International Journal of Intelligent Machines and Robotics (8 papers in press)
Robust Industrial Vision System for Mechanical Parts Recognition
by Tushar Jain, Meenu , H.K. Sardana
Abstract: Automated recognition of mechanical parts is a task in manufacturing that has been automated at a comparatively slow pace. Nearly all of the existing object recognition systems, with the exception of very few experimental systems have been designed to recognize a single object. In this paper, this problem is solved in a great manner so that the same process can handle different 2-D recognition applications. Color images are used during object recognition. The Fourier descriptor method has been adopted for recognition of mechanical parts. This method recognizes an object by extraction of features from an object image. The objects may be classified using Artificial Neural Network (ANN). For training and testing in either case, the features are extracted by presenting the object in different orientations. A feed forward neural network structure that learns the characteristics of the training data through the back-propagation learning algorithm is employed. The emphasis is put on the choice of network architecture and setting of different parameters. The study also considers the effects of various user-defined parameters and noting their effect on classification accuracy. The effect of orientation angle of the object and sample size on overall accuracy is also considered on the used classifier.
Keywords: Automated Recognition; Mechanical parts; Object recognition systems; Image Processing; Artificial Neural Network; Intelligent Machines; and Robotics.
Optical dynamic balancing of shaking force and shaking moment for planar mechanisms
by Samiksha Agarwal, Vikas Bansal
Abstract: An optimization technique is applied in dynamic balancing of planner mechanism in which gen etic algorithm is applied. By using genetic algorithm the driving torque, shaking force and shaking moment are minimized. An equivalent (equimomental) system (point masses) are developed which is dynamic equivalent to another system, where some sets of forces and moments can produce the same linear velocity, angular velocity, angular acceleration or linear acceleration. Point masses are represented by the shaking force and shaking moment, choose as design variables. Design variables of an optimization problem are changed into single objective function and apply genetic algorithm. Best results obtained through Genetic algorithm than conventional algorithm. The standard problem of four bar mechanism shows its effectiveness.
Keywords: Optimization; Equimomental system; Dynamic balancing; Shaking force; Shaking moment; Genetic algorithm.
Prediction of Diabetic Retinopathy based on a Committee of Random Forests
by Hedieh Sajedi
Abstract: Diabetic retinopathy is an ocular disease generated by complications of diabetes, and it must be discovered quickly for effective cure. By early diagnosis of retinal fundus disease, ophthalmologists can cure the disease or reduce its deterioration, thereby preventing the patients from vision loss. Using enlarged images, ophthalmologists can diagnose diabetic retinopathy. In this paper, Committee of Random Forests (CRF) for detection of diabetic retinopathy is proposed. In this approach, we use k-means clustering algorithm and random forest classification method to create a new classifier. CRF has been tested on the Diabetic retinopathy Debrecen dataset, in which 94.76% accuracy is reached in a disease or no disease setting.
Keywords: Diabetic retinopathy; Machine learning; Hybrid method; K-means clustering; Random forest classification.
Development of Machine Vision System for Automated Mechanical Objects Recognition
by Tushar Jain, Meenu , H.K. Sardana
Abstract: Object recognition is a type of pattern recognition. Object recognition is widely used in the manufacturing industry for the purpose of inspection. Machine vision techniques are being applied in areas ranging from medical imaging to remote sensing, industrial inspection to document processing, and nanotechnology to multimedia databases. In this paper, mechanical objects recognition used in manufacturing industry is taken into consideration. Mechanically manufactured parts have recognition difficulties due to manufacturing process including machine malfunctioning, tool wear, and variations in raw material. This paper considers the problem of recognizing and classifying of such objects. These objects are kept in different orientations for invariant rotation, translation and scaling. The features are extracted using Fourier descriptor technique. Artificial Neural Network (ANN) with back-propagation learning algorithm is used to train the network and for classification of five different objects. This paper shows the effect of learning rate and momentum on the classification accuracy of objects.
Keywords: Mechanical objects; Machine vision system; Image processing; Fourier descriptor; Object recognition; Artificial neural network; and Learning rate and momentum.
Development of a Low-Cost Robotic Pan Flute
by Praneel Chand, Kishen Kumar, Kishan Kumar
Abstract: Music is a popular form of entertainment for human beings. Instruments are often played by experts in studios or at live performances. Digital or digitised (computer) music is also popular amongst todays listeners. However, there are subtle differences between the sound of a real instrument and digital music. Moreover, the physical movement of real instruments creates a different atmosphere. Hence, this paper addresses the development of a robotic pan flute that enables a real instrument to be played by non-expert musicians. The various sub-systems of the low-cost prototype solution are detailed and its performance is evaluated. Experimental results indicate that the system can produce the desired (theoretical) music notes and is robust to variations in compressed air pressure over the tested range.
Keywords: Music; automation; real instruments; pan flute; low-cost; microcontroller; LabVIEW.
Conversation Model between Human and Machine Using Machine Learning
by Manasi Khanaj
Abstract: Human and machine can interact with each other, in which chatbot plays an important role. The machine can identify the sentences of human and can able to make a decision by its own to respond to question asked by person. The response approach can be used as matching the sentence spoken by user. From sentence that is given as input, machine will be able to get the sentences that are similar enough, and higher matched points gained the more reference to similar sentences. The bigram technology can be used for sentence resemblance calculation. The chatboat trained data are saved in the database. The core and interface are major contents of chatboat and can be accessible in database. The database is used to store knowledge storage of machine and interpreter employs stored programs of procedures and functions and used for pattern-matching through trained dataset.
Keywords: — chatbot; sentence; pattern-matching; trained dataset; Query Processing; Keyword Matching.
Robotic Simulation of Human Brain Using Convolutional Deep Belief Networks
by P.S.Jagadeesh Kumar, Yanmin Yuan, Yang Yung, Mingmin Pan, Wenli Hu
Abstract: Collective endeavours in the fields of computational neuroscience, software engineering, and biology permitted outlining naturally sensible models of the human brain in light of convolutional deep belief networks. While satisfactory devices exist to mimic either complex neural systems or their surroundings, there is so far no mechanism that permits to productively setting up a correspondence amongst brain and its mathematical model. Deep robotics is another stage that intends to fill this gap by offering researchers and innovation engineers in distinguishing human brain diseases by enabling them to associate human brain models to itemized re-enactments of automated programming. In this manuscript, deep robotics utilizing convolutional deep belief networks were exploited to recreate human brain in distinguishing brain diseases. Prediction accuracy of the three noteworthy ideal models, for example, Artificial Neural Networks, Machine Learning and Deep Learning were looked at in distinguishing brain related diseases, such as, Alzheimer's disease and Parkinson's sickness. Customary on the numerical analysis, convolutional deep belief networks outclassed neural back-propagation networks and convolutional neural networks in estimating Alzheimer's disease and Parkinson's sickness.
Keywords: Alzheimer’s Disease; Convolutional Deep Belief Networks; Deep Robotics; Human Brain Simulation; Neuroscience; Parkinson’s Sickness.
CONTACT POINTS DETERMINATION AND VALIDATION FOR GRASPING OF DIFFERENT OBJECTS BY A FOUR-FINGER ROBOTIC HAND
by Eram Neha, Mohd. Suhaib, Sudipto Mukherjee
Abstract: Grasping and manipulation of objects with the multi-fingered robotic hand is required in order to replace human hands in performing various tasks. It is desirable to analyse a robotic hand in comparison with the human hand in terms of stability and various grasp properties. Therefore, in order to attain a stable grasp, the contact points and the grip configuration must be selected in accordance with the grasp stability. In this paper, grasp analysis of the four-fingered tendon actuated robotic hand is carried out. The hand is simulated to grasp different objects in order to determine the contact points of the fingertips on the surface of these objects. Simulation of the hand is done to check the grasp capabilities prior to experimentation while grasping different objects and obtain the contact points for the same. MATLAB SimMechanics tool is used to perform the 3D visualization of the robotic hand where the fingers are controlled in order to grasp objects at the contact points. The obtained contact points are validated using the kinematics and geometric collision detection. These contact points are further utilized to determine the amount of weight required by the tendon to produce the flexion motion in order to grasp the object at the contact points.
Keywords: Contact Points; Tip Prehension; Flexion motion; tendon actuated; SimMechanics.