• Journal of the Korea Society for Simulation
• /
• v.20 no.1
• /
• pp.51-59
• /
• 2011

This paper is concerned with locomotion of dog-like quadruped robots that can adapt to various terrains, mainly dealing with implementation methods and characteristics of static and dynamic gaits. To this end, a 12-DOF robot is built in house, motional trajectories of its body and feet are generated mimicking biological life, and the corresponding leg joint angles are analytically obtained by inverse kinematics. Such joint angle data are then applied to the robot's ADAMS model for computer simulations so that the planned walk and trot gaits are both confirmed dynamically stable. However, contrary to the simulation results, previous trot patterns showed unstable behavior during experiments. This problem led us to analyze the reason, and in the course we discovered the importance of maximally utilizing the concept of WSM rather than ZMP and therefore reducing the gait period to secure the stability of dynamic gaits such as trot.

• Journal of the Korea Society for Simulation
• /
• v.25 no.1
• /
• pp.53-61
• /
• 2016

This paper introduces 'EgresSIM', which is microscopic evacuation simulation software. EgresSIM developed in this paper is a three-dimensional (3D) pedestrian evacuation simulator based on the improved model advanced from the floor field model(FFM), a microscopic pedestrian model. This software can simulate large size buildings that consist of a number of floors, stairs, rooms, and exit doors. Moreover, this software can arrange several hundreds or thousands of pedestrians in indoor space and check their movements through the 3D viewer in real time, as well as produce detailed results about evacuation situations such as which paths are employed by individual pedestrians, how long does it takes to evacuate, and how many evacuees are gathered at each of the exit doors. Building data needed in the simulation are constructed as XML files according to pre-defined indoor data models and information of simulation results is also created as XML log files. A moving pattern of pedestrians can be represented in many ways by adjusting the sensitivity parameters of two walk models supported by EgresSIM. Thus, evacuation simulation can be done based on many assumptions of situations such as movement to the nearest exit door or blackout after outage.

• Journal of Biomedical Engineering Research
• /
• v.24 no.2
• /
• pp.133-140
• /
• 2003

In this study, muscle activity was measured using surface EMG (sEMG) during a voluntary maneuver (ankle dorsiflexion) in the supine position was compared pre and post gait training. Nine patients with incomplete spinal cord injury participated in a supported treadmill ambulation training (STAT), twenty minutes a day, five days a week for three months. Two tests, a gait speed test and a voluntary maneuver test, were made the same day, or at least the same week, pre and post gait training. Ten healthy subjects' data recorded using the same voluntary maneuvers were used for the reference. sEMG measured from ten lower limb muscles was used to observe the two features of amplitude and motor control distribution pattern, named response vector. The result showed that the average gait speed of patients increased significantly (p〈0.1) from 0.47$\pm$0.35 m/s to 0.68$\pm$0.52 m/s. In sEMG analysis, six out of nine patients showed a tendency to increase the right tibialis anterior activity during right ankle dorsiflexion from 109.7$\pm$148.5 $mutextrm{V}$ to 145.9$\pm$180.7 $mutextrm{V}$ but it was not significant (p〈0.055). In addition, only two patients showed increase of correlation coefficient and total muscle activity in the left fide during left dorsiflexion. Patients' muscle activity changes after gait training varied individually and generally depended on their muscle control abilities of the pre-STAT status. Response vector being introduced for quantitative analysis showed good Possibility to anticipate. evaluate, and/or guide patients with SCI, before and after gait training.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.3
• /
• pp.35-42
• /
• 2021

In this paper, we propose a deep learning-based person re-identification method using a three-dimensional RGB-Depth Xtion2 camera considering joint coordinates and dynamic features(velocity, acceleration). The main idea of the proposed identification methodology is to easily extract gait data such as joint coordinates, dynamic features with an RGB-D camera and automatically identify gait patterns through a self-designed one-dimensional convolutional neural network classifier(1D-ConvNet). The accuracy was measured based on the F1 Score, and the influence was measured by comparing the accuracy with the classifier model (JC) that did not consider dynamic characteristics. As a result, our proposed classifier model in the case of considering the dynamic characteristics(JCSpeed) showed about 8% higher F1-Score than JC.