Building intelligent robotic systems at the intersection of control theory, computer vision, and multi-robot coordination. ROS2, MATLAB/Simulink, Python, and C from simulation to hardware.
Target Roles
PID, FDI, and controller reconfiguration for fault-tolerant UAV systems. Validated from MATLAB/Simulink digital twin through Crazyflie hardware.
YOLOv8 and OpenCV pipelines for weed detection on ROSMaster X3. ArUco-based maze detection for 6-DOF robotic arm control.
Centralized autonomy stack for 5 TurtleBots. Hungarian task allocation and A* motion planning validated across 10 scenarios.
Hands-on experience with Crazyflie 2.1 (fault-tolerant control), Parrot Mambo (autonomous vision landing), and P80 heavy-payload multirotor (ArduPilot PID tuning).
ROS2 stacks, OpenCV pipelines, and sensor fusion nodes. Embedded C for low-level motor control and hardware interfacing.
Digital twin development in Simulink for autonomous drone landing. Multi-robot Gazebo environments with full collision avoidance.
| Parameter | Value | Variance | Status |
|---|---|---|---|
| FDI Recovery Time | < 50ms — Crazyflie 2.1 motor fault injection | ± 2ms | [ ACTIVE ] |
| CV Detection Accuracy | 97% @ 12 FPS — YOLOv8n on ROSMaster X3 | ± 0.5% | [ PEAK ] |
| Navigation Success Rate | 99% — 5 TurtleBots, 10 Gazebo scenarios | 0.01% | [ OPTIMIZED ] |
| Autonomous Landing CEP | 8 cm radius — Parrot Mambo on moving platform | ± 1.2 cm | [ VALIDATED ] |
| Maze Plan & Execution | 60 s total — myCobot600 8x8 ArUco grid (A*) | ± 3 s | [ NOMINAL ] |
| UAV Stability Gain | +5% — ArduPilot PID tuning on P80 multirotor | ± 1% | [ NOMINAL ] |
Experience Record: 2023 to Present
Aatram
Co-founded an emotion-first anti-procrastination app for students and young professionals. Designed an adaptive nudge engine leveraging Apple Intelligence for context-aware notifications. Built a Momentum Board with streak alternatives — Bounce-Back Score, Monthly Chapters, and momentum tracking — replacing punitive streak systems with psychologically grounded progress mechanics. Integrated evidence-based focus techniques including implementation intentions, WOOP, affect labeling, and temptation bundling as Pomodoro alternatives.
Honeywell — Arizona State University Collaborative Research
Architecting Sky Speak, an AI-powered pilot training platform for Honeywell Aerospace's Anthem ecosystem. Integrating voice stress analysis with CBTA competency grading to assess pilot cognitive load during ATC communications. Built a React instructor dashboard with real-time communication analytics and performance visualization.
Arizona State University — Intro to MATLAB Programming; Controls & Systems Lab
Supporting 90+ students across 5 lab sections for MATLAB Programming and Controls & Systems courses at ASU's Fulton Schools of Engineering. Leading review sessions, running office hours, and grading labs covering PID control and systems analysis.
Marut Drones (IIIT-Hyderabad)
Integrated electronic and mechanical subsystems across multi-rotor platforms at one of India's leading agricultural UAV companies. Configured ArduPilot and Mission Planner for a P80 heavy-payload drone, improving production efficiency by 15% and flight stability by 5% through PID tuning.
Project Archive: 2024 to 2026
FDI Recovery
<50ms
Platform
Crazyflie 2.1
Validation
ROS2 Sim & Hardware
fault = fdi.check(
motor_states,
tf2.TimePointZero
)
if fault.detected:
ctrl.reconfigure(
fault.failed_id,
cmd_wrench
)Fault-tolerant flight control stack with FDI and controller reconfiguration. ROS2 test harness for log replay, automated flight scenarios, and stability bounds tracking on Crazyflie 2.1 hardware.
Robots
5 TurtleBots
Success Rate
99%
Scenarios
10 validated
def allocate(robots, tasks):
C = build_cost_matrix(
robots, tasks
)
assign = hungarian(C)
for r, t in assign:
plan_path(r, t, costmap)Centralized ROS2 autonomy stack in Gazebo for 5 TurtleBots. Hungarian task assignment and A* motion planning with costmap-based collision avoidance. Validated across 10 scenarios with 99% mission success.
Accuracy
97%
Throughput
12 FPS
Dataset
3000 samples
def detect(self, frame):
results = model.predict(
frame, conf=0.5
)
for box in results.boxes:
cls = box.cls.item()
if cls == WEED_ID:
self.flag(box.xyxy)YOLOv8 model trained on 3000 samples, deployed on ROSMaster X3 at 97% accuracy and 12 FPS. Integrated ROS inference and visualization nodes for real-time robotic decision-making in precision agriculture.
Landing Accuracy
8 cm CEP
Sim Trials
25 validated
Platform
Parrot Mambo
def track_target(self, frame):
target = self.detect_marker(frame)
error = target.pos - self.uav.pos
cmd = self.pid.compute(error)
self.publish_velocity(cmd)Guidance and control logic built in MATLAB/Simulink (digital twin, 25 trials). Onboard vision tracking on Parrot Mambo achieves closed-loop landing within 8 cm of a moving platform.
Grid
8×8
Plan Time
60 s
Arm
6-DOF myCobot600
grid = occupancy_grid(
frame, aruco_markers
)
path = astar(grid, start, goal)
for waypoint in path:
arm.move_to_waypoint(waypoint)OpenCV pipeline for maze detection using 2 ArUco markers on an 8×8 occupancy grid. A* path computed and executed via myCobot600 6-DOF motion control; 60 s planning time. Applicable to industrial pick-and-place workflows.
Certifications, Publications & Education
Core & Pro
Universal Robots
Machine Learning & Deep Learning
MathWorks
Reinforcement Learning
Hugging Face
Robot Operating System
Udemy
M.S. Robotics & Autonomous Systems
GPA: 3.89 / 4.0
B.E. Electrical & Electronics Engineering
Hyderabad, Telangana
Available for full-time roles, internship roles, collaboration, and research roles
Current Status