The artificial human olfactory system mimics the human olfactory system using a sensor array and AI algorithms to identify and quantify volatile compounds. This technology has shown promise in various fields, including agriculture, food and beverage, cosmetics, healthcare, and environmental monitoring. However, it remains underdeveloped due to the complexity of odors and the limitations of current sensors in distinguishing specific odors. Odor is a mixture of gases and volatile organic compounds, presenting significant classification challenges due to their low concentrations and complex chemical structures. Our study utilizes multiple sensors to convert gas molecular signals into electrical signals, identifying specific gases and their physicochemical characteristics. This paper introduces a novel study leveraging a unique dataset, comprising 20 distinct olfactory signatures captured with 11 sensors. Additionally, we have developed an artificial olfactory system capable of sensing the concentration of various gases and volatile organic compounds, learning the unknown basis of the olfactory signature based on the received time-series data, and finally classifying it as a smell. Developing such a system involves overcoming challenges related to complex gas interactions, sensor accuracy, and the analysis of the sensor-generated data. We evaluate a range of deep learning models, including 1DCNNs, ResidualCNNs, RNNs with attention, GRU, and LSTM, to classify these signatures across a carefully prepared dataset.