MSc Autonomy Technologies · Robotics, Sensor Fusion and Control
ROS 2 · Camera/LiDAR/Radar Data · State Estimation · Engineering Software
I am an MSc student in Autonomy Technologies at FAU with a BSc in Electrical and Electronics Engineering from Atilim University. My primary portfolio focuses on robotics, autonomous systems, sensor data, control, embedded systems and reproducible engineering software.
I am targeting robotics, autonomy and Robot Deployment Engineer roles. The evidence below comes from deterministic simulation, synthetic data and public datasets; planned work and validation limits are labelled explicitly.
Currently building: Autonomous Sensor Fusion Lab and FaultNav ROS 2.
| Project | What is implemented | Status and boundary |
|---|---|---|
| Autonomous Sensor Fusion Lab | Deterministic nuScenes table traversal, timestamp diagnostics, calibrated sensor-to-ego/global transforms, cross-time LiDAR/radar camera projection, bird's-eye visualization and synthetic CI fixtures. | Active, v0.1 foundation. A real nuScenes mini run is still pending; no detection, tracking or physical sensor validation is claimed. |
| FaultNav ROS 2 | Exact differential-drive motion, quantised encoder and seeded IMU simulation, fault injection, encoder-derived odometry, ROS 2 odometry/TF, reports and automated tests. | Active. Controlled software simulation only; EKF, physics simulation, SLAM, Nav2 and hardware work remain future milestones. |
| Automatic Control Laboratory | Five nonlinear/state-space control studies, LQR, observers, saturation and anti-windup, independent Python references and portable C99 runtimes. | Maintained. MATLAB, Python and C software validation; no hardware or production-controller claim. |
| Industrial Quality Anomaly Monitor | Deterministic synthetic manufacturing data, global and machine-aware robust baselines, Isolation Forest, a shared-dataset comparison runner, reports, Docker and tests. | Active. Synthetic comparison evidence, not real-factory performance. |
| Power Electronics Manufacturing | A tested Python shear-curve workflow covering preprocessing, features, batch validation, review-oriented outlier evidence, capability gates and two-parameter Weibull analysis. | Maintained, private repository. Generic/synthetic data only; no proprietary or production validation evidence. |
- Embedded BMS and CAN Simulator — specification and milestone plan only. The intended C++ battery model, protection state machine and virtual CAN workflow do not yet have an executable foundation.
- Radar / ISAR Classification Pipeline — specification and milestone plan only. Dataset processing, baselines and evaluation are not yet implemented.
| Area | Demonstrated tools and methods |
|---|---|
| Robotics and autonomy | Python, ROS 2 interfaces, differential-drive modelling, sensor/fault simulation, odometry, TF and deterministic scenarios |
| Autonomous-driving data | nuScenes devkit, camera/LiDAR/radar calibration, coordinate transforms, timestamp diagnostics, projection and bird's-eye visualization |
| Control and embedded foundations | MATLAB, state-space methods, LQR, observers, numerical integration, saturation, anti-windup, portable fixed-size C99 and CMake/CTest |
| Manufacturing analytics | NumPy, pandas, scikit-learn, robust statistics, Isolation Forest, curve analysis, process diagnostics and Weibull modelling |
| Engineering workflow | pytest, Ruff, GitHub Actions, Docker, reproducible CLIs and reviewable CSV/JSON/SVG/PNG artifacts |
- Separate ground truth, measurements, estimates and evaluation data.
- Record units, coordinate frames, timestamps and assumptions explicitly.
- Use deterministic scenarios and tests before publishing measurements.
- Keep failure handling and rejected data visible instead of silently cleaning it away.
- Distinguish software evidence from real-world validation.
Engineering evidence should be measurable, testable and reproducible.
github.com/seneserisen