BITS Faculty Publications

Permanent URI for this communityhttp://localhost:4000/handle/123456789/1867

Browse

Has files: NoSubject: search.filters.subject.Artificial Neural Networks

Search Results

Now showing 1 - 10 of 14
  • No Thumbnail Available
    Item
    A Phase-wise Analysis of Machine Learning based Human Activity Recognition using Inertial Sensors
    (IEEE, 2020) Kala, Prateek
    Advancements in sensor technology, expanded analytical skills, and advancement in the field of Machine Learning (ML) and Deep Learning (DL), have all resulted in a substantial increase in popularity and wearable sensor performance for Human Activity Recognition (HAR). The real challenge is to spot the events in an unsupervised environment. In this study, we have tried to build a user-friendly, as well as an effective HAR system using an inertial sensor for eight everyday activities performed. The data was collected by our research team, using a single inertial sensor in a fully unsupervised setup. The eight tasks include: standing, sitting, sleeping, running, walking, cycling, upstairs and downstairs. This paper aims to present a detailed analysis and comparison for three primary aspects of a general HAR which contributes to the overall system performance. This involves analyzing the effects of pre-processing, comparing several extraction and selection methods for generating features from time-series data, and finally building and validating the performance of various classification methods to obtain the best combination of the three. The classification methods included in this study are Logistic Regression, K-Nearest Neighbors, Support Vector Machines and Artificial Neural Networks. After choosing the best parameters and techniques, we achieve a remarkable performance for recognizing the eight activities with an overall accuracy of 93.6%.
  • No Thumbnail Available
    Item
    Bypassing traditional molecular dynamics with artificial neural networks
    (AIP, 2023-05) Aneesh, A.M.
    An attempt has been made to speed up molecular dynamics simulations using machine learning. LAMMPS package was used to generate data for training the ML model which was programmed in PyTorch. The fidelity of the data generated by LAMMPS was first validated by simulating the evaporation of an Argon droplet in its own vapor. Results from simulations were compared with the D2 Law of droplet evaporation and a reasonably good agreement between theory and simulation was observed. Training and testing datasets consisted of per-timestep snapshots from 6 simulations of equilibration of up to 100 atoms in a periodic box. These were converted to images of dimension (10,10,6), such that 100 pixels of dimension (1,1,6) stored the coordinates and velocity components (x,y,z,vx,vy,vz) of up to 100 atoms. A Symplectic Recurrent Convolutional Hamiltonian Neural Network (SRCHNN) was proposed in which a conserved scalar analogous to the Hamiltonian of a system of interacting atoms was modeled using a Convolutional Neural Network. Using Hamilton's equations of motion, time derivatives of positions and velocities were obtained by taking the symplectic gradient of the Hamiltonian, calculated using backpropagation. Symplectic time integration with the Leapfrog algorithm was employed for predicting trajectories using the calculated time derivatives. The model was trained in a recurrent manner with sequences of particles’ positions and velocities. The performance of SRCHNN was tested against the length of the sequence used for training, ranging from 1 to 6. The mean square error (L2 loss) between the true and predicted output states did not decrease significantly with larger training sequence lengths. The percentage error between the predicted and true number of droplet particles was least for the smallest sequence length of 1; while the percentage errors between the droplet and ambient temperatures were roughly the same for all training sequence lengths. The SRCHNN was able to predict 15 future states in sequence within acceptable degrees of accuracy and a 3.1x speedup over LAMMPS was observed.
  • No Thumbnail Available
    Item
    Comparative Study of Estimated Surface Roughness Using GA and PSO Techniques for Milling of Thin-Walled Structures
    (Springer, 2022-04) Bera, T.C.
    Thin-walled structures, due to their lightweight, have found significant applications in the aerospace industry. For the manufacturing of any component, its surface quality index is of prime importance. A very well-known measure of this surface quality is surface roughness. For a product of high quality, the surface roughness value is often desired to be minimum. However, the machining parameters for the production of such surfaces often rely on the engineer's experience and expertise, which always do not lead to the best possible results. In this study, a neural network was first created for surface roughness estimation, then evolutionary algorithms such as Genetic Algorithm and Particle Swarm Optimization were used to minimize the surface roughness value. During this process, the impact of milling parameters such as rake angle, nose radius, and approach angle on the surface roughness value was also studied with the aid of surface plots of surface roughness developed by taking two parameters at a time and holding the third parameter constant.
  • No Thumbnail Available
    Item
    Smart control of electric lamp using artificial intelligence based controller
    (IEEE, 2015) Dasgupta, Mani Sankar
    There is need and scope of controlling the power consumed by conventional electric lamps in presence of some natural light. An artificial intelligence based control system has been developed to control a lamp dimmer circuit with bidirectional triode thyristor. The light present in the room is sensed and voltage supplied to the lamp is controlled by varying the time constant of the circuit through change of resistance of a multi-turn potentiometer with a stepper motor. The resistance set in the lamp dimmer circuit is in accordance with signals from an adaptive Neural Network running in MATLAB ® . The artificial neural network (ANN) runs in real time in MATLAB ® environment to control the time constant of the lamp dimmer circuit for controlling of power consumption. Under test conditions, energy savings up to 35% is achieved.
  • No Thumbnail Available
    Item
    Performance evaluation of a CO2 scroll expander for work recovery using artificial neural network
    (Taylor & Francis, 2017-10) Dasgupta, Mani Sankar
    CO2 trans-critical refrigeration systems operate in sub-critical zone for major part of the up-time even for warm climate regions. Recovery of expansion work from CO2 refrigeration systems is viewed as a workable solution to tide over the challenge of typical low coefficient of performance of such systems. Some of the barriers for wide spread implementation of expanders are; relatively low work recovery and high initial investment. In the present study, the functioning of a scroll work recovery expander under sub-critical condition is investigated in an open-loop setup using CO2 as working fluid. The scroll expander itself is obtained through conversion from a scroll compressor with minimal additional investment. Influence of the various operating parameters like mass flow rate, suction pressure, pressure ratio, and rotational speed on the overall performance of the system are examined. An artificial neural network is then trained in Statistical Package for the Social Sciences (SPSS) platform with part of the experimental data and the same is validated with remaining data. It is observed that, the deviation between the shaft speed and shaft work for the model based prediction and experimental results are within ±7.5% and ±11.1%, respectively. The developed artificial neural network will be useful for predicting performance of work recovery scroll expander in closed loop operation with CO2 refrigeration system in sub-critical zone.
  • No Thumbnail Available
    Item
    A Comparative Analysis of Surface Roughness Prediction Models Using Soft Computing Techniques
    (Springer, 2020-07) Garg, Girish Kant; Sangwan, Kuldip Singh
    Surface roughness is one of the significant index to measure the product quality of the machined parts. The objective of this work is to contribute towards the development of prediction models for surface roughness. In this work, the predictive models were developed for turning operations using soft computing techniques; support vector regression (SVR) and artificial neural network (ANN). The turning experiments are conducted to obtain the experimental data. The developed predictive models were compared using relative error and validated using hypothesis testing. The results indicate that both techniques provide a close relation between the predicted values and the experimental values for surface roughness and are appropriate to predict the surface roughness with significant acceptable accuracy. It is found that ANN performs better as compared to SVR.
  • No Thumbnail Available
    Item
    Predictive Modeling for Power Consumption in Machining Using Artificial Intelligence Techniques
    (Elsevier, 2015) Sangwan, Kuldip Singh; Garg, Girish Kant
    The objective of this work is to highlight the modeling capabilities of artificial intelligence techniques for predicting the power requirements in machining process. The present scenario demands such types of models so that the acceptability of power prediction models can be raised and can be applied in sustainable process planning. This paper presents two artificial intelligence modeling techniques - artificial neural network and support vector regression - used for predicting the power consumed in machining process. In order to investigate the capability of these techniques for predicting the value of power, a real machining experiment is performed. Experiments are designed using Taguchi method so that effect of all the parameters could be studied with minimum possible number of experiments. A L16 (43) 4-level 3-factor Taguchi design is used to elaborate the plan of experiments. The power predicted by both techniques are compared and evaluated against each other and it has been found that ANN slightly performs better as compare to SVR. To check the goodness of models, some representative hypothesis tests t-test to test the means, f-test and Leven's test to test variance are conducted. Results indicate that the models proposed in the research are suitable for predicting the power.
  • No Thumbnail Available
    Item
    Predictive Modelling for Energy Consumption in Machining Using Artificial Neural Network
    (Elsevier, 2015) Sangwan, Kuldip Singh; Sangwan, Kuldip Singh
    The energy efficiency is important evaluation criterion for new investment in machinery and equipment in addition to the classical parameters accuracy, performance, cost and reliability. Even the users in the automotive industry demand new acquisitions of energy consumed by a machine tool during machining. Large interrelated parameters that influence the energy consumption of a machine tool make the development of an appropriate predictive model a very difficult task. In this paper, a real machining experiment is referred to investigate the capability of artificial neural network model for predicting the value of energy consumption. Results indicate that the model proposed in the research is capable of predicting the energy consumption. The present scenario demands such type of models so that the acceptability of prediction models can be raised and can be applied in sustainable process planning during the manufacturing phase of life cycle of a machine tool.
  • No Thumbnail Available
    Item
    Predictive Modelling and Optimization of Machining Parameters to Minimize Surface Roughness using Artificial Neural Network Coupled with Genetic Algorithm
    (Elsevier, 2015) Sangwan, Kuldip Singh; Garg, Girish Kant
    This paper develops a predictive and optimization model by coupling the two artificial intelligence approaches – artificial neural network and genetic algorithm – as an alternative to conventional approaches in predicting the optimal value of machining parameters leading to minimum surface roughness. A real machining experiment has been referred in this study to check the capability of the proposed model for prediction and optimization of surface roughness. The results predicted by the proposed model indicate good agreement between the predicted values and experimental values. The analysis of this study proves that the proposed approach is capable of determining the optimum machining parameters.
  • No Thumbnail Available
    Item
    Optimization of Machining Parameters to Minimize Surface Roughness using Integrated ANN-GA Approach
    (Elsevier, 2015) Sangwan, Kuldip Singh; Garg, Girish Kant
    The surface roughness is a widely used index of product quality in terms of precision fit of mating surfaces, fatigue life improvement, corrosion resistance, aesthetics, etc. Surface roughness also denotes the amount of energy and other resources consumed during machining. This paper presents an approach for determining the optimum machining parameters leading to minimum surface roughness by integrating Artificial Neural Network(ANN) and Genetic Algorithm (GA). To check the capability of the ANN-GA approach for prediction and optimization of surface roughness, a real machining experiment has been referred in this study. A feed forward neural network is developed by collecting the data obtained during the turning of Ti-6Al-4 V titanium alloy. The MATLAB toolbox has been used for training and testing of neural network model. The predicted results using ANN indicate good agreement between the predicted values and experimental values. Further, GA is integrated with neural network model to determine the optimal machining parameters leading to minimum surface roughness. The analysis of this study proves that the ANN-GA approach is capable of predicting the optimum machining parameters.