linkedin.com/in/pawan-vijayanagar/
I’m a Mechanical, Computer, and Electrical Engineer who loves building things that blend electronics, mechanics, and code. I work with embedded C, MATLAB, and Python, and I’m comfortable creating and jumping between hardware like microcontroller–sensor fusion setups, the oscilloscope, a soldering iron, composites, 3D printing, CNC machining and ofc the toaster oven.
I’m big on rapid prototyping and enjoy the process of debugging electronics with digital lab gear. Some of my personal projects include setting a past internet record for the longest-flying camera drone under 0.6 kg, plus a few others you can check out below. Always up for mixing tech, creativity, and a bit of problem-solving fun.
Cheers!
F 2025 - Extended Kalman Filter AHRS
I built an AHRS (Attitude and Heading Reference System) that fuses an IMU, a barometric pressure sensor, and an auxiliary UART feed with GNSS data (Lat, Long, Elevation). The module exposes an SPI interface for output.
On code, CMSIS on C, I've implemented an extended Kalman filter (and SPI ODs (Data Object Dictionaries), device code etc...) that runs update step at 1 kHz: gyroscopes propagate the attitude state based on integration, while the accelerometer (magnetometer, to be added) updates and makes 'corrections' to drift in estimated attitude. Altitude is estimated primarily from the pressure sensor, with optional corrections (to be implemented) from vertical acceleration and GNSS. The STM-based controller performs the floating-point matrix ops fast enough to sustain the 1 kHz update loop. As per the current world politics and for obvious reasons, I'm refraining from posting details or code as these devices could be easily misused.
F 2024 - SAE Electronic Control Unit
S 2023 - Helicopter Flight Controller
I completed a weekend project where I designed and coded a flight controller for a flybarless helicopter. Unlike airplanes, which have inherent aerodynamic stability, helicopters are aerodynamically unstable. As the rotor system creates multiple dynamic forces, such as torque, transient lift, gyroscopic precision, and dissymetry of lift, they require constant, precise control inputs to maintain stable flight.
The controller uses a BNO055 IMU's euler angles and controls the swashplate, as well as its gyroscope to control the tail rotor's yaw rotation rate. I had to integrate the various sensor inputs to achieve accurate and responsive control. It took a lot of experimentation and testing to ensure the controller worked correctly. I'm really proud of the final product, and I believe that the code I wrote could be a valuable resource for others interested in building their own flight assist controllers. This project showcased my technical skills in control systems and my understanding of both the sensors and the physics of helicopter flight.
F 2022 - Machine Learning Paper - VCS
Paper Title: Defence against Voice-Controlled System Adversarial Attacks
Abstract: This document is the final paper for CSE598 Machine Learning Security and Fairness taught by Dr. Chaowei Xiao. The idea behind this paper is to develop a defense against adversarial attacks using Machine Learning techniques on Voice Controlled Systems like and not limited to Google Assistant, Siri, and Amazon Alexa. We used the 8k Audio dataset to create a perturbed dataset using audio attack noise found on https://adversarial-attacks.net/ for VCS to be able to classify non-attack audio from attack audio (perturbed audio) using a programmed ML algorithm using PyTorch.
Link to Paper:
F 2022 - Machine Learning Paper - HMM
Paper Title: Inferring the Topology of Transmembrane Proteins with Hidden Markov Models of Various Complexities
Abstract: Our work revolves around Tamposis et al. 2019 [1]. This paper describes a hidden Markov model (HMM) method to incorporate unlabeled data for predicting amino acid positions in transmembrane proteins. We analyze this method and their prior work in Bagos et al. 2006 [2] to understand how to construct and train HMMs. We then implement a simple, two-state HMM described in Bystroff and Krogh 2008 [3] and connect these results to the two works to predict an Amino Acid position in the TM protein chain. The above graph predicts intersection points where Amino Acids could be found in the unlabelled protein data in question.
S 2021 - LiDAR Mapping Rover
Developed an autonomous robot that is capable of mapping the entire Tempe campus and classrooms using LiDAR technology and a developed mesh network. The above video shows a mapped environment of Sun Devil Football Stadium, ASU. Generated and post-processed by my teammate, Omar Alkathib.
S 2021 - MiniSat
Built a mini-satellite that can track and position itself with a simulated earth's magnetic field from a Helmholtz coil. The system can also simulate star tracking and positioning using light intensity based on directionality. I modeled the system equations and used the MATLAB SISO tool to simulate the system dynamics to find an optimal controller using the root locus method. I implemented this controller TF in discrete time (embedded C) to control the system and follow the simulated magnetic field (demo shown in video).
S 2021 - ESP32 based GoPro Protune Remote (RTOS)
I developed a GoPro remote to combat GoPro's removal of ProTune from the GoPro mobile application and made it open-sourced using an ESP32 that was programmed with wifi, and a basic realtinme to run multiple schedules at once (Oled screen, wifi, menu selection, GoPro setting change). The developed DIY method was appreciated by many fellow FPV drone enthusiasts who used GoPros to record drone footage. The device was also featured on Joshua Bardwell's youtube live stream! Be sure to check out the code linked below
Live Stream: https://youtu.be/Xaq7Zhghdo8?t=505
F 2020 - Remote Sensing of Wind using UAS
I invented a tool as an inspiration from my MEE441 - Wind Energy class at ASU that I took over fall. In this class, we learned aspects of engineering windmill blades using fluid mechanics, developed VADs from Doppler LiDAR data, and wrote BEM code (Blade Element Momentum Method) to analyze the efficiency of windmill blades from scratch using wind data from a doppler LiDAR. The developed tool is a device that calculates live wind data using the attitude (angle) of the drone and sends live wind data at the set altitude to the ground station through a programmed long-range link. The upside of using drone technology compared to a doppler LiDAR would be portability - 1ton vs 250grams and human safety; as the system can be flown to a remote location and collect live wind data from 10km away, not possible with a doppler LiDAR. Take a look at the video and processed MATLAB plots for better understanding.
Paper Title: Remote Sensing of Wind using UAS
LoRa Mesh Network & Cart Tracker v2
 |
| Planned Static Node zones |
 | ||
A hidden Static Node during testing
|
- Production Relay Nodes & Cart Tracker -
 |
| Node Front View |
 |
| Node Right-Rear View |
 |
| Production Node secured to a parking structure light pole to relay IoT Sensor data |
 |
| Cart with deployed Tracker that lasts for 3 months on a single charge and indefinitely on Solar |
Note: LoRa (short for Long Range) is the physical layer of the wireless communication protocol, which is responsible for transmitting data over long distances using low-power radio frequencies. It is a proprietary technology developed by Semtech Corporation and is designed to be used in a wide range of applications, including IoT, smart cities, and industrial automation.
LoRaWAN (short for Long Range Wide Area Network) is the network protocol that sits on top of LoRa. It is an open standard for wireless communication that is designed specifically for IoT devices and networks. As LoRaWAN is open source, its network is typically operated by network providers or independent users, who manage the gateways and provide access to the network for public devices. This allows any LoRaWAN-enabled devices to connect to the internet and communicate with other devices and servers.
The main difference between LoRa and LoRaWAN is that LoRa is a physical layer protocol, while LoRaWAN is a network protocol. LoRa is responsible for transmitting data over long distances using low-power radio frequencies, while LoRaWAN is responsible for managing the communication between devices and gateways, as well as providing internet connectivity.
In the case of the LoRa mesh network that I built and programmed for Arizona State University (ASU), it was built on top of the LoRa technology, not the LoRaWAN protocol. This means that the network is managed and operated by ASU and will only accept incoming data streams of its own IoT devices. This allows for more flexibility and control over the network, as well as the ability to customize it to the specific needs of the university.
ASU IoT Pole Project v4
 |
| MK-I, MK-II, MK-III/MK-IV (Left to Right) |
MK-II, MK-III, and MK-IV were showcased in image #2. I designed and developed MK-IV from the ground up, integrating all sensors and microcontrollers onto a single PCB, creating a standalone system capable of being mounted on poles or installed on ASU building walls. The system utilized either a USB serial interface for pole installations or a proprietary LoRa off-grid mesh networking system, which I developed, to enable communication without direct internet access. The development of MK-IV was completed within two months, but deployment was postponed due to manufacturing and approval delays. For more details about this project, please refer to the section below titled Pole Project MK-I.
S 2020 - 1-Heure v2.0
I redesigned and built a mini version of 1-Heure (v1.0), an efficient 3D-printed quadrotor that was tested to hover for 65 minutes (1 hour and 5 minutes) on a single charge. This design was engineered to withstand extreme weather conditions, including temperatures exceeding 110°F (43°C). Its flight time surpasses that of any consumer drone currently available, combining endurance with resilience for reliable performance in demanding environments.
ASU Cart Project - v1.0
I contributed to the Cart Project at ASU, which helps the university track campus carts in real time for asset tracking and enhanced student interaction. This system relies on GPS data transmitted via LoRa wireless communication protocol, with updates sent to AWS every few seconds. Much like platforms such as Uber and Lyft, this tracking data was integrated into ASU’s Disability Access and Resources Transportation (DART) web app, providing students with real-time cart location updates for improved accessibility and campus interaction. For this project, I developed the code for the trackers, designed the PCB and housing, and conducted real-world testing to ensure defined functionality and range.
Presentation: https://drive.google.com/file/d/1IMTb9X79Dz0iWGf3Pkc-J4ErhABNT8IW/view?usp=sharing
S 2020 - Stabilized Spoon Project
I further developed a robotic, self-stabilizing spoon designed to counteract tremors or shakes. This spoon was engineered to aid individuals living with Parkinson's disease, a central nervous system disorder that affects movement and often causes tremors. The system was built to provide greater control and ease during eating, improving the quality of life for those affected.
ASU IoT Pole Project v1.0
As an IoT Developer for ASU's University Technology Office, I designed the housing, electrical schematics, and PCBs for a project that will soon be deployed across multiple emergency poles throughout ASU campuses. These emergency poles integrate weather sensors, LiDAR, and cameras, all connected to AWS to monitor real-time weather conditions and automatically alert authorities during emergencies.
F 2019 - ASU Devils Invent
Stood 1st place along with 2 teammates at Devil's Invent; an annual ASU engineering event that brings engineering students of all levels of education to compete for the best and most innovative solution to a problem statement by building and pitching a prototype model in less than 24 hours.
https://www.linkedin.com/posts/pawan-vijayanagar_arizonastateuniversity-develop-idea-activity-6586125185525653504-NbYG
S 2019 - 1-Heure v1.0
 |
| Stress Test Conducted to determine Factor of Safety |
I designed, built, and tested an efficient, weatherproof, 3D-printed quadrotor that successfully hovered for 64 minutes (1 hour and 4 minutes) in extreme conditions. Throughout the building and testing phases, I encountered and resolved numerous challenges, particularly due to testing in extreme heat exceeding 110°F (43°C). Despite these conditions, the aircraft performed exceptionally well, demonstrating impressive endurance and validating both my design choices and the durability of the system.
S 2019 Electric IoT Longboard
I designed and built an efficient electric skateboard, powered by a custom-built 0.273 kWh 18650 battery pack. The system included a mobile GPS tracking unit, which operated seamlessly over the university’s Wi-Fi network and was controlled by an ESP8266 board. This electric longboard was capable of achieving a top speed of 31 mph and offered a range of up to 18 miles on a single charge, combining performance with real-time location tracking for added functionality.
F 2017 - Jan 2018
I designed and 3D-printed a modular aircraft using SketchUp, which could be fully assembled or disassembled pretty quickly. While the model had not yet undergone flight testing, it was lightweight yet durable, optimized for endurance missions.
The aircraft featured a Clark Y airfoil, renowned for its balance of lift and low drag, which was integrated into both the wing and fuselage structures. For the V-tail, I selected the NACA 0010 airfoil, chosen for its symmetrical profile and stability characteristics. The precise airfoil coordinates for both configurations were sourced from the UIUC Airfoil Coordinates Database following a thorough aerodynamic performance analysis.
This design approach ensured not only aerodynamic efficiency but also ease of transport and rapid assembly, making it ideal for field deployment and long-duration operations.
S 2017
An old prototype of the spoon from 2017.
F 18th, 2016
 |
| An X-ray view of the designed prototype |
 |
| Laser cut |
 |
| A cage with a robotic clamp was designed for a competition |
Won two consecutive 1st places at the IIT Bombay Tech Fest (Indian Institute of Technology) for the UAV competition, individually defeating over 200 University teams from all over India.
S 2016
 |
| The wind tunnel |
 |
| A KFM-2 airfoil being tested |
Built a wind tunnel to test the aerodynamic characteristics of unique (nonstandard) airfoils and wing designs compared to a standardly shaped airfoil. The results of the experiment were plotted and the graphs were thereafter analyzed. A thesis was then written for my IBDP coursework with the following research question: How does the shape of an airfoil impact the drag, lift, and stall characteristics of an aircraft?
Link to the written paper: https://drive.google.com/file/d/1DJbqDnr-TTcnnZ1HhGI92vfSkFj-d0E5/view?usp=sharing
F 2015
 |
| The autonomous fixed wing |
 |
| Capable of self-flying |
Designed and built a fully autonomous fixed-wing aircraft capable of cruising long-range missions as well as surveying large areas of land quickly, which is helpful for farming and mining purposes.
In the same year, I got a special mention award as well as won the 2nd runners-up at the IIT (Indian Institute of Technology) Tech Fest 2015 <the largest Technology Fest in the whole of Asia> under the UAV competition wherein I individually competed with IIT and engineering college teams from all over the country as well as internationally.
F 2014
 |
| An electric box controller with an embedded Arduino |
Successfully built an Autonomous multirotor for Geo-mapping along with a self-programmed high-tech ground station.
S 2013
 |
| My first aircraft airborne |
 |
| Built by me at a pretty young age |
I was one of the first to publish a DIY article on building a Tricopter, which has now garnered over 100,000 views and remains one of the first results to appear on Google, dating back to 2014.
No comments:
Post a Comment