Showing papers by "Anupam Shukla published in 2012"

Location Estimation Based on Semi-Supervised Locally Linear Embedding (SSLLE) Approach for Indoor Wireless Networks

Abstract: Location aware computing is popularized and location information use has important due to huge application of mobile computing devices and local area wireless networks. In this paper, we have proposed a method based on Semi-supervised Locally Linear Embedding for indoor wireless networks. Previous methods for location estimation in indoor wireless networks require a large amount of labeled data for learning the radio map. However, labeled instances are often difficult, expensive, or time consuming to obtain, as they require great efforts, meanwhile unlabeled data may be relatively easy to collect. So, the use of semi-supervised learning is more feasible. In the experiment 101 access points (APs) have been deployed so, the RSS vector received by the mobile station has large dimensions (i.e. 101). At first, we use Locally Linear Embedding to reduce the dimensions of data, and then we use semi-supervised learning algorithm to learn the radio map. The algorithm performs nonlinear mapping between the received signal strengths from nearby access points and the user’s location. It is shown that the proposed scheme has the advantage of robustness and scalability, and is easy in training and implementation. In addition, the scheme exhibits superior performance in the nonline-of-sight (NLOS) situation. Experimental results are presented to demonstrate the feasibility of the proposed SSLLE algorithm.

Modified A* Algorithm for Mobile Robot Path Planning

Abstract: Robot path planning is about finding a collision free motion from one position to another. Efficient algorithms for solving problems of this type have important applications in areas such as: industrial robotics, computer animation, drug design, and automated surveillance. In this paper, a modified A* algorithm is used for optimizing the path. Different from the approaches that only choose the shortest routes, this method estimates the energy consumption and chooses the most energy efficient routes. As mobile robots are powered by batteries, their energy is limited. Therefore, how to minimize energy consumption is an important problem. The basic idea is to minimize unnecessary stops and turns for mobile robots that cause acceleration and deceleration and consumes significant energy. Simulation results are presented on various environments with different levels of complexity depending on the density of the obstacles. The effectiveness of the proposed approach is evaluated in terms of number of movement steps, path length, energy consumption, number of turns and time. The experimental results show that our approach can provide effective path by reducing the number of turns compared to A*, thus saving energy. All paths generated were optimal in terms of length and smoothness.

Multi-Robot Exploration in Wireless Environments

Abstract: This paper presents a multi-robot exploration approach for application in wireless environments. The challenges generally faced by a robot team are to maintain network connectivity among themselves, in order to have an accurate map of the environment at each instant and have an efficient navigation plan for moving toward the unexplored area. To address these issues, we focus on the integration of such connectivity constraints and take navigation plan problems into account. A modified A* based algorithm is proposed for planning the navigation of the robots. A communication protocol based on the concept of leader-follower is developed for maintaining network connectivity. Mobile robots typically use a wireless connection to communicate with the other team members and establishes a Mobile Ad Hoc NETwork among themselves. A communication route is established between each robot pair for exchanging local map data, in order to achieve consistent global map of the environment at each instant. If the routes have multiple hops, this raises the problem of message delaying because time delay accumulates per hop traveled. The purpose of the proposed Leader Follower Interaction Protocol is to reduce the total number of hop counts required for all transmissions between robot pairs. This is different from the centralized approach where the leader is a fixed base station. The role of leader in the proposed approach switches from one robot to others as network’s wireless topology changes as robots move. Simulation results show the effectiveness of communication protocol, as well as the navigation mechanism.

Analysis & modeling multi-breeded Mean-Minded ant colony optimization of agent based Road Vehicle Routing Management

Abstract: In this paper Road Vehicle Routing Management is being analyzed and modeled considering multi-parameter scheme and a new modified Mean-Minded ant colony optimization (ACO) heuristic is used to optimized the different options that several vehicle system can avail to reach its destination The model has taken care so that the busy roads are avoided and congestion never arises The aim of this work is to uniformly distribute the traffic and the movement of vehicles through some selected points is enumerated to see the distribution of vehicles in all paths Some modification of ant-colony optimization algorithm is made and instead of running one breed of ants, here multi breeds are being initialized to demarcate multi - objective and multi - capacitive vehicles The pheromone density no longer depends on the number of ants, but is actually a function of the parameters which it is seeking, instead of the traditional pheromone trail function used So in a nutshell the pheromone evaporation functions will a different one and evaporation criteria will be how much the ant is happy while passing through that road Analogy can be derived as a road with scattered food of different type and several types of ants are passing, and each time they see food of their liking they eat them and spread pheromone to attract more insects of its types, however that eaten food is refilled and the supply will never end The results obtained showed that the ACO has been successful up to a certain extent in channeling the traffic in various routes of the system irrespective of its kind and considering the road network as a dynamic system with varying parameters, the vehicle distribution has been near uniform except fluctuations arising due to dynamicity error

Robot Path Planning using Dynamic Programming with Accelerating Nodes

Abstract: We solve the problem of robot path planning using Dynamic Programming (DP) designed to perform well in case of a sudden path blockage. A conventional DP algorithm works well for real time scenarios only when the update frequency is high i.e. changes can be readily propagated. In case updates are costly, for a sudden blockage the robot continues moving along the wrong path or stands stationary. We propose a modified DP that has nodes with additional processing (called accelerating nodes) to enable different segments of the map to become informed about the blockage rapidly. We further quickly compute an alternative path in case of a blockage. Experimental results verify that usage of accelerating nodes makes the robot follow optimal paths in dynamic environments.

Gender Identification in Human Gait Using Neural Network

Abstract: Biometrics is an advanced way of person recognition as it establishes more direct and explicit link with humans than passwords, since biometrics use measurable physiological and behavioural features of a person. In this paper gender recognition from human gait in image sequence have been successfully investigated. Silhouette of 15 males and 15 females from the database collected from CASIR site have been extracted. The computer vision based gender classification is then carried out on the basis of standard deviation, centre of mass and height from head to toe using Feed Forward Back Propagation Network with TRAINLM as training functions, LEARNGD as adaptation learning function and MSEREG as performance function. Experimental results demonstrate that the present gender recognition system achieve recognition performance of 93.4%, 94.6%, and 94.7% with 2 layers/20 neurons, 3 layers/30 neurons and 4 layers/30 neurons respectively. When the performance function is replaced with SSE the recognition performance is increased by 2%, 2.4% and 3% respectively for 2 layers/20 neurons, 3 layers/30 neurons and 4 layers/30 neurons.The above study indicates that Gait based gender recognition is one of the best reliable biometric technology that can be used to monitor people without their cooperation. Controlled environments such as banks, military installations and even airports need to quickly detect threats and provide differing levels of access to different user groups.

PERDURABLE: A Real Time Commit Protocol

Abstract: Real Time Systems (RTS) are those systems, for which, correctness depends not only on the logical properties of the produced results but also on the temporal properties of these results [27]. Typically, RTS are associated with critical applications in which human lives or expensive machineries may be at stake. Examples include telecommunication systems, trading systems, online gaming, chemical plant control, multi point fuel injection system (MPFI), video conferencing, missile guidance system, sensor networks etc. Hence, in such systems, an action performed too late (or too early) or a computation which uses temporally invalid data may be useless and sometimes harmful even if such an action or computation is functionally correct. As RTS continue to evolve, their applications become more and more complex, and often require timely access and predictable processing of massive amounts of data [21]. The database systems especially designed for efficient processing of these types of real time data are referred as distributed real time database system (DRTDBS). Here, data must be extremely reliable and available as any unavailability or extra delay could result in heavy loss. Business transactions being used in these applications in the absence of real time could lead to financial devastations and in worst case cause injuries or deaths [20].

Multi Robot Exploration through Pruning Frontiers

Abstract: In this paper, an approach to multi robot exploration is presented. One of the key issues in multi robot exploration is how to assign target locations to the individual robots and how to better distribute the robots over the environment. The proposed technique applies a well-known unsupervised clustering algorithm (k-means) in order to fairly divide the space into as many disjoint regions as available robots. Hungarian Method is used for the assignment of robots to the individual regions with the task to explore the corresponding area. To drive the robots around the environment, a frontier ‘regions on the boundary between open space and unexplored space’ based navigation strategy is used to decide where to move next, according to the data collected so far. Furthermore, we discuss improvements to the frontier based exploration strategy, by pruning the frontier cells that further reduces the computational time. The numbers of candidate locations are evaluated based on three criteria: number of unknown cells, number of known cells and real path travelling cost. Simulations are presented to show the performance of the proposed technique. This method can best be applied in search and rescue operations, partitioning helps to explore different regions of the workspace parallely by different robots instead of concentrating efforts in particular spot, pruning helps to make movement decisions much faster, the result is that the potential victims in a region will not have to wait much longer.

Robotic path planning using hybrid genetic algorithm particle swarm optimisation

Abstract: The problem of robotic path planning has always attracted the interests of a significantly large number of researchers due to the various constraints and issues related to it. The optimisation in terms of time and path length and validity of the non-holonomic constraints, especially in large sized maps of high resolution, pose serious challenges for the researchers. In this paper we propose hybrid genetic algorithm particle swarm optimisation (HGAPSO) algorithm for solving the problem. Diversity preservation measures are introduced in this applied evolutionary technique. The novelty of the algorithm is threefold. Firstly, the algorithm generates paths of increasing complexity along with time. This ensures that the algorithm generates the best path for any type of map. Secondly, the algorithm is efficient in terms of computational time which is done by introducing the concept of momentum-based exploration in its fitness function. The indicators contributing to fitness function can only be measured by exploring the path represented. This exploration is vague at start and detailed at the later stages. Thirdly, the algorithm uses a multi-objective optimisation technique to optimise the total path length, the distance from obstacle and the maximum number of turns. These multi-objective parameters may be altered according to the robot design.

Distributed location estimation system using WLAN received signal strength fingerprints

Abstract: Location Estimation has become important for many applications of indoor wireless networks. Received Signal Strength (RSS) fingerprinting methods have been widely used for location estimation. The accuracy and response time of estimation are critical issue in location estimation system. Most of the location estimation system suffers with the problem of scalability and unavailability of all the access points at all the location for large site. In this paper, we have proposed a distributed location estimation method, which divide the location estimation system into subsystems. Our method partition the input signal space and output location space into clusters on the basis of visibility of access points at various locations of the site area. Each cluster of input signal space together with output location subspace is used to learn the association between RSS fingerprint and their respective location in a subsystem. We have compared our results with benchmark RADAR method. Experimental results show that our method provide better results in terms of accuracy and response time in comparison to centralized systems, in which a single system is used for large site.

View variations effect in gait recognition and performance improvement using fusion

Abstract: This paper presents the view variations effect in gait recognition. Here different variation is created based on the walking in different angle of person with respect to particular line. Initially we are showing the result of what is the recognition rate when we are taking different view training and testing image. These results show that recognition through gait is affected by view variation. So we are using fusion strategy to improve the results. Here different view image are fuse using PCA algorithm. We are also showing the result of fusion based gait recognition in different view. All experiments are performed on CASIA gait data base. Our results show that fusion based gait recognition is more effective when we want recognition people in different view. Gait recognition is one kind of biometric technology that can be used to monitor people without their cooperation, it is also difficult to conceal.

Hybrid computing based intelligent system for breast cancer diagnosis

Abstract: Breast cancer is one of the major causes of death in women which accounts one out of eight. As primary cause is still unknown, early detection increases better treatment and improves total recovery. We present some novel hybrid approaches for classification of breast cancer. Artificial Neural Network (ANN) which suffers credit assignment problem can be avoided by modular and evolutionary artificial neural network which achieves simple and small individual neural network. Ensemble of ANN is to obtain a more reliable and accurate ANN. Evolutionary Neural Network (ENN) is used for optimization of neural network learning and design. The best accuracies achieved for diagnosis are around 99% using breast cancer datasets.

A novel approach for path planning, area explore and area retrieval of an autonomous robot using the low-cost infrared sensors

Abstract: This paper presents a set of novel techniques along with the algorithms that are proposed to analyze various physical characteristics of arena autonomously by a robot. The proposed methodology and algorithms identify such characteristics on the basis of sensing intense and sudden change in physical characteristics of the area which it being explored. The ultimate function of these novel techniques is to efficiently explore the arena and successfully reconstruct it after completing whole exploration process. It would provide not only be a cost-effective solution for the problem of traversing the arena of large dimensions but also would be a swift and uncomplicated method to do so. With the major objective of proposing an inexpensive solution to the problem, the proposed methodology is designed to work with low-cost infrared sensors in performing all the necessary and inter-related tasks of path planning, area explore and area retrieval. The proposed methodology was implemented in an arena containing rapidly changing physical characteristics with the target of traversing whole arena and finally providing the output based on how accurately changes detected to give an inferential analysis of whole arena. In the process, we also came with a novel approach to plan the path with low cost infrared sensors instead of using high cost long range cameras. The methodology was able to give 82% percent correct results with respect to the parameters set for the manual testing.

Breast Cancer Data Prediction by Dimensionality Reduction Using PCA and Adaptive Neuro Evolution

Abstract: In this paper a new approach for the prediction of breast cancer has been made by reducing the features of the data set using PCA principal component analysis technique and prediction results by simulating different models namely SANE Symbiotic, Adaptive Neuro-evolution, Modular neural network, Fixed architecture evolutionary neural network F-ENN, and Variable Architecture evolutionary neural network V-ENN. The dimensionality reduction of the inputs achieved by PCA technique to an extent of 33% and further different models of the soft computing technique simulated and tested based on efficiency to find the optimum model. The SANE model includes maximum number of connections per neuron as 24, evolutionary population size of 1000, maximum neurons in hidden layer as 12, SANE elite value of 200, mutation rate of 0.2, and number of generations as 100. The simulated results reflect that this is the best model for the prediction of the breast cancer disease among the other models considered in the experiment and it can effectively assist the doctors for taking the diagnosis results as its efficiency found to be 98.52% accuracy which is highest.

MDABC: Motif Discovery Using Artificial Bee Colony Algorithm

Abstract: Motif discovery is one of the most popular problems in molecular biology. There are many solutions provided by researchers. In this paper we have adopted the features of the Artificial beecolony algorithm ABC and designed a sequential algorithm to solve motif problem and named it MDABC. Artificial bee colony algorithm is a population based heuristic search technique used for optimization problem. We have performed experiments with the nucleotide sequences of homo sapiens human and mouse viz. CDRT4, MACF1, Zfa, TNFRSF19 and TICAM2. The ABC algorithm applied to the CDRT4, MACF1, Zfa, TNFRSF19 and TICAM2 DNA sequence for determining the motif of length 10, 20 and 30 using the maximum number of cycle MCN or the maximum number of generation is equal to 250, 500 and 1000. Our result shows a promising superior performance of the algorithm. As we will see, our results surpass the results obtained by other approaches proposed in the literature.

Guided reproduction in differential evolution

Abstract: Differential Evolution (DE) is a vector population based and stochastic search optimization algorithm. DE converges faster, finds the global minimum independent to initial parameters, and uses few control parameters. DE is being trapped in local optima due to its greedy updating approach and inherent differential property. In order to maintain the proper balance between exploration and exploitation in the population a novel strategy named Guided Reproduction in Differential Evolution(GRDE) algorithm is proposed. In GRDE, two new phases are introduced into classical DE; first phase enhance the diversity while second phase exploits the search space without increasing the function evaluation. With the help of experiments over 20 well known benchmark problems 3 real world optimization problems; it has been shown that GRDE outperform as compared with classical DE.

Efficient strategy for co-ordinated multirobot exploration

Abstract: Area exploration is the key behind many researches in robotics. Numerous exploration problems have been solved based on the concept of frontiers that can be defined as the boundary between the explored and unexplored cell. In this paper we considered the problem of energy efficient exploration with a team of robots. An approach has been proposed that chooses the next frontier based on the direction strategy which simultaneously takes into account the location of other robot as well. Whenever a frontier has to be assigned to a specific robot, the utility of the unexplored area visible from this position is increased so that at a time not more than single robot moves to the same cell. Based on the direction penalty is calculated for each target points. Then the frontier having minimum utility and penalty has been chosen as the next target point. The robot moves to that frontier cell using energy efficient A* algorithm. The energy efficient A* gives optimal results taking into account energy consumed for stops and turns. Java based platform is used to run the simulation. Proposed algorithm has been tested on various test maps. The result shows that our technique accomplishes the mission quickly as compared to single robot energy efficient exploration and effectively distributes the robots over the environment.

An energy efficient platform for mobile robot

Abstract: Exploration of an unknown environment by autonomous mobile robot is a fundamental concern in mobile robotics. Today most of the mobile robots are powered by batteries so their energy and operation times are limited. Therefore, how to minimize energy consumption and obstacle avoidance becomes an important problem. Frontier-based method is known to be most efficient for robot exploration system. In this paper, we propose a novel energy efficient platform based on the notion of frontiers for mobile robot. The robot selects the next frontier to visit based on the robot's current direction and relative location of frontier cells. Our frontier selection method can reduce duplicate coverage and thus save energy and reduce the distance also. Distance to frontier are computed using energy efficient A* algorithm. We estimate the energy consumption and choose the most energy-efficient route to move to that frontier considering energy of stop and turnings. We tested the algorithm on various types of environment maps. The experiments were conducted assuming equal velocity for the robot during whole exploration. All paths generated were optimal in terms of energy consumption and turns. The robot was easily able to escape a variety of obstacles and reach the goal in an optimal manner.