Darshak Panseriya

Developed real-time UWB CIR analysis tools (Python/PyQt, FFT, jitter, statistics) enabling instant signal-level insights. Executed UWB radar & ranging validation for automotive functions (CPD, Kick Sensor, Intrusion Detection), including antenna diversity, bumper influence, and stress testing. Automated CANoe–CAPL–Python test workflows, scalable data pipelines (100k+ files), and labeling systems (radar + video sync), reducing manual effort by >80%. Tech: Python, MATLAB, CANoe, CAPL, OpenCV.

Developed and validated YOLOv8-pose perception pipeline on full FSOCO dataset (>11k images). Built custom AP-vs-IoU and keypoint-aware validation beyond standard mAP, including confusion-matrix and RMSE analysis. Implemented distance-based gating (5–20 m) and mask-based keypoint error estimation, revealing 25–35% recall losses and dataset gaps. Automated self-labeling, reducing manual effort by ~100 hours. Tech: Python, YOLOv8, OpenCV, Pandas, Matplotlib.

Led conception and development of a micromobility vehicle platform from architecture to production. Designed and integrated suspension, brakes, steering, and drivetrain systems. Performed vehicle kinematics and dynamics analyses to ensure performance and robustness. Executed CAD design, packaging, space optimization, and component-level FEA. Delivered production drawings and BoM supporting manufacturing readiness. Tech: SolidWorks, ANSYS, Adams, Advanced Excel.

Implemented object-based sensor fusion using mono-camera and 3D automotive radar for road user detection. Trained and validated YOLOv8 detectors with dataset balancing and augmentation to improve robustness. Developed validation metrics and performed data association between image detections and radar clusters. Designed tracking and motion prediction concepts using EKF-based track management. Tech: Python, YOLOv8, OpenCV, Pandas, Matplotlib.

Developed and optimized suspension, chassis, and vehicle dynamics for a Formula Student race car. Analyzed 3.5M+ tyre data points to derive load cases and dynamics models. Designed and validated suspension kinematics, F1-style decoupled heave–roll system, and adjustable blade anti-roll bar. Executed component and chassis FEA achieving up to 15% weight reduction. Led testing, data acquisition, and on-track setup validation. Tech: SolidWorks, ANSYS, MSC Adams, Lotus Suspension Analysis, Advanced Excel.

Led design and development of an SAE competition vehicle platform, including chassis and drivetrain systems. Planned project schedules and coordinated subsystem integration. Performed CAD-based packaging, space optimization, and component stress simulations. Delivered production-ready design documentation, BoMs, and built the vehicle. Cleared technical and safety inspections in the team’s debut year, securing 36th place, establishing a scalable vehicle architecture for future iterations.