Examined a position using JavaScript to calculate 1000 moves per second

Evaluated future positions using a Minimax algorithm to select the move that results in the greatest point and positional advantage

Used alpha beta pruning to allow the Minimax algorithm to reduce move selection at layer 3 by 50% and allow looking 5 layers further with a time limit of 3 seconds

Selected hardware to capture and process images using weighted pros and cons while staying within a $600 dollar budget

Field tested camera and algorithm using a beach bonfire resulting in positive identification of a fire at 1300ft with only 3 false positives over 15 images.

Included feedback from accelerometer and magnetometer to correct gyroscopic error and bias drift, making attitude estimation resistant to quick movements and extended use.

Corrected gyroscopic error and bias drift using feedback from accelerometer and magnetometer to create an estimator resistant fast movement and extended use.