98 to 0.99. It is necessary to upgrade the joint angle calculation algorithm through future research. Additionally, the possibility of clinical application for actual stroke patients will be reviewed.

Herein, we describe the development of a wearable lower limb rehabilitation robot that can perform walking movement according to the walking pattern trajectory. The robot can adjust the left and right widths of the waist and the front and rear widths of 100 and 20 mm, and the length of the thigh link and calf link by 100 and 80 mm, respectively, so that stroke patients of different heights and weights can use it in hospitals. For manufacturing the lower limb rehabilitation robot, the right exoskeleton was safely designed through structural analysis, and the motor and reducer constituting the hip joint actuator were calculated. The fabricated lower limb rehabilitation robot was divided into its own characteristic experiment and wearing characteristic experiment. Its own characteristic experiment was an experiment by the robot itself, and the wearing characteristic experiment was an experiment conducted after a person wears the robot. Through these two experiments, angular deviation of the walking pattern was analyzed. Results of the analysis confirmed that the wearable walking characteristic test was performed within 3.1° based on the self walking characteristic test result. Therefore, the fabricated lower limb rehabilitation robot can be used for gait training in stroke patients.

This paper introduces PongBot, a quadruped robot developed for preparation in the Dronebot Challenge held in Jangseonggun, Jeollanam-Do, South Korea in November 2020. The Dronebot Challenge, hosted by the Army Headquarters, is a competition to demonstrate that drones and robots can be useful for military purposes. In 2020, this competition consisted of a total of 8 events and we participated in the ‘Traveling on rough terrain’ event, which consisted of various terrains, such as, slopes, unpaved roads, and streams. PongBot is a quadruped robot that uses an electric motor and can walk for more than an hour on various terrains. Also, according to the rules of the competition, the robot had a system which could be remotely controlled from a ground control station. In addition, by applying the SLAM algorithm, the robot operator received information about its surrounding environment, thereby deriving records to facilitate the operation. The performance of this robot system and SLAM algorithm was verified through this competition.

The current method of gait analysis has several limitations for determining gait stability, such as a complicated preparation process, repeated experimental procedures that are time-consuming, and financial burden of experiments. This study investigated whether gait stability could be analyzed using only the COM-COP (Center of Mass-Center of Pressure) inclination angle connecting COM and COP. COM and COP coordinates were obtained from a motion analysis system for a total of 40 elderly and young subjects. The COM-COP inclination angle that changed in real time during level walking was then analyzed to obtain gait stability on each of sagittal and frontal planes using these coordinates. As a result, the gait symmetry index on the sagittal plane did not show a statistically significant difference between young and elderly subjects (First Step, p = 0.189; Second Step, p = 0.711). On the frontal plane, elderly subjects showed 0.39 degrees (p = 0.058) and 0.5 degree (p = 0.03) larger side-to-side sway angles in the first and second steps than young subjects, respectively. Gait stability can be analyzed using a more simplified experimental method with minimum amount of data in future gait analysis.

In this paper, an integrated ankle torque sensor and mechanism (Foot Link) of a Tendon driven-type wearing walking aid robot were designed. The foot link consists of an ankle torque sensor and a mechanism connected to the footrest. The size of the sensing part of the ankle torque sensor was designed through structural analysis and assembled by attaching a strain gauge. As a result, the reproducibility error and the nonlinearity error were within 0.04%, respectively. And the calibration result of the ankle torque sensor, reproducibility error, and non-linearity error were identified to be within 1%, respectively. Therefore, it is proposed that the ankle torque sensor presented in this paper can be used to measure the torque acting on the tendon-driven walking aid robot.

The purpose of this study was to suggest the method for automated locomotion modes (Level Walking, Stair Ascent, Stair Descent) detection based on the Radial Basis Function Support Vector Machine (RBF-SVM) for the hip gait assist robot. The universal hip gait assist robot had a limit in detection of the walking intention of users because of the limited sensors’ quantity. Through the offline training, using MATLAB, we trained the collected gait data of users wearing the hip gait assist robot and obtained the parameter of the RBF-SVM model. In the online test, using LabVIEW, we developed the algorithm for the locomotion modes decision of individuals using the optimized parameter of the RBF-SVM. Finally, we executed the gait test for three terrains through the walking environment’s test platform. As a result, the locomotion modes decision rate for three terrains was 98.5%, 99%, and 98% respectively. And the decision delay time of algorithm was 0.03 s, 0.03 s, and 0.06 s respectively.

In this paper, the design and fabrication of the calf-link with knee joint torque sensor of a tandem-driven walking-assist robot is described. Tendon-driven walking-assist robots should be designed and constructed with a wire wheel and a torque sensor, as one body to reduce the weight of the calf link. The torque sensor consists of four plate sensing parts crossed 90° around the wire wheel. Structural analysis was performed to determine the size of the torque sensor sensing part, and a torque sensor was built by attaching a strain gauge to the sensing part. As a result of the characteristics test, the reproducibility error and the nonlinearity error of the manufactured torque sensor were less than 0.03% and 0.04%, respectively. As a result of the calibration, the reproducibility error and the nonlinearity error were less than 0.08%, respectively. Thus, it is considered that the knee joint torque sensor of the calf link can be attached to the tandem-driven walking-assist robot.