I am an Electrical and Computer Engineering graduate from the University of Toronto. I'm interested in computer hardware, electronic circuit designs and robots. I've completed minors in Mechatronics and Robotics and Engineering Business.

I am involved with multiple robotics teams and completed multiple technical projects. I've been programming and taking on technical work for about 5 years fascinated by troubleshooting complex problems, working in team environments and being able to develop technology with lasting impacts on society.

I am an active member on the University of Toronto Robotics for Space Exploration team (RSX) being in the organization for about 5 years. I joined RSX in my first year as an electrical team member. In my second year, I took on greater responsibilities being the Junior Electrical Lead. For the past year, I've also been a member on the University of Toronto Aerospace team and more specifically, I worked in the firmware systems division under the Space Systems subteam. And in my third year, I became the Electrical Lead for 2 years. In my final year of engineering, I became the co-president for the Robotics for Space Exploration team.

Please be advised that the source code for all of these projects cannot be released due to copyright issues and intellectual property theft. I've tried my best to use images to help backup the content I've wrote in the README portions of the documentation.

Feel free to get in touch with me by referring to the contact section below.

• Implemented a live trumpet replicator SoC using System Verilog, C Programming on the Xilinx Artix-7 (Nexys 4 DDR and Video) FPGA boards collaborating in a team of 4 to create hardware with Vivado (+IP Integrator) and software on the MicroBlaze processor using SDK
• The project will listen to audio from the user using a SPI MEMS microphone; process the audio data in real-time using digital signal processing algorithms that include sampling, filtering, equalization, applying the Fast Fourier Transform and verification with MATLAB; and identify the note being played by determining the respective frequency (pitch)
• The positive impact of the project will be with learning music and there will be a live animated trumpet output to an HDMI monitor (from FPGA). This graphical interface will help each beginning trumpeter see how to play a music piece

• Implemented and embedded a computer system C Programming on the Intel® DE1-SoC board with ARM Cortex A9 processor to create a game that helps users learn about the various organs in the human body in an amusing manner
• Employed a combination of logic instructions, interrupts, registers, input/output devices including a keyboard, timer,
7 Segment Displays, and VGA display to develop the software finite state machine for the game control

• Used Logisim (remote circuit development software) to implement 7400 series dual in-line package (DIP) chips to study circuits with logic gates and understand how they can be used to manipulate and produce signals as desired

• Designed a hexdecoder circuit using Verilog with multiplexers, modular hierarchy, and a 7 segment display to display hexadecimal values according to four switch signals (four bits)
• Utilized Intel Quartus Prime to implement the Verilog design on FPGAs (Field Programmable Gate Array) describing the hexdecoder
• Simulated the hardware design using ModelSim software and verified the output waveforms to ensure that the tool synthesizes the circuit correctly the first time since it can be time consuming sometimes

• Designed a 7-to-1 multiplexer using Verilog case statements in an always block
• Developed a ripple carry adder circuit that adds two 4-bit numbers using four instances of a full adder module
• Built and simulated an ALU by implementing all required operations including addition, subtraction, sign extension, etc., and connected the outputs to a 7-to-1 multiplexer
• Using 3 select bits to drive the multiplexer select lines, the output value for the ALU can be chosen and then displayed on to LEDs and a set of HEX displays

• Developed a gated D latch using Verilog that depends on the level of clock and placed two of them in series with opposite edges of clock to create an edge triggered D flip flop (also known as a primary-secondary flip flop) in which the output changes when the clock edge rises and also implemented an active low, synchronous reset
• Designed an 8-bit register by having 8 flip flops in parallel to one another
• Extended the functionality of the ALU by storing the output into an 8-bit register and implemented feedback into the second input of the ALU using the four least significant bits of the register output
• Designed and simulated a left/right 8-bit rotating register with parallel load using 8 instantiations of the flip flop module in which shift occurs by copying the bites to the next flip flop to the left or right on successive clock cycles. The parallel load enables the user to enter a preset set of 8 binary bits

• Implemented a series of 8 T-flip flops in series using Verilog with additional combinational logic to develop an 8-bit counter with an enable signal and an active low asynchronous reset
• Developed a rate divider (counter) that would use an industrial clock to increment a counter at varying speeds by having two counters. The rate divider creates pulses at the required rates by counting down certain numbers of clock cycles from the CPU clock frequency to achieve the desired speeds

• Designed a finite state machine using Verilog that would identify certain sequences of binary patterns for the purpose of identifying specific numbers from a long sequence of binary bits
• Emulated a typical calculator by implementing a finite state machine for the control to load numbers entered by the user into registers. It also accounts for the wait state as the user clocks the information
• Used another separate finite state machine for the data path. Based on enables signals from the control path, the numbers entered by the user are stored in multiple registers. Then, it is relayed into Arithmetic Logic Units using multiplexers depending on the preferred arithmetic operation
• Finally, the output is stored in register and output to LEDs and Hexadecimal displays at run-time

• Developed a RAM memory block using Quartus IP Catalog and Verilog to store the pixels in a frame buffer which is then used by the VGA Controller to display images on to a monitor
• Implemented a control finite state machine to take in user input and create enable signals for the data path
• Designed a finite state machine that would take X, Y and RGB colour inputs to develop squares on the screen by modulating counters to pixelate (4 x 4) 16 more squares going in a down-right direction

• Simplified plant incubation for busy people to enhance the nurturing process for healthy plants
• Developed livestreaming support for real-time camera feed to server using Java, ESP32 & Android Studio, enabling users to monitor plants via an Android app
• Designed the electronics with a transformer, step-down regulators and fuses to deliver the correct power safely to all downstream devices
• Implemented serial (UART) communication between Arduino Uno and ESP32 to inform primary controller which plant to harvest
• Programmed the 3-axis gantry system using stepper motors with C++ & Arduino for precise movement to harvest plants autonomously once they are detected to reach maturity via periodic scans using an ultrasonic sensor
• Achieved University of Toronto ECE Capstone Certificate of Distinction

• Implemented a SLAM algorithm in ROS using Gmapping and Rviz visualization software with C++ , ROS, TurtleBot, OpenCV to allow a TurtleBot to traverse and explore a maze and generate a map of the environment (using graph theory)
• Programmed a node-based depth-first search algorithm to systematically explore unvisited areas
• Avoided obstacles by utilizing sensor data from laser and bumper sensors, backtracking to previous nodes
• Utilized the ROS AMCL and move_base packages to effectively navigate the robot to five objects placed at known locations on the map. Tested algorithms using Gazebo simulation software. Gyroscope sensor data was used to calculate odometry data to gather positional and orientation measurements for efficient localization
• Used OpenCV framework to identify the objects in the FOV of an RGB camera using image recognition with SURF algorithm to detect features, FLANN algorithm to compare the object’s features to the template, and Lowe’s ratio test to filter good matches

• Developed a Convolutional Neural Network (CNN) machine learning (ML) model with Python that successfully classifies brain tumours (abnormal cell developments in skull) based on MRI images to aid medical doctors make accurate diagnosis
• Collected a dataset of 7500 images with Meningioma, Pituitary, Glioma and No Tumour pre-processing the images with cv2 library for gray scale, histogram equalization, Gaussian blur (noise) and balancing the set resizing to smallest size
• The encoder is stacked with 5 convolutional layers with 3x3 kernels each followed by ReLU activation function, batch normalization and max pooling to extract features implemented in Python with Pandas, Numpy and PyTorch libraries
• Flattened the feature maps from the last convolutional layer into a 2-layer Artificial Neural Network (ANN) to classify the images and the final prediction probabilities are determined by applying a SoftMax function for the distribution
• Performing hyperparameter tuning, achieved a model with 94.4% testing, 95.3% validation and 99.9% training accuracy beating the Random Forest Decision Trees algorithm from Scikit-learn ML library in Python

• Developed a web application using Python, HTML/CSS, ReactJS, Flask, Bootstrap to help post-secondary students explore and search courses for their time in university, create a course plan that would follow the requirements to become an engineer such as kernel, depth and elective courses, and programmed a recursive map to visually see course prerequisites
• Adopted agile development practices using Jira.
Link: https://github.com/ECE444-2022Fall/curriculus
• Developed a Flask App to detect if a snippet of text is Fake News or not using machine learning (ML) and with Heroku to make the model accessible as a service using a REST API call. Deployed the ML model to the cloud using AWS elastic beanstalk for ease and scalability, and measured API latency over 100 calls to the REST server
• Achieved a score of 94% (A+) in the course and 4th place out of a class of about 120 students

• Developed software application using C++ that would allow users to navigate through cities around the world in a user-friendly manner similar to Google Maps
• Added mouse hover functionality to aid users with how to use the software application, zooming and highlighting capabilities based on the user's search input, real-time search suggestions and autocompletion based on keyboard input from user, and pop-up message boxes for incorrect/successful user input operations and help button
• Implemented Libcurl functionality for users to view real-time delays on highway 401 and ability to view subway stations on the map
• Used a toggle button to show hidden second search bar to find path between two intersections. The path will be highlighted and have arrows for direction.
The path determined by the application will be the shortest path implemented using Dijkstra’s Algorithm
• In order to solve the delivery truck problem with multiple pickups and dropoffs of packages, a greedy algorithm is implemented to determine the local best path optimizing the time by introducing a 2-D matrix to store the values to be reused and using 2-Opt Algorithm to change the delivery order to be efficient
• Used timers from Chrono Library to evaluate the speed of our program and boost its performance

• Developed software using C++ that would implement an artificial intelligence (AI) to destroy as many asteroids (randomly generated by a server) as possible to save the Starship Galaxy Explorer from destruction
• Designed an advanced linked list that observes asteroids in the ships vicinity and updates each asteroid in the list based on how it move in the x and y directions, it’s health after each shot by the ship and whether the asteroid is destroyed or moved away from the ship’s visible range
• The AI would traverse this list of asteroids and consider factors including proximity of asteroids to the ship, the reload time of each shot, the asteroids’ remaining health to decide on which asteroid to shoot at
• Once decided, the AI uses its method functionality to maneuver the ship controller by rotating the ship and firing a shot at the asteroid. All these operations repeat at every clock cycle of the CPU

• Developed software using C++ to allow users to store information about shapes in a memory (dynamic) efficient database
• Designed a complex linked list with pointers to dynamically allocate and remove memory for each shape
• Implemented a command line with a graphical user interface for the user to ensure that there are no errors in the asserted commands and provided specific feedback as to what the particular error was parsed
• Inherited and reused code from others and extended the functionality (using object-oriented programming) to ensure that different shapes with different properties (information) can also be added to the database
• Created commands that allow users to create shapes, create shape subgroups, draw all or subgroups of shapes, move shapes from one subgroup to another, and remove shapes from the database by implementing methods in my linked list for efficient list traversal, list insertion, deletion of a node, and deep assignments

• Developed software using C programming that would replicate the game of reverse to allow people to play the game at home during the COVID-19 pandemic; this method proved to be a quicker and cost-effective way of playing the game
• Programmed an artificial intelligence that can defeat human players using game trees, minimax algorithm and recursive implementation
• Ensured that the AI can make moves within one second of calculations allowing the program to be efficient

• Utilized Java to create the classical Mastermind game to make it easier and efficient to play the game taking less time and it’s free.
• Implemented Graphical User Interface (GUI) for the game board and pins to make it visually appearing for the player

• Utilized Java to create an application that simulated an MP3 Player to allow customers to listen to music of their preference
• Executed features including GUI for the menu, MP3 Player case design and selection of songs to make it user-friendly to navigate through and find songs efficiently

• Developed software using Java that would replicate a traditional Chess game to be a more efficient way of playing the game taking less time and no need for the physical board and pieces
• Added features including a virtual game board, menu and pieces lost for the GUI to make easier for the user to see and have a better experience

• Using an Arduino UNO, programmed a police alarm independently with C++ to detect when a thief would enter a room using a sensor and display a message on an LCD screen calling 911 while outputting a programmed police alarm
• Programmed a lighting system using Arduino and an Integrated Circuit chip to enable lights to turn on/off according to logical operators within the chip and binary systems
• Utilized Arduino to program a seven-segment display in C++ that would be able to simulate
• Timers such as the ones at Ice Hockey games to show time elapsed

• Developed a device with SPDT (single pole double throw) switches, motors and other electrical components to create a device that would independently operate through a maze

• Utilized Infrared Sensors and programming in C++ to maneuver a small robot that would be used to stay in an arena, detect other Sumo Bots and engage in combat with them

Email: [email protected]

LinkedIn: http://www.linkedin.com/in/krishanth-suthaharan