Knowledge Connect @ SASTRA
Not a member yet
201 research outputs found
Sort by
Digital Technology Usage Behaviour Among Unorganised Retailers In Tamil Nadu
No full textUnorganised retail, consisting of small local shops offering various products, holds a significant market share of 81.50 %, according to the India Retailer Report (2023). Despite the burgeoning e-commerce and organised retail sectors, unorganised retail remains a preferred shopping choice for many consumers.
In response to the sector\u27s potential, several startups have emerged, aiming to support kirana stores. While these startups have primarily focused on horizontal solutions to enhance efficiency, there is a growing need for tailored solutions addressing the specific requirements of unorganised retailers.
Introducing the Dukantech (Dukan in hindi meaning ‘Shop’) model as a solution, this approach enables local sellers to digitize their shops, thereby improving customer service and operational effectiveness
Deep Learning Technique for the Classification of Stress among the Students Using Physiological Biomarkers with a Hybrid Feature Approach
Get PDFAdolescence is a crucial part in life, and the presence of stress, anxiety, depression, and health issues during this stage is a great concern. This research aims to analyze and predict the cognitive stress in students during the examination period using EEG biomarkers. In this study, raw EEG data is acquired under two different experimental conditions, before and after examination, from 14 subjects with an eight-channel Enobio device. After preprocessing of the EEG signal, the brain rhythms such as theta, alpha, and beta sub-band energies and EEG band ratios such as neural activity, heart rate, arousal index, vigilance index and cognitive performance attentional resource index (CPARI) extracted for before and after examination conditions using db4 as mother wavelet with six level decomposition. The extracted features are analyzed before and after exams, as well as gender-wise categories, using SPSS software.
Machine learning techniques are introduced to classify two states as stress and non-stress states. The raw EEG signals are acquired from 25 subjects under two conditions, each with a 3-minute duration: non-stress (relax mode) and stress (mental task), using an EEG device. A total of eleven EEG features were extracted using the discrete wavelet transform technique. The evaluation metrics are compared for different classifier algorithms. The validation of proposed model is carried out using benchmark data from the physionet database.
Further, to improve the performance metrics of classifier algorithms, the proposed model includes hybrid features (both time-frequency and time domain features). EEG data are collected from 110 students under two conditions: relaxing and performing the mental task, each lasting 5 minutes. After pre-processing, the wavelet-based time-frequency features, such as three relative sub-band energies and eight EEG band ratios, are computed. Also, five-time domain-based statistical features such as mean, root mean square, standard deviation, skewness, and peak-to-peak value are considered.
A comparison study is carried out to analyze the performance of the classification model for wavelet-based features alone (time-frequency) and hybrid features (both time-frequency and time-domain features). The result shows that better performance metrics are obtained for the cubic SVM classifier using hybrid features. The validation of the proposed model is carried out with a benchmark database from the physionet, which consists of 36 subjects’ EEG data under relax and task state. Also, we study the importance of fixed and sliding window concepts; the result reports better performance metrics for the sliding window approach using hybrid features.
A deep learning technique is introduced for the classification of stress and non-stress (relax) states using hybrid features with a 2-second sliding window approach. The CNN-BLSTM model reported the highest accuracy for the classification of stress and non-stress states. Finally, an experimental study for the multiclass classification of stress levels such as relaxed, low, medium, and high-stress is carried out by acquiring EEG signals from 35 subjects by performing different levels of task, each with a 5-minute duration. A sliding window approach of 2 seconds with 50% overlap is used for feature extraction. Hybrid features are considered and fed to CNN and BLSTM for classification, and better performance metrics are obtained for the proposed model. In a comparison of both machine learning and deep learning approaches for the classification of stress and non-stress states, better performance metrics are reported for the deep learning method. This proposed model could be used to support the health care providers in early diagnosis of stress among students to avoid suicidal thoughts and provide necessary treatment or counselling in advance
An Application of Soren Kierkegaard\u27s Three Stages of Life to the Marginalised Transgender Persons in India
Get PDFAccording to Soren Kierkegaard, “Life is not a problem to be solved, but a reality to be experienced”. Human beings are completely free to make choices in their life and live in accordance to those. However, they have to be responsible as well as committed to their choices for an authentic existence. Transgender persons are also human beings though they identify themselves different from the gender binary.
This research has proposed to explore the existential struggles and faith of transgender persons in the modern society with reference to the application of three stages of life of Soren Kierkegaard to the Indian English novels namely I am also a Human: The Inner Voice of Soul by Jasbir Singh and Raasathi: The Other Side of a Transgender by Sasindran Kallinkeel.
In addition to this, the research has also extended to an empirical study conducted in the selected districts of Tamil Nadu namely Thanjavur and Trichy for exploring the practical challenges of transgender persons in making their existence. Thus, this research investigates the existential strife of transgender persons as represented both in the literary works and in the society to justify the idea that transgender persons are also human beings and that literature serves its purpose rightly
Techno-Economic Analysis of Solar PV Systems Under Various Environmental Conditions
Get PDFWorld Energy Council has predicted that the global electricity demand will peak in 2030. The need for renewable energy has grown rapidly due to the rising energy demand in this ever-growing population. Green house gas emissions are predominantly due to the energy generation and consumption of fossil fuels. On the other hand, fossil fuels supply nearly 74% of India\u27s energy demand, making India one of the world\u27s largest coal consumers.
According to World Resource Institute, India ranks fourth in the world in terms of carbon emissions. Investments in renewable energy sources and energy-saving technologies should be part of a solution to lower these emissions. Amongst sustainable energy sources, energy from the sun can be considered one of the vital sustainable solutions for meeting the energy demand. Solar energy conversion technologies can be classified into two primary sources: solar thermal and solar photovoltaic (PV). This work mainly focused on PV systems.
Variations in atmospheric conditions, such as solar irradiance and temperature, significantly impact solar PV output power. Therefore, it\u27s crucial to understand how temperature and solar radiation affect PV panels. However, the data sheets provided by module manufacturing firms only reveal the electrical properties of the PV module under the standard test conditions (STC) of 1000 W/m2 solar radiation level, 25oC cell temperature, and 1.5 air mass rate.
Therefore, the electrical characteristics of the PV module that differ from the STC are unknown. Hence, there is a need to study the various parameters affecting the solar PV module\u27s solar irradiance and temperature and how they affect the system\u27s economics. This research has been focused on the techno-economic performance of solar PV.
The first phase of the study presents the economic analysis of the cleaning frequency of solar PV. By performing an experimental investigation on a solar photovoltaic module under clean, dust, and shadow conditions, the reduced electrical power produced was estimated compared to the clean panel. From the results, it is clear that there is a substantial decrease in output power (40 % in the case of dust panels and 80 % in the case of shadow panels) compared to that of the cleaned solar panel. The effect of dust can be mitigated only by cleaning the PV module, whereas the proper placement of PV modules can avoid the effect of shadow. The frequency of cleaning has to be reduced as much as possible so that the process of cleaning the PV module can be done at a cost-effective rate.
Measuring the dust accumulating every time on the panel is challenging in deciding the cleaning frequency. In order to overcome this challenge, the experiment is carried out with three different dust levels in order to establish a correlation between the output power and the dust level. Using this correlation, the frequency of cleaning has been determined for different dust levels.
Solar photovoltaic system efficiency depends on the wavelength of the solar radiation. The second phase of the work consisted of an experimental investigation utilising filter papers of different colours to assess their economic benefit. Five different colour filters have been used, and their corresponding temperatures and output powers have been measured. The economic gains of the PV system with and without colour filters have been evaluated with an initial degradation rate of 0.5% per year and 0.8% per year to demonstrate the improvement in output power under the influence of colour filters.
The third phase of the study presents the techno-economic analysis of numerous passive cooling strategies to determine the best passive cooling approach for PV modules. The analysis deals with various cooling techniques such as coir pith material, Phase Change Material (PCM), and greenhouse net cooling. Hybrid Optimization of Multiple Energy Resources (HOMER) Pro software has been utilized to compare the economics of the system with passive cooling techniques under study to show the efficacy of the cooling techniques
Enhancing Security and Privacy for Smarter Environment through a robust Cyber-Physical System Framework
Get PDFAs digital computing paradigm and practices have emerged in disciplines, devices with processors and sensors were rudimentary, performing independent tasks with limited power. The first computer processors were slow, bulky and consuming high energy, as sensors in thermometers and pressure gauges provide original, independent measurements without effective communication, 1999. It often required powerful, energy-efficient processors and advanced sensors to enable seamless communication and sophisticated data processing.
These devices, since smart home systems to industrial automation tools, which continuously collect, analyse and share data via the internet, facilitating if real-time management, predictive maintenance, and improved seamless experience are used, transforming everyday objects into intelligent, connected systems but IoT devices face greater security risks and attacks. These vulnerabilities are often caused by weak encryption, inadequate authentication methods and irregular software updates, which put critical data and critical systems at potential risk.
The proposed mechanism aims to provide a secure authentication mechanism between IoT nodes and gateways utilizing lightweight authentication techniques and provides a robust mechanism for secured key generation, and key exchange to increase the performance and confirm the integrity of communication with less time consumption and computation costs compared to other existing approaches.
Lightweight authentication protocols enable efficient data exchange among resourceconstrained devices, while robust authentication mechanisms at the gateway level ensures the protection of sensitive information and critical network operations. Hence the proposed system employs an HG-based hashing technique, which offers enhanced resistance to attacks during the key-sharing process. This multi-dimensional approach improves the robustness and security of the authentication process, ensuring a higher level of protection for IoT networks.
Furthermore, an important feature of the proposed scheme is that it can generate a unique random number using a chaotic pseudorandom number generator (PRNG) for each authentication instance, thus improving security by ensuring that each assembly is unique and confusing, ensuring the same range of safe types. The proposed system is designed to operate with minimal computing and communication costs, and maintain low storage requirements, which are critical for resource-constrained environments, and makes the system particularly suitable for IoT and CPS applications Comparative analysis with existing approaches shows the proposed System reachability security is robust, resulting in dynamic security assets that can adapt to multiple threats and operating conditions occur.
The performance of the proposed system is evaluated on different hardware platforms, where real-time performance is critical. The design of the system ensures that it can meet the demands of real-time applications, delivering better results than existing solutions. This improvement is attributed to the resource efficiency of the system and its robust security system, which does not compromise performance. Furthermore, chaos-based PRNG-based random number generation enhances security by eliminating guesswork in the authentication process. This feature is especially important in environments where devices are constantly communicating and exchanging critical information to withstand covert inspection and inspection attacks i.e., probe attacks.
In summary, the proposed system enhances machine communication performance and security through a sophisticated authentication scheme with efficient resource management It effectively addresses the identity-based authentication issues and reveals which good results have been shown in real-time execution, adapted to practical situations compared to existing methods. The integrated storage cost significantly improves the effectiveness of the proposed scheme for secure and efficient machine communication in modern technology
Performance Analysis and Evaluation of Image Processing Techniques, DCNN and BSN model for Detection and Classification of Diabetic Retinopathy
Get PDFDiabetes is a disorder that arises when blood sugar level increases. Insufficient secretion of insulin hormone is the ground for the evolution of diabetes, and it affects most of the critical organs in our body. Diabetes causes leakage of blood in the retinal blood vessels, inducing an eye disease, Diabetic Retinopathy (DR). The visual identification of micro features in fundus images makes the clinicians complex and challenging tasks. Therefore, timely diagnosis and proper treatment can prevent eye blindness. This research is proposed to identify and classify diabetic retinopathy using an Image Processing technique, Deep Convolutional Neural Network (DCNN), Biological Sensor Network Model (BSN), and the integrated method. Finally, the performances of all approaches are compared.
In the image processing technique, the workflow pre-processes the fundus images. Pre-processing comprises green channel extraction, wiener filtering, and contrast enhancement. Subsequently, the morphological operation is performed on the pre-processed image to extract retinal blood vessels. Graph cut image segmentation is performed to the resultant image to detect lesions, namely microaneurysms, hemorrhages, and exudates. From the segmented image, various features like statistical (Mean, Variance, and Standard Deviation), texture (GLDM), and Histogram of Oriented Gradients (HOG) features are extracted. Further, Cascaded Rotation Forest (CRF) classifier is trained with the extracted features. During the testing stage, the classifier detects the lesion and the severity level of diabetic retinopathy. Accuracy, sensitivity, and specificity are 98.00%, 96.00%, and 98.66%.
A deep convolutional neural network consists of many layers to facilitate extracting features and classifying fundus images into normal images, early-stage DR images (Non-Proliferative Diabetic Retinopathy), and advanced-stage DR images (Proliferative Diabetic Retinopathy). Furthermore, it classifies NPDR according to microaneurysms, hemorrhages, cotton wool spots, and exudates, and the presence of new blood vessels indicates PDR. The accuracy, sensitivity, and specificity of this approach are 98.88%, 96.66%, and 99.32%, respectively. The biological sensor model uses Electroretinogram signals (ERG) and the signal processing algorithm to detect diabetic retinopathy.
The workflow consists of pre-processing the recorded one-dimensional electroretinogram signals to remove noises. Fast Fourier Transform (FFT) extracts spectral information of signals. Then, Mel Frequency Cepstral Coefficients (MFCC) features are extorted. These features train the Support Vector Machine (SVM) classifier. During the testing stage, extraction of MFCC features is done for signal under testing. The classifier will predict and classify the input ERG signal from the extracted features into normal and diabetic retinopathy. The accuracy, sensitivity, and specificity of this approach are 98.50%, 97.00%, and 100%, respectively.
A novel integrated system using DCNN and BSN model is proposed to identify and classify diabetic retinopathy. A comparison of all the methods showed that the integrated method achieved a high accuracy rate of 99.33%, a sensitivity of about 98.80%, and a specificity of about 100%. Using fundus image and the Electroretinogram signal of a patient, it is possible to increase system effectiveness in detecting and classifying diabetic retinopathy
Mobile Robot Path Planning Optimization Problem Using Multi- Objective Genetic Algorithm
No full textMobile Robot Path Planning problem (MRPPP) is the most prominent research area employed in different real-time environments. The research domain of robotics offers abundant opportunities for researchers in various engineering fields with different dimensions. The automation of physical movements of the robots with intelligence to make dynamic decisions for interacting with the environment opens up a lot of challenges to the research community.
A variety of approaches are employed to solve the Mobile Robot Path Planning Problem (MRPP), which is to derive a feasible collision-free path to reach the destination from the given starting point by avoiding obstacles. The solution for the MRPP is not only influenced by the physical design of the robot but also by the environment it explores. In the field of mobile robot path planning, a lot of research has been done on both static and dynamic working environments. The determination of an optimal path from the start to the target point, while avoiding collisions in mobile robot path planning with the obstacles in an environment is an NP-hard problem.
Various traditional approaches for solving the path planning problem have been developed using deterministic and non-deterministic approaches. The optimal path, if it exists, is guaranteed by deterministic techniques. However, as the complexity of the working environment rises, so does the time complexity for the techniques. On the other hand, non-deterministic approaches are developed to produce optimal paths with constraints. The principles of the Genetic Algorithm, with many modifications and hybridization with other algorithms, are frequently used to select an optimal path for robot navigation. The major challenges are identified and addressed to explore the solution.
a. The GA is a well-accepted evolutionary algorithm for the MRPP, which requires a large initial population to predict the quality of the path. However, because of the computational complexity, the performance is reduced. In the Environment Specific Strategy(ESS), the performance is improved by reducing computational cost which is achieved by determining the initial population size. Hence the proposed methodology is adapted to address this issue by prescribing three strategies decided based on the density of the obstacles present in the environment.
b. The specific goals of the research work is not only the length of the path is considered as the prime objective as commonly suggested objective, but the smoothness of the path and safety of the path is also considered to improve the quality of the path. This objective is achieved by hybridizing APF and MOGA, exploiting the advantage of both methods and overcoming the drawbacks. In addition to that the derived path is smoothened by applying the three-phase technique.
c. When in the multi-objective case, it is very difficult to customize the level of preferences for the different objectives based on the demand for the environment. The GA technique with a single primary objective is used as the preprocessing technique to apply Fuzzy TOPSIS. Decision makers state the level of preference by linguistic variables rather than absolute values when using the Fuzzy TOPSIS approach.
d. The MOGA approach is applied to determine the optimal path is very complicated for three-dimensional environments. The proposed 3DLAP-MOGA is accomplished by constructing an occupancy matrix for the 3D space and employing MOGA. The methodology 3DLAP-4OGA is additionally considering vertical movement as one of the objectives which give more quality optimal path.
e. The MRPS-MOGA is another methodology proposed where the objectives are added to consider the distance from the obstacles while traversing, the number of obstacles that interacted with the robot and the traveling time in addition to the distance traveling.
So the objectives for the research work are consolidated to address the issues in the MRPP problem as reducing the computational cost, improving the quality of the optimal path, provision to customize the preference levels of objectives and extending the MOGA methodology to 3D space. The suggested ESS has undergone two stages of analysis and proof in order to both quantitatively and qualitatively support the results. In the case of quantitative analysis, the size of the population is compared for the various environments and analyzed which directly influences the computational cost. In qualitative analysis, the length of the path is considered as a parameter to be compared to determine the quality of the paths obtained by the different strategies.
In the proposed DPA, the average of objective values of paths produced by the proposed algorithm is compared with well-known algorithms such as A*, Dijkstra, EGA(enhanced GA) and APF for 8 distinct maps. The findings indicate that the suggested hybrid algorithm is delivering superior outcomes at the level of the individual objectives. The fuzzy-based TOPSIS method was implemented and analyzed with different combinations of the preference level of DM. Then for the quality confirmation of the proposed method, best and worst cases are analyzed for the different pref levels. In addition to that the computational advantage of the proposed methodology is proved by comparing the computational score of GA, Fuzzy TOPSIS and hybridized algorithm.
In the 3DLAP-MOGA technique, Cumulative Normalized Performance Score (CNPS) and Cumulative Normalized Environment Score (CNES) are two scales that are used to measure performance with respect to environmental parameters. To justify the above results another metric is called assessment point. In the extension of this method, 3DLAP-4OGA, the results are analyzed using z-score estimation for standardization which indicates the deviation from the mean. The greater performance was illustrated by the MRPS-MOGA approach from the results implementing the method. The different parameters are analyzed to prove the performance of the methodology. The experimental results of these proposed methodologies revealed that the above approaches are useful for finding an optimal feasible collision-free path for the given environment
A Spatial Data Framework for Indoor Positioning using Machine Learning Techniques
No full textThe last few years have seen an increase in interest in indoor positioning and localization as potential research and development areas. WiFi is a strong substitute that supports positioning based on indoor floor plans. In this thesis, the Principal Featured - Kohonen Deep Structure (PF-KDS) model is developed to position WiFi devices more accurately and efficiently for indoor floor planning. Initially, spatial data analysis is conducted using the Principal Feature Enhanced Auto-Encoder algorithm, extracting principal features for dimensionality reduction.
Following this, the Kohonen Self- Organizing Deep Structured Learning technique is devised for precise position estimation by considering a new path loss model incorporating wall influences on Received Signal Strength Indication (RSSI), thereby enhancing device positioning accuracy. The second aspect introduces the Gaussian Distributive Feature Embedding-based Deep Recurrent Perceptive Neural Learning (GDFE-DRPNL) for improved position estimation with reduced errors.
GDFE-DRPNL selects primary features using Gaussian Distributive Feature Embedding and employs Deep Recurrent Multilayer Perceptive Neural Learning with a Deming Regressive Trilateral Positioning Model to compute device positions, resulting in enhanced accuracy. The third research introduces the Linear Geometric Projective Convolutional Deep Belief Network (LGPCDBN) to enhance position estimation accuracy with minimal errors. Dimensionality reduction is initiated through Linear Helmert–Wolf blocked Sammon projection, followed by Geometric Levenberg–Marquardt Convolutional Deep Belief Network for precise device positioning using geometric triangulation methods.
Additionally, the Enhanced River-Formation Dynamics for Multi-node Routing Protocol (ERFD-MRP) is proposed for efficient sensor data collection in Wireless Sensor Networks (WSN) based on River Formation-based Dynamics, ensuring energy-efficient multi-hop routing. Performance evaluations of the proposed methods are conducted using Java language implementation and Indoor Positioning and Indoor Navigation (IPIN) 2016 competition dataset, demonstrating LGPCDBN\u27s superior positional accuracy and ERFD-MRP\u27s higher throughput, packet delivery ratio, lower delay, and energy consumption compared to existing approaches
Exploring the role of RB1 in cellular differentiation using patient-derived stem cells
No full textThe Retinoblastoma (RB1) gene, predominantly recognized as a tumour suppressor, plays a crucial role in early growth and development. Mice lacking the retinoblastoma gene do not survive and exhibit defects in the differentiation of various tissues, underscoring its involvement in cellular differentiation. This suggests that the functions of the RB1 gene extend beyond its role as a tumour suppressor, encompassing cell cycle regulation, maintenance of genomic stability, cellular senescence, and cellular differentiation.
While the tumour suppressor function of the RB1 has been extensively studied, its role in cellular differentiation remains incompletely understood. Existing literature, primarily based on animal models, has provided inconclusive results on RB1’s role in differentiation, highlighting the need for human studies. This study aimed to investigate the role of RB1 in cellular differentiation using patient-derived stem cells from retinoblastoma patients.
Mesenchymal stem cells (MSCs) isolated from the orbital adipose tissue of the retinoblastoma patients were cultured and characterised as Orbital Adipose Mesenchymal Stem cells (OAMSCs). The RB1+/- OAMSCs were differentiated into osteogenic and adipogenic lineages to investigate the role of RB1. Furthermore, to study the role of RB1 retinal differentiation, OAMSCs were reprogrammed to induced pluripotent stem cells (iPSCs) and then subjected to stepwise retinogenesis. The study revealed that the xvi
monoallelic loss of RB1 does not alter the mesenchymal phenotype of the MSCs, but impacts its proliferation and differentiation. In adipogenesis, the monoallelic loss of RB1 accelerated and increased the differentiation, with a bias towards brown adipocytes; in osteogenesis, it increased differentiation and proliferation but maintained genomic integrity, possibly preventing or delaying tumourigenesis.
However, in retinogenesis, RB1 monoallelic loss did not impact proliferation or differentiation but influenced metabolism, as evidenced by enhanced glycolytic activity and ATP production. This metabolic adaptation likely meets altered energy demands and cellular requirements in RB1+/- retinal organoids.
These findings highlight that RB1 exerts different roles in distinct cell types. While promoting proliferation and differentiation in adipogenic and osteogenic lineages, RB1\u27s monoallelic loss in retinal cells primarily influences metabolic processes rather than differentiation. This underscores the multifaceted role of RB1 in cellular differentiation and emphasizes the importance of cell-type-specific studies to fully elucidate RB1’s diverse functions
ZnO Based Chemiresistive Sensor For Ammonia Detection
No full textThis work aims on the development of nanostructured sensing elements based on ZnO for detecting ammonia, specifically designed for applications in poultry and farm fields. To achieve this objective, we prepared sensing elements, including ZnO, CuO, WO3, NiO, ZnO/CuO, ZnO/WO3, ZnO/NiO, and rGO/ZnO, using the spray pyrolysis technique. Structural, morphological, and compositional analyses were studied using X-ray Diffraction (XRD), Field Emission Microscope (FESEM), and X-ray Photoelectron Spectroscopy (XPS), respectively. The vapour/gas detection studies of fabricated sensing elements were carried out in a home built sensing chamber.
The single oxides, ZnO, CuO, WO3, demonstrated ammonia detection in the range of 5-100 ppm at 350 °C. Mixed oxides, ZnO/NiO, ZnO/CuO, ZnO/WO3, and rGO functionalized ZnO, exhibited an enhanced sensitivity of 0.5 ppm at 300 °C. This improved sensitivity was attributed due to the formation of p-n junctions and increased oxygen vacancies in the materials. The sensors were further tested in a mixed gas environment with major poultry and farm field gases such as methane (500 ppm), carbon dioxide (5000 ppm) and nitrous oxide (25 ppm). The concentrations of these gases were kept at their Permissible Exposure Limit (PEL) and Recommended Exposure Limit (REL). The ammonia concentration was maintained below the permissible limit at 20 ppm.
To address output voltage drift in the sensors, pattern recognition techniques were employed, classifying the concentration of gases into four levels: low range of ammonia (0.5 ppm to 8 ppm), medium range of ammonia (greater than 8 ppm and less than 25 ppm), high range of ammonia (25 ppm to 100 ppm), and other gases. The features extracted from two group of sensors was tested using Random Forest (RF), Support Vector Machine (SVM), and Naive Bayes algorithms, with RF demonstrating the highest classification accuracy of the dataset. Consequently, a sensor array was fabricated using ZnO, CuO, WO3, ZnO/CuO, and ZnO/WO3, achieving an RF accuracy of 96.8%.
Two prototypes of ammonia detectors were developed. The ZnO-based detector exhibited an ammonia detection range of 5-100 ppm at 350 °C with continuous monitoring. The second prototype, based on a sensor array utilizing the RF classifier, displayed different concentration of ammonia levels through LED indication. The findings of the proposed work focus on ammonia detection below trace levels by enhancing the base material composition and subsequently post processing the signals by pattern recognition technics. In this way a systematic approach to develop an effective and selective real time ammonia detectors for poultry and farm field applications, combining ZnO based sensing elements and pattern recognition algorithms for accurate and reliable detection can be accomplished