SEARCH RESULT

The number of people with some kind of physical disabilities in the world is around 1 billion. The rate of people who need physiotherapy increases with the aging world population every passing day. Physiotherapy may not be completed with exercises made only in hospital. Therefore, patients should do the exercises given by the physiotherapist at home. The main problem here is that only 31% of the exercises performed by the patients are done correctly. Nowadays, in addition to traditional treatment methods there are studies on telerehabilitation to solve such problems. The aim of this study is to ensure that the shoulder rehabilitation exercises are performed at home by the patients and to provide physiotherapists with meaningful data about the exercises. In this study, Improved Shoulder Physiotherapy Application (ISPA) was presented using Microsoft Kinect 2 for shoulder rehabilitation exercises. ISPA is a hardware and software product that uses the joints on the patient's skeletal system. In the proposed system, the angular values are calculated using the joint points taken from Kinect 2 and the patients are simultaneously guided to do shoulder rehabilitation exercises correctly.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Burakhan Çubukçu Uğur Yüzgeç Raif Zileli Ahu Zileli

Neural networks (NNs) is used to solve many engineering and science problem. Generally, feedforward architecture is preferred and gradient-based learning algorithms are extensively operated to tune all parameters of NN iteratively. This training method is a conventional one, but training process takes a long time due to the slowness of gradient-based learning algorithms. This slowness has been an important drawback in their applications. To overcome this disadvantage, extreme learning machine (ELM) concept introduced to science community in near past. Essentially, ELM is a data-driven learning algorithm for single-hidden layer feedforward neural networks (SLFNs). This algorithm provides extremely fast learning speed. In this study, performance of SLFNs learned by ELM algorithm is investigated on the problem of highly nonlinear dynamic system identification. As a result of studies on selected benchmark problems in the literature, it has been seen that ELM may not provide a good generalization success due to randomly chosen the number of hidden nodes and weight parameters for inputs in SLFN. For both the training and the test data set, very poor results have been obtained and observed surprisingly during the above-mentioned studies. Here, a simple heuristic approach has been proposed in this study in order to eliminate this bad situation and the findings obtained with this approach are discussed. Based on the obtained experimental results, it has been shown that the proposed approach determines the optimal the number of hidden nodes and a reasonable random selection of input weights required for a good generalization performance.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Cihan Karakuzu Uğur Yüzgeç

Unmanned Aerial Vehicles (UAVs), are air aircraft that can fly autonomously without human control or that can be controlled by a ground station. In parallel with the technological developments of the present day, it has been observed that the use of UAVs is also greatly increased. In this context, new working areas were needed together with the widespread use of UAVs. As a result of the researches, the autonomous control works was seen to become the foreground. Within the scope of this study, an autonomous landing system has been implemented for UAVs. The images taken with the help of the camera located at the bottom of the UAV are subjected to image processing techniques on the Linux based operating system to determine the landing track. In the scope of the study, the landing station detected is a circle with a "H" in the middle. The letter "H" on the landing station can be changed color according to the location. For example; the landing station can be red in football ground, can be blue in empty ground. Images taken from the camera are subject to the image processing technique. First of all, the image is removed from noises. Then, it is classified according to the “H” letter color on landing station. Pixels which are same with "H" letter are changed to white color, others changed to black color. So that, it is determined how many shaped has been found. The shapes are ordered according to their size and compared with the landing station. If the similarity rate is over 75%, the target is determined.Once the determination, the altitude of UAV will gradually decrease and the image acquisition and measurement process will be repeated until it reaches a desired position. After arriving at the desired position (Eg z<=20 cm), the UAV performs descent by stopping engines.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Çağrı Çiçekdemir H. UCGUN Uğur Yüzgeç M. KESLER

In daily life we may encounter many different optimization problems. There are many different algorithms developed for solving these problems. The abundance of optimization problems in every field, such as science, engineering, economics, health, communication, production, etc., have made these algorithms are more popular. These algorithms are divided into two main groups as mathematical based and meta-heuristic based algorithms. The meta-heuristic algorithms are inspired by the nature events in general. In this study, we deal with the Vortex Search (VS) algorithm, which is one of the meta-heuristic algorithms. This algorithm is inspired from the vortex pattern produced by the vortical flow of stirred fluids. We proposed some improvements on vortex search algorithm to increase the performance of the original VS algorithm. To show the performance of Improved Vortex Search (ImpVS) algorithm, we used well-known meta-heuristic algorithms, such as Differential Evolution (DE) algorithm, Particle Swarm Optimization (PSO) algorithm, Artificial Bee Colony (ABC) algorithm, Simulated Annealing (SA) algorithm and original Vortex Search (VS) algorithm. The obtained Benchmark results show that the proposed ImpVS algorithm has got the competitive performance for the optimization problems.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Gökhan Uçar Uğur Yüzgeç

Today, the use of meta-heuristic algorithms in solving non-linear and complex problems is becoming popular. These algorithms are inspired by natural selection, swarm intelligence, physical events and nature. Crab mating optimization algorithm is one of the meta-heuristic algorithms which was developed by V.R. Chifu in 2014. The crab mating optimization algorithm imitates the mating behavior of crabs in nature. One of the biggest shortcomings of this algorithm is that it has got the quite long running times for optimization problems. In this study, to increase the running speed of the original crab mating optimization algorithm, the mating process of the crabs in population was modified. A male crab mates with only a female crab in the new mating procedure. Therefore, the name of the proposed meta-heuristic algorithm comes from this new mating process as a Monogamous Crab (MC) mating algorithm. Vehicle routing problem (VRP) is the popular combinatorial optimization problem. VRP relates to the most appropriate route design to be delivered to a range of customer service by a number of fleets. In this study, the developed MC algorithm is adapted for the vehicle routing problem. The performance of MC algorithm was compared with those of metaheuristic algorithms. The results show that MC algorithm provides promising and competitive performance.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Burakhan Çubukçu Uğur Yüzgeç

AntLion Optimization (ALO) algorithm is one recent of the meta-heuristic algorithms that was developed by Mirjalili in 2015. ALO algorithm imitates the antlion's hunting behaviour in its larvae phase. The long run time of ALO algorithm is the biggest disadvantage of this algorithm. To overcome this deficiency, we proposed some improvements on the mechanisms of the original ALO algorithm. In order to improve the ALO algorithm, firstly, the random walking distance was changed as twenty percent of maximum iteration instead of the maximum iteration number in the original ALO algorithm. We defined new movements between boundaries around the antlion on the phase of trapping antlion pits. In addition, the boundary checking process, the catching prey and rebuilding the pit were developed. The parallel machine scheduling problem (PMS) is defined that it is a set of independent jobs to be scheduled on a number of parallel machines. Scheduling process optimizes the production job sequences in terms of the different patterns. When there are the similar type of machines to be existing in multiple numbers, the jobs can be scheduled over these parallel machines at the same time. To show the performance of improved ALO (IALO) algorithm, some of well-known meta-heuristic algorithms were used in comparison works. The obtained PMS results show that the proposed IALO algorithm has very competitive results in terms of the mean, best, worst cost and standard deviation metrics.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Haydar Kılıç Uğur Yüzgeç

Path planning problem plays an important role in mobile robot works. The robotic systems use intelligence algorithms to plan the path of the robot from one point to the other point. The main goal of path planning is to find the allowable movements of a robot in an environment with obstacles. These motions involve a path free of collision from the start position to the target position. In this study, Gray Wolf Optimization (GWO) algorithm was adapted to solve robot path planning problem. GWO algorithm imitates the hunting behavior and social leadership of gray wolves in nature. The leadership hierarchy consists of four grey wolf groups: alpha, beta, delta, and omega wolves. This algorithm comprises hunting mechanism with three stages: searching for prey, encircling prey, and attacking prey. In the test simulations of the robot path planning, we used a map with three circular obstacles. GWO algorithm was adapted to this problem. While finding the candidate solutions in path planning, three coordinate points are used between start and target points. For each iteration, these coordinate points are updated by GWO algorithm. If the solution point is in the obstacle zone, then violation is added to the cost function. The performance of GWO algorithm was evaluated with those of meta-heuristic algorithms for solving the robot path planning problem. The results obtained by GWO algorithm show that the optimal path is found for used test map.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Lokman DOĞAN Uğur Yüzgeç

Quadratic Assignment Problem (QAP) is based the facilities allocation, and it is a difficult combinatorial optimization problem. The objective of this problem is to make total assignment cost minimum while being assigned facilities to locations that are already known. To solve this problem, the different methods are used in the literature. Meta-heuristic algorithms are ones from these methods and in this study, we present a new version of recent antlion optimization algorithm for QAP. AntLion Optimization (ALO) algorithm was developed by Mirjalili in 2015. ALO algorithm is based on the hunting behaviour of the antlion. This algorithm comprises five stages: random walking mechanism, constructing trap, trapping in the antlion’s pit, sliding ants in the pit, catching the prey and reconstructing the pit. Although ALO algorithm is successful in benchmark functions of multi dimensions, it has got some drawbacks. The most notable improvement is the use of tournament method instead of roulette wheel method. In ALO algorithm, the antlion is chosen from the population by roulette wheel method for using in each ant's random walking model. The roulette wheel method is more successful in maximization problems. In the minimization problems, the tournament selection method is more efficient method than the other selection methods Therefore, we used the tournament selection method in this study instead of the roulette wheel method on random walking mechanism. This proposed algorithm has been called the tournament selection based antlion optimization algorithm (TALO). To evaluate the performance of TALO algorithm, we used well-known meta-heuristic algorithms. The results provide the proposed TALO algorithm has the best performance in comparison with those of the other algorithms.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Haydar Kılıç Uğur Yüzgeç

In this study, the Opposition based learning Spiral Optimization Algorithm (OBLSOA) is presented for training Multi-Layer Perceptron (MLP). The main idea of Spiral Optimization Algorithm (SOA) is based on the dynamic step dimension in its spiral path trajectory. The primary opposition based learning (OBL) concept first was come from the Yin-Yang symbol in the ancient Chinese philosophy. According to OBL concept, if a candidate point is far from the solution, the opposite point of this candidate can be closer to the solution than that point. We applied OBL concept to spiral optimization algorithm for training MLP. OBLSOA comprises two main stages: the first is the opposition-based learning population initialization and the other is opposition-based learning generation jumping. To evaluate the performance of the proposed OBLSOA, we used eight standard datasets including four classification datasets (XOR, balloon, Iris, breast cancer) and three function-approximation datasets (sigmoid, cosine, and sine). The performance proposed OBLSOA was compared with the original SOA for all datasets in terms of the Mean Square Error (MSE) metric. The training and test results show that the proposed OBLSOA is able to be provide very competitive and effective in training MLPs.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Uğur Yüzgeç Cihan Karakuzu