I was really interested in hearing about ML accelerators being built on different kinds on hardware called FPGAs. So, I decided self-teach myself about digital design by building a small NN inference accelerator in VHDL. I was fortunate to be selected in Blackbird's Protostars program to help fund a PYNQ-Z2 board.
VIEW ON GITHUB ↗
Skyguard was a really interesting project designed as a local airspace awareness system designed for drone flyers operating in outdoor areas. It detects aircraft's ADS-B signals using an RTL-SDR reciever connected to a Raspberry Pi. The Raspberry Pi then talks using MQTT to an embedded network that computes appraching aircraft's risk to the airspace and publishes its findings to wireless actuator nodes and a dashboard.
VIEW ON GITHUB ↗
A challenging project that was very satisfying to demo! Implemented BLE-based indoor localisation using iBeacon data fusion. Essentially, there was a lab with many different airtags (iBeacons) that were transmitting packets. The Mobile Node (placed on a person) would report RSSI measurements whilst walking around the lab to the Base Node. The Base Node uses a Kalman Filter to estimate the localisation of the Mobile Node to a wireless display and a dashboard.
VIEW ON GITHUB ↗
Built as part of my internship at Hand Plus Robotics in Malaysia. Learnt how challenging it can be to integrate hardware (cameras) systems together, ROS2, and the cool cameras RealSense makes. Built a vision system for an autonomous welding robot.
VIEW ON GITHUB ↗
Developed an embedded FreeRTOS-based control system for a remote-controlled microscope on the STM32F429ZI platform. Produced modular driver libraries to manage physical and digital command inputs, radio transmission to the microscope using the NRF24L01+ module, and display outputs to RGB LEDs and a seven-segment display.