• 한국정밀공학회지
• /
• 제28권12호
• /
• pp.1425-1433
• /
• 2011

We investigated an early rehabilitation training system that increase the intensity of patient rehabilitation training to shorten the time it takes for patients to progress to a secondary rehabilitation training stage by allowing patients incapable of self-ambulation. It consisted of tilting bed, unstable platform using strong springs and training program for lower limb rehabilitation. We performed experimental study on the muscular activities of tibialis anterior(TA), soleus(SO), gastrocnemius(GA) in the lower extremities during training of straight line, circle, quardrangle pattern during tilting angle of $30^{\circ}$, $60^{\circ}$. The muscle activities were higher during tilting angle of $30^{\circ}$ than $60^{\circ}$. In straight line pattern, the muscle activities were higher by SO, GA and TA during medio-lateral direction, however, by TA, SO and GA during anterio-posterior direction. In circle and quardrangle pattern, the muscle activities were higher by TA, SO and GA during clockwise and counterclockwise direction. The results indicate that the early rehabilitation training system could be applied to improve the lower extremity muscular strength for elderly and patients, especially, stroke.

• 한국전문물리치료학회지
• /
• 제25권2호
• /
• pp.53-61
• /
• 2018

Background: Stroke patients show abnormal walking patterns due to brain injury. In order to have the desired walking pattern, appropriate stimulation is required to activate the central pattern generator. For this reason, our study performed treadmill ambulatory training with rhythmic auditory stimulation. However we did not consider the influence of visual feedback. Objects: The purpose of this study was to compare the gait abilities in chronic stroke patients following either treadmill walking training with rhythmic auditory stimulation and visual feedback (TRASVF) or treadmill walking training with rhythmic auditory stimulation (TRAS) alone. Methods: Twenty-one stroke patients were divided into two groups: A TRASVF group (10 subjects) and a TRAS group (11 subjects). They received 30 minutes of neuro-developmental therapy (NDT) and walking training for 30 minutes, five times a week for three weeks. Temporal and spatial gait parameters were measured before and after the training period. The Biodex gait trainer treadmill system measured gait parameters. Results: After the training periods, the TRASVF group showed a significant improvement in walking speed, the step length of the affected limb, and time on each foot of the affected limb when compared to the TRAS group (p<.05). Conclusion: The results of this study showed that the treadmill walking training with rhythmic auditory stimulation and visual feedback improved individual gait ability more than the treadmill walking training with rhythmic auditory stimulation alone. Therefore, visual feedback should be considered along with rhythmic auditory stimulation training.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제34권1호
• /
• pp.11-17
• /
• 2007

패턴 분류에 많이 사용되는 기법 중의 하나인 메모리 기반 추론 알고리즘은 단순히 메모리에 저장된 학습패턴 또는 초월평면과 테스트 패턴간의 거리를 계산하여 가장 가까운 학습패턴의 클래스로 분류하기 때문에 테스트 패턴을 분류하는 기준을 설명할 수 없다는 문제점을 가지고 있다. 이 문제를 해결하기 위하여, 메모리 기반 학습 기법인 RPA를 기반으로 학습패턴들에 내재된 규칙성을 표현하는 IF-THEN 형태의 규칙을 생성하는 점진적 학습 알고리즘을 제안하였다. 하지만, RPA에 의해 생성된 규칙은 주어진 학습패턴 집합에만 충실히 학습되어 overfitting 현상을 보이게 되며, 또한 패턴 공간의 과도한 분할로 인하여 필요 이상으로 많은 개수의 규칙이 생성된다. 따라서, 본 논문에서는 생성된 규칙으로부터 불필요한 조건을 제거함으로써 ovefitting 현상을 해결함과 동시에 생성되는 규칙의 개수를 줄일 수 있는 점진적 규칙 추출 알고리즘을 제안하였으며, UCI Machine Learning Repository의 벤치마크 데이터를 이용하여 제안한 알고리즘의 성능을 입증하였다.

• 한국전문물리치료학회지
• /
• 제10권1호
• /
• pp.29-35
• /
• 2003

Body weight support treadmill training is a new and promising therapy in gait rehabilitation of patients with hemiplegia. The purpose of this study was to identify the effects of body weight support treadmill training on gait and standing balance in patients with hemiplegia. Eighteen patients with hemiplegia participated in the study. A 10 m-timed walk test, measurements of step length and standing balance score were administered. Intervention consisted of body weight support treadmill training five times a week for 2 weeks. The data were analyzed by paired t-test. Body weight support treadmill training scoring of standing balance, step length and 10 m-timed walk test showed a definite improvement. Body weight support treadmill training offers the advantages of task-oriented training with numerous repetitions of a supervised gait pattern. The outcomes suggest that patients with hemiplegia can improve their gait ability and standing balance through body weight support treadmill training.

• Archives of Plastic Surgery
• /
• 제42권6호
• /
• pp.695-703
• /
• 2015

Background It is still difficult to prevent partial or full-thickness flap necrosis. In this study, the effects of a cream containing menthol and methyl salicylate on the viability of random-pattern skin flaps were studied. Methods Forty female Sprague-Dawley rats were divided into two equal groups. Caudally based dorsal random-pattern skin flaps were elevated, including the panniculus carnosus. In the study group, 1.5 mL of a cream containing menthol and methyl salicylate was applied to the skin of the flap, and saline solution (0.9%) was used in the control group. Upon completion of the experiment, flap necrosis was analyzed with imaging software and radionuclide scintigraphy. Histopathological measurements were made of the percentage of viable flaps, the number of vessels, and the width of the panniculus carnosus muscle. Results According to the photographic analysis, the mean viable flap surface area in the study group was larger than that in the control group (P=0.004). According to the scintigrams, no change in radioactivity uptake was seen in the study group (P>0.05). However, a significant decrease was observed in the control group (P=0.006). No statistically significant differences were observed between the groups in terms of the percentage of viable flaps, the number of vessels, or the width of the panniculus carnosus muscle (P>0.05). Conclusions Based on these results, it is certain that the cream did not reduce the viability of the flaps. Due to its vasodilatory effect, it can be used as a component of the dressing in reconstructive operations where skin perfusion is compromised.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2006년도 심포지엄 논문집 정보 및 제어부문
• /
• pp.129-131
• /
• 2006

The performances of neural network systems mainly depend on the kind and the number of input patterns for its training. Hence, the kind of input patterns as well as its number is very important for the character recognition system using back-propagation network. The more input patters are used, the better the system recognizes various characters. However, training is not always successful as the number of input patters increases. Moreover, there exists a limit to consider many input patterns of the recognition system for cursive script characters. In this paper we present a new character recognition system using the back-propagation neural networks. By using an additional neural network, an input pattern generation method is provided for increasing the recognition ratio and a successful training. We firstly introduce the structure of the proposed system. Then, the character recognition system is investigated through some experiments.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2007년도 심포지엄 논문집 정보 및 제어부문
• /
• pp.407-408
• /
• 2007

In this paper, we present a new method for classifying object using LVQ (Learning Vector Quantization) with flexible output layer. The proposed LVQ is a supervised learning method that dynamically generates output neurons and initializes automatically the weight vectors from training patterns. If the classes of the nearest output neuron is different from the class of the training pattern, a new output neuron is created and the given training pattern is used to initialize the weight vector of the created neuron. The proposed method is significantly different from the previous competitive learning algorithms in the point that the output neurons are dynamically generated during the learning process.

• 제어로봇시스템학회논문지
• /
• 제3권6호
• /
• pp.594-601
• /
• 1997

A new algorithm is proposed for training the multilayer perceptrion(MLP) in pattern classification problems to accelerate the learning speed. It is shown that the sigmoid activation function of the output node can have deterimental effect on the performance of learning. To overcome this detrimental effect and to use the information fully in supervised learning, an objective function for binary modes is proposed. This objective function is composed with two new output activation functions which are selectively used depending on desired values of training patterns. The effect of the objective function is analyzed and a training algorithm is proposed based on this. Its performance is tested in several examples. Simulation results show that the performance of the proposed method is better than that of the conventional error back propagation (EBP) method.

• The Journal of Korean Physical Therapy
• /
• 제17권1호
• /
• pp.108-122
• /
• 2005

The gait training strategy in very important things for central nervous system(CNS) injury patients. There are many method and strategy for regaining of the gait who had CNS injury. A human being has central pattern generator(CPG) is spinal CPG for locomotion. It is a neural network which make the cyclical patterns and rhythmical activities for walking. Sensory input from loading and hip position is essential for CPG stimulation that makes the central neural rhythm and pattern generating structure. From sensory input, the proprioceptive information facilitate proximal muscles that controlled in voluntarily from cortical level and visual and / or acoustical information facilitate distal muscles that controlled voluntarily from subcortical level. Gait training method can classify that is functional level and structural level. Functional level includ level surface gait, going up and down the stair. It is important to facilitate a guide tempo in order to activate the central pattern generators. During the functional test or functional activities, can point out the poor period in gait that have to be facilitate in structural level. There are many access methods with patient position and potentiality. The methods are using of rhythmic initiation, replication and combination of isotonic with standing position. Clinically using it on weight transfer onto the stance leg, loading response, loading response and pre-swing, terminal stance, up and downwards stairs.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2005년도 ICCAS
• /
• pp.308-313
• /
• 2005

For a real-time quadruped robot running control, there are many important objectives to consider. In this paper, the running control architecture based on global states, which describe the cyclic target motion, and central pattern is proposed. The main goal of the controller is how the robot can have robustness to an unpredictable environment with reducing calculation burden to generate control inputs. Additional goal is construction of a single framework controller to avoid discontinuities during transition between multi-framework controllers and of a training-free controller. The global state dependent neuron network induces adaptation ability to an environment and makes the training-free controller. The central pattern based approach makes the controller have a single framework, and calculation burden is resolved by extracting dynamic equations from the control loop. In our approach, the model of the quadruped robot is designed using anatomical information of a cat, and simulated in 3D dynamic environment. The simulation results show the proposed single framework controller is robustly performed in an unpredictable sloped terrain without training.