Puneet Kumar

I am a recent B.Tech graduate in Electronics and Communication Engineering from VIT University and currently pursuing a Master’s in Communication and Signal Processing at Arizona State University. I am seeking opportunities to apply and expand my expertise in this domain while contributing to innovative projects and advancing my professional skills.

• Date of Birth: 01/24/03
• Nationality: Indian
• Gender: Male

Channel Estimator for OFDM Systems with insufficient redundancy using AIML Techniques.

A channel estimator for OFDM systems with insufficient redundancy uses AI/ML techniques to enhance signal recovery and reduce error rates. Machine learning algorithms, such as neural networks, analyze complex channel behaviors, compensating for limited redundancy by accurately estimating channel state information (CSI) and improving overall OFDM

communication reliability and efficiency using MATLAB and Python.

(https://drive.google.com/file/d/1Ky83XSrhlIZTpk7knVzZdrKsrZ1X60wx/view?usp=share_link)

Multichannel DUC DDC IP Design for 5G system

Built a multichannel Digital Upconverter (DUC) and Digital Downconverter (DDC) IP design is essential for 5G systems, enabling efficient signal processing across multiple frequency bands. This design facilitates data conversion and channelization, enhancing flexibility, bandwidth, and performance in 5G infrastructure by supporting simultaneous multichannel transmission and reception using MATLAB

and VIVADO.

(https://drive.google.com/file/d/1cB6MmLKLS3EOAYIHrXPwb44ZVFQTny0P/view?usp=share_link)

Binary Classification of Forest–Non Forest Area using Satellite Imaging and Coordinate Mapping

Developed a simple machine learning model to classify land cover into forest and non-forest regions using satellite imagery and geospatial coordinates. Performed image preprocessing, feature extraction, and coordinate mapping to align spatial data. Trained and evaluated multiple classification algorithms to achieve high accuracy, enabling efficient monitoring of deforestation and land-use changes for environmental analysis. I have written a paper on this project with the results obtained

(https://drive.google.com/file/d/1BKbcJqfj6ETnE1_ee7bNG5IREZoEPe7W/view?usp=share_link)

IoT-based Water Monitoring System using Neural Networks

Designed and implemented a smart water quality monitoring solution integrating IoT sensors to measure parameters such as pH, turbidity, and temperature in real time. Processed collected sensor data using a neural network model for anomaly detection and predictive analysis. Achieved reliable classification of water quality levels, enabling proactive maintenance and resource management for sustainable water usage. I have written a paper on this project with the results obtained

(https://drive.google.com/file/d/1uNbt4A-o6G2UMYvacgoRkmUkxHhD2rGs/view?usp=share_link)

Temperature and Humidity Controlled Fan using Arduino Board

Built a temperature and humidity controlled fan using an Arduino board that automatically adjusts its speed based on environmental conditions. With sensors like DHT11 or DHT22, it measures temperature and humidity levels, activating or adjusting the fan as needed. This setup improves energy efficiency and provides a comfortable environment in various applications using Arduino UNO.

(https://drive.google.com/file/d/1tp2OoLaeiM87RivFkz6b96mH3wb6GPNP/view?usp=share_link)

IoT-based Smart Precision Agriculture using Drones

Developed a precision farming solution integrating IoT sensors and drone technology to monitor crop health, soil conditions, and environmental parameters. Utilised drone-based imaging for high-resolution field mapping and applied neural networks and data analytics to optimise irrigation, fertilisation, and pest control. Enabled data-driven decision-making to improve crop yield, reduce resource wastage, and enhance sustainability in agricultural practices.

(https://drive.google.com/file/d/1D_Euebn1USchgwuefgAOo5nAyyN4fKf1/view?usp=share_link)

Accurate Indoor Positioning Using Wi-Fi Signals and Deep Learning Techniques

Designed and implemented a comparison between 5 deep learning–based systems to estimate precise indoor locations using Wi-Fi RSSI data. Collected and preprocessed signal datasets, applied models including Transformers, CNNs, ANN, GAN and Autoencoders, and optimised hyperparameters to improve mean absolute error to ~2–3 meters under the Free-Space Path Loss and Log-Normal Shadowing model. Evaluated performance across multiple architectures, demonstrating superior accuracy compared to traditional methods, and ensured the approach remained scalable and device-agnostic for real-world deployment. I have written a paper on this project with the results obtained, and the paper has been published in NMITCON.

(https://drive.google.com/file/d/175zkem3skZ-nHlxBPi0lVw4mZBtmSwa2/view?usp=share_link)