Hrushikesh Manjunath Thotamsetty

Designed and built a deformable modular robot prototype (DATOM) integrating scissor-lift actuation, magnetic latching (permanent + electromagnet), and infrared communication within a dodecahedral ABS shell. Created 3D models in Creo Parametric and SolidWorks through multiple design iterations, resolving volume constraints that made linear actuators geometrically infeasible. Programmed Arduino Nano control system for inter-module IR communication, verified via oscilloscope waveform analysis. Designed complete power management system: 9V input, dual voltage regulation (5V/6V), MOSFET-controlled electromagnet switching, total system draw 9.55W.

Derived analytical solutions for Couette flow between concentric cylinders under three boundary conditions using Navier-Stokes equations. Built computational meshes in GMSH, ran simulations in SU2, and post-processed velocity fields in ParaView. Validated computational results against analytical solutions, achieving close agreement with minor boundary-layer discrepancies. Generated normalised velocity profiles in MATLAB for direct theoretical-computational comparison.

Systematically disassembled a 13-component consumer product (toy garbage truck with circuit board, lights, and sound system), measured each part using calipers and radius gauges, and digitally reconstructed the complete assembly in Creo Parametric. Applied advanced CAD operations including revolve, chamfer, shell, and mirror to replicate complex curved surfaces. Achieved close dimensional agreement through iterative measurement and modelling refinement. Produced full bill of materials, exploded assembly views, and 2D technical drawings.

Low-cost vibration anomaly detection system targeting SME manufacturers who cannot justify enterprise predictive maintenance solutions. Built around MPU6050 accelerometer + Arduino Nano, streaming real-time FFT data to a Python dashboard. Addresses a real gap: enterprise systems cost thousands; this prototype demonstrates the same core principle under £50. Directly extends unresolved DATOM hardware noise problem — applying frequency-domain analysis to distinguish actuator wear signatures from background vibration.