SATYAN BHOGE

Worked on Battery Management System of logistic UAV for mountainous region, used various AI algorithms for improving efficiency and overall temperature balance of batteries, used Supervised Learning algorithms like LSTMs(Long Short- Term Memory Networks) and random forests on historic data to predict failures and power distribution. Increased efficiency of battery by 15% and temperature deviation reduced to ±1-2°C which is 4 times stable than traditional BMS. Built a strong foundation in Deep Learning like CNN models like U-Net and mask R-CNN segment which helps to detect navigable and non navigable zones while working on Collision Avoidance, Increased accuracy to 93% by using Kalman filters and deep sensor fusion for handling different sensor's data. Developed innovative approach to stabilization of payload by applying YOLO8 and Model Predictive Control which predicts future movement of UAV and adjusts gimbal accordingly which greatly impacts on stabilization of overall UAV.

Developed an Mobile application in a team of 5 members, The goal of an application was make a offline trajectory predictor app which can be used by soldiers and army staff on field by ease. Worked on a feature called trajectory prediction, we have taken help of ML models such as LSTM for missile trajectory and gaussian probabilities which helps to estimate uncertainties in the real environment. Accuracy obtained was about 79%. Design different types of Aerodynamic Models to simulate them in virtual environments like ANSYS fluent and use different CAD softwares.