Projects
Reinforcement Learning on the PIC18
Grid-X is a grid game with an oscilloscope display using the PIC18F87K22 microprocessor. The aim of the game is to get from one grid to the goal, accumulating the highest number of points possible. Reinforcement learning was implemented to allow to character to optimise and find the best path possible. The PIC18 is an 8-bit microprocessor with a clock speed of 16 MHz, hence special care had to be taken to implement 16-bit arithmetic for the reinforcement learning.
View Project
Heat Sink Optimisation
Used finite difference methods to solve the heat diffusion partial differential equation to model a system with a microprocessor with a heat sink. The widths and heights of the heat sink were optimised to achieve an optimal configuration for heat loss. Convection cooling was added to the model to achieve sufficient cooling of the microprocessor to suitable working temperatures.
View Project
Astronomical Image Processing
Analysed astronomical data from the SWIRE dataset to conduct a galaxy survey on the number of counts of galaxies. The detected galactic objects were profiled using the Sersic profile for galaxies and a Moffat point source function (PSF) for stars.
View Project
Millikan's Oil Drop Experiment
This traditionally entails sitting in front of a microscope and recording the movement of oil droplets under the application of an electric field to calculate the velocities and thus deducing the charge on each droplet. Instead, I decided to record videos of the moving oil drops under the electric field and utilised image recognition and particle tracking to automate the process of data collection. This allowed the collection of a larger data set which enabled the analysis of small systematic errors.
View Project
Investigating Aberration in Lenses
Built a framework for the simulation of optical rays through lenses. Optimisation was also done using scipy's optimise framework to find the best curvature which minimised spot radius. Worked on an independent extension on correcting chromatic aberration using combinations of lenses.
View Project
Charged Particles in the Magnetosphere
Simulation of charged particles in the Earth’s magnetic field using Python, tracking their trajectory and investigating the conditions required for the Northern/Southern lights to occur, which happens when the energetic particles enter the Earth’s atmosphere.
View Project