1. Inability to plan under uncertainty of dynamic environments: Conventionally, global planners rely on a complete map of the environment in order to calculate the ideal and collision-free path between the starting point and the ending point prior to execution of the robot. The original plans of those conventional algorithms must be revised accordingly if a dynamic environment is encountered (Dijkstra, 1959; Hart et al., 1968). In practise, environment of robots often includes various hazard sources that robots must avoid, for example landmines, fire in rescue duty, and war enemies. Since it is impossible or expensive to acquire their accurate locations, decision- makers know only their action ranges in most cases (Zhang et al., 2013). Mobile robots must be able to evade both static and moving obstacles (Ferguson et al., 2006). Algorithms such as sampling-based methods (Khaksar et al., 2012) are not suitable for online planning when involving moving obstacles, due to the fact that these methods are designed based on a static environment model. These models are time- consuming when applied to a dynamic environment (involving interpolation cycle during each update, see (Huptych & Röck, 2015)). Therefore, classical path planning methods such as Visibility Graph (Lozano-Perez, 1987), Voronoi Diagrams (Leven & Sharir, 1987), Grids (Weigl et al., 1993), Cell Decomposition (Regli, 2007), Artificial Potential Field (Khatib, 1985), Rule Based methods (Fujimura, 1991) and Rules Learning techniques (Ibrahim & Fernandes, 2004) are not practical (Mohanty & Parhi, 2013). Occasionally, these algorithms are optimized to handle a specific problem at the expense of sacrificing the performance of other parameters such as increasing of the computational cost of the algorithm.
In other literatures, other researchers like Phinni , uses a genetic-neuro-fuzzy approach to deal with the obstacleavoidance problem. He proposes to generate collision free path and have proper motionplanning as well as obstacleavoidance scheme. In the developed neuro-fuzzy approaches, training is done off-line with the help of Genetic Algorithm (GA), and the result shows that the system able to extract automatically the fuzzy rules and the membership functions in order to guide a wheeled mobilerobot. The membership function for distance is generated by using Takagi and Sugeno Approach. His results seem to lack in hardware implementation and verification since he only produces simulation results.
In , authors developed a path planner mechanism which is composed of Cellular Automata (CA) and Ant Colony Optimization (ACO) methods for multiple mobile robots. The proposed method is applicable in a dynamic environment and in real time applications. It also reduces the complexity and thus improves performance. A fuzzy-logic-assisted interacting multiple model (FLAIMM) technique is introduced in . Authors designed two extended Kalman filters (EKF) to solve the problem of the dynamics of mobile robots whereas an adaptive neurofuzzy inferencesystem (ANFIS) was used for predicting the slip. The proposed approach has been tested using a robot navigating in indoor environment which proves that the proposed work enhances the location accuracy. In , authors used the AdaptiveNeuroFuzzyInferenceSystem (ANFIS) Controller for mobilerobotnavigation in a cluttered environment with different conditions and various types of objects. The proposed ANFIS has successfully tested in simulation and in real experiment. The mobilerobot has avoided obstacles and reached to the desired goal in less time.
When autonomous robots navigate within outdoor environments (open space), they have to be endowed the ability to move through the environments, to move without collision with obstacles, and follow the direction with the help of (compass, GPS or camera). The basic concept of this project report is to design robot which can move to the north direction and avoid any obstacle on its way without human guidance or control. Arduino Mega 2560 was used as the “brain” to control the system of the robot [1-3, 6]. Figure 1 shows the prototype system.
Abstract — The aim of the present study is to explore applicability of artificial intelligence techniques such as ANFIS (AdaptiveNeuro-FuzzyInferenceSystem) in forecasting flood for the case study, Dharoi Dam on the Sabarmati river near village Dharoi in Kheralu Taluka of Mehsana District in Gujarat State, India. The proposed technique combines the learning ability of neural network with the transparent linguistic representation of fuzzysystem. ANFIS models with various input structures and membership functions are constructed, trained and tested to evaluate the models. Statistical indices such as Root Mean Square Error (RMSE), Correlation Coefficient (R), Coefficient of Determination (R 2 ) and Discrepancy Ratio (D) are used to evaluate performance of the ANFIS models in forecasting flood. This objective is accomplished by evaluating the model by comparing ANFIS model to Statistical method like Log Pearson type-III method to forecasting flood. This comparison shows that ANFIS model can accurately and reliably be used to forecast flood in this study.
IV. EXPERIMENTAL RESULTS Three experiments were performed. Of particular concern was the case where a dead-end existed and the robot could be trapped, such as with a U-shaped obstacle pattern. The first experiment used a cluttered environment but with a clear path to the goal, as shown in Fig. 6. The goal is shown with a triangle. The robot’s path is shown with circles. Obstacles are indicated with filled squares. The second experiment used multiple obstacles in a small U-shaped pattern, as shown in Fig. 7. The radius and the depth of this U- shape were no more than 30 cm. The third experiment used a combination of obstacles in a big U-shaped pattern, as shown in Fig. 8. Each environment was 2 m by 2 m in size. Navigation was attempted in these environments from a fixed starting point. In the figures shown, the robot started from the lower left corner and proceeded toward the goal located in the upper right corner.
The ANFIS model is trained using rainfall-runoff data of Kuantan catchment. The accuracy of the models is measured using RMSE and R Squared. The good values of these two measures have been obtained from the 4-input model indicating that this is the best input combination for the rainfall-runoff model of Kuantan catchment. The result also shows that ANFIS has the potential to be used for flood forecasting generally, or rainfall-runoff modelling specifically.
There is no doubt about the numerous researches that have been carried out in the area of artificial intelligence in medicine. A tremendous advancement in software and hardware industries has provided opportunities for computer scientist researchers to explore every aspect of artificial intelligence methods and techniques in building relevant systems, and devices for medical practioners. Karthikkalyan (2014) demonstrated a framework built with ANN and multilayer perceptron that classified liver diseases with improved accuracy performance. The model inputs were obtained from liver ultrasound images and employed MLP to detect the traits of any liver disease. In the study, the training of the dataset yielded 90% while testing validation yielded 95%. Olaniyi and Adnan (2013) designed an artificial neural network based model that used back propagation neural network based model that used backpropagation method of 10 hidden neurons and radial basis neural network of 20 hidden neurons, with only 6 inputs and one output to diagnose liver diseases into healthy and unhealthy liver. The dataset was obtained from BUPA in UCI Machine Learning Repository and the proposed model achieved 63% accuracy using backpropagation and 70% accuracy using radial basis technique. A similar work was proposed in Bahramirad et al. (2013), with the same input attributes and technique, the accuracy of 73% and 67% were obtained for radial basis and backpropagation technique respectively.
Currently, analysis of the patient’s polysomnography (PSG) is considered as the effective diagnosis of the sleep apnea . PSG requires overnight recordings of several electrophysiological signals during a night’s sleep such as electrocardiogram, respiratory effort, air- flow, etc. in sleep laboratories using specific systems and participating personnel . To derive respiration from electrocardiography (ECG) which is a simple, low cost and non-invasive recording is an alternative way. The correctness of such an idea using the comparison exam- ple recordings of the ECG derived respiration (EDR) with common respiration measurements are showed in Moody et al. work . Thus, different methods have been proposed to derive respiratory signal from the ECG [8-10]. Numerous studies show that EDR methods using
results but they are usually hampere by the fact that they consume long computing time because of the requirement for repetitive power flow calculations. Online voltage security assessment is a very useful but not yet becomes a widely used tool that measures the distance from the current operating condition at any time to the critical point. ADAPTIVENEUROFUZZYINFERENCE SYSTEMhave recently received widespread attention from researchers for this application. Most of ANFIS applications have been implemented using multi-layered feed-forward neural networks trained by back propagation because of their robustness to input and system noise, their capability of handling incomplete or corrupt input data. However, in typical power systems there are voluminous amount of input data. Then, the success of ANFIS applications also depends on the systematic approach of selecting highly important features which will result in a compact and efficient ANFIS. In this part, several voltage stability indicators are calculated. It should be mentioned here that this paper aims at implementing these already proposed indicators by ANFIS. The capability of monitoring proximity to voltage collapse was tested beforehand, but unfortunately due to space limitation and scope of this paper the complete results cANFISot be presented.
Now a day’s Robotics is part of today’s communication & communication is part of advancement of technology, so we decided to work on ROBOTICS field, and design something which will make human life today’s aspect. Primary things required for mobilerobot is obstacleavoidance and path finding in unknown environment and not to get damaged. There are some very important methods for mobilerobot like wall following, edge detection, line following. One of the important use is wall following robot on a floor to clean floor. The most commonly employed method for obstacleavoidance is based on simple object detection using Infrared Ray sensor. A disadvantage with obstacleavoidance based on edge detecting is the need of the robot to stop in front of an obstacle in order to provide a more accurate measurement.
Predicting the rainfall is a complex process due to the incidence of rain is non linear and dynamic. Many studies have tried to modeling rainfall forecasting to predict accurately. Such as statistical model approach like Multiple Linear Regression and artificial intelligence approaches like FuzzyInferenceSystem , Artificial Neural Networks (ANN) and AdaptiveNeuroFuzzyInferenceSystem (ANFIS). Comparison of these approaches show ANFIS and GA predict more accurate than other methods. And ANFIS that use hybrid training method give better results . Another study shows that ANFIS is better than ANN in forecasting rainfall monthly . Since the popularity of ANFIS in predicting rainfall, further investigation is needed in forecasting rainfall using ANFIS.
A Line follower is a machine that can follow a path. The path can be visible like black tape on a white surface (or vice-versa) or it can appear invisible like a magnetic field . The essential method to build a line follower is sensing a line and manoeuvring the robot to stay on course, while constantly tuning its error (wrong moves) using feedback mechanism and forms a simple yet effective closed loop system. Line following robot is mobilerobot that widely used in different are, especially in industry field. These robots function as material carrier to deliver the product from one place to another where the conveyor and rail not possible. Apart from line following capabilities, this robot should also capability to navigate. Sensor positioning also plays an importance role in optimizing robotnavigation performance. Over the past, there are many method for controller has been develop to increase the performances of the robot in term of navigation such as PID controller, PI controller, neutral network and fuzzy logic.
Abstract: In the present generation, surroundings molding has become a primary task in the navigation of mobile robots. It is very important for an autonomous robot to automatically build an abstract over surroundings as it is not possible for human intervention all the time. A robot must work on both the localization and mapping to adapt itself to the surroundings. In this work, a surrounding molded algorithm has been developed which purely based on the Extended Kalman Filter (EKF). The mapping and localization issue, called Simultaneous Localization and Mapping (SLAM), is an operation of developing an abstract map on the surroundings to identify the robot coordinates. The SLAM algorithm mainly begins with a mobilerobot in an undetermined position without previous knowledge of the surroundings on the map.
This is to certify that the thesis entitled “Navigation of Mobil RobotusingFuzzy Logic” is the bona fide work of Krushna Shankar Sethi and SanjeevPothen Jacob under the Guidance of Dr. D.R.K.Parhi for the requirement of the award of the degree of BACHELOR OF TECHNOLOGY specialization “Mechanical Engineering” andsubmitted in the Department of Mechanical Engineeringat National Institute of Technology Rourkela, During the period 2012-2013 .
Abstract: Mobile robots have been successfully used in many fields due to their abilities to perform difficult tasks in hazardous environments, such as robot rescuing, space exploring and their various promising applications in the daily lives. Robot path planning is a key issue in robotnavigation which is a kernel part in mobilerobot technology. Robot path planning is to generate a collision-free path in an environment while satisfying some optimization criteria. Mobilerobot path planning is a nondeterministic polynomial time (NP) problem, traditional optimization methods are not very effective to it, which are easy to plunge into local minimum. In this research work, an evolutionary algorithm to solve the robot path planning problem is devised. A method of robot path planning in partially unknown environments based on A star (A*) algorithm was proposed. The proposed algorithm allows a mobilerobot to navigate through static obstacles and finds its path in order to reach from its initial position to the target without collision. In addition, the environment is partially unknown for the robot due to the limit detection range of its sensors. The robot processor updates its information during the motion. The simulations are performed in different static environments, and the results show that the robot reaches its target with colliding free obstacles. The optimal path is generated with this method when the robot reaches its target. The simulation results are developed by MATLAB environments.
In order to navigate in unknown environments the robot has to get information, process them and providing an optimal solution for the task. One particular issue for reaching the target position is avoiding the obstacles encountered in robot’s path. There is a lot of research work carried out so far in this field but we will focus into implementing a fuzzyinferencesystem that has a high accuracy and adapts better to our requirements. The fuzzyalgorithm used for obstacleavoidance is a Mamdani inferencesystem. In the fuzzification stage the systems requires to map the inputs from two ultrasonic sensors to values ranging from 0 to 1 using a set of input membership functions. The two ultrasonic sensors are placed on Nao’s chest. Each sensor is equipped with transmitter and receiver, with a distance of 7 cm between the two sonars transmitters. The sonars effective cone is 600, frequency=40kHz and the detection range varies between 0.25 to 2.55 m, with a resolution of 1cm.
The visual tracking system installed on the follower enables the follower to follow the leader robustly in a variety of environment. Despite this advantage, there are quite a number of disadvantages for using vision system. One of the main disadvantages of the vision system is that it is costly to build. This is due to the high computational power required to process the raw data obtained from vision sensor. This reason makes the vision tracking system less preferable to be used.
Heart signals allow for a comprehensive analysis of the heart. Electrocardiography (ECG or EKG) uses electrodes to measure the electrical activity of the heart. Extracting ECG signals is a non- invasive process that opens the door to new possibilities for the application of advanced signal processing and data analysis techniques in the diagnosis of heart diseases. With the help of today’s large database of ECG signals, a computationally intelligent system can learn and take the place of a cardiologist. Detection of various abnormalities in the patient’s heart to identify various heart diseases can be made through an AdaptiveNeuro-FuzzyInferenceSystem (ANFIS) preprocessed by subtractive clustering. Six types of heartbeats are classified: normal sinus rhythm, premature ventricular contraction (PVC), atrial premature contraction (APC), left bundle branch block (LBBB), right bundle branch block (RBBB), and paced beats. The goal is to detect important characteristics of an ECG signal to determine if the patient’s heartbeat is normal or irregular. The results from three trials indicate an average accuracy of 98.10%, average sensitivity of 94.99%, and average specificity of 98.87%. These results are comparable to two artificial neural network (ANN) algorithms: gradient descent and Levenberg Marquardt, as well as the ANFIS preprocessed by grid partitioning.
In this section, the structure of our system and function- alities of each component are described. The structure of ANFIS-based implicit authentication system is presented in Fig. 2 which includes an activity monitoring module, a scor- ing module, a reference computation module and an ANFIS module. The learning capability of ANFIS allows for training of specific fuzzyinference model based on given input-output user data without manual interference. Therefore, this trained fuzzy model represents well the user’s behaviour and can then be used to make final authentication decision to provide access control. The deployment of the system can be divided into two phases: training phase and deployment phase. During the training phase, we use generated input (scores, references) and output (classification labels) variables to train ANFIS module offline. When the training is completed the deployment phase starts, this enables the system to infer the authenticity of the current user behaviour by taking in real-time computed scores and references.