Projects

Architected a production multi-camera 3D perception system integrating MonoDETR-based 3D object detection, open-vocabulary detection, semantic segmentation, and Depth Anything V2 metric depth estimation. Achieved sub-30 ms end-to-end inference on a single NVIDIA Jetson Orin SoC via a full PTQ/QAT INT8–FP16 TensorRT/ONNX pipeline: 3.2× model size reduction, 8× inference speedup, <1% mAP degradation.

Led the development of a unified visual perception engine using a DINOv3 backbone for on-device deployment. Established cross-modal alignment between visual and language representations for zero-shot generalization across unseen object categories. Deployed on Neubility's autonomous delivery robots in urban environments.

Designed a hardware-aware multiscale transformer for inertial motion tracking. NanoMST achieves 4.7× compute reduction over LSTM baselines using only 298K parameters with 8-bit quantization support, enabling real-time TinyML deployment on embedded devices. Published in IEEE Internet of Things Journal (IF: 8.9).

An AIoT-enabled inertial localization system achieving sub-meter accuracy across diverse environments without environment-specific retraining, combining deep domain adaptation with multi-sensor fusion. Published in IEEE Internet of Things Journal (IF: 8.9).

A hybrid ConvNeXt-Transformer architecture with formal ε-DP privacy guarantees for distributed multimodal sensor fusion, maintaining localization accuracy while protecting user data. Published in IEEE Transactions on Systems, Man, and Cybernetics: Systems (Early Access).

A Visual-Inertial SLAM hardware accelerator on PYNQ-Z1 FPGA achieving real-time navigation at 30 FPS with optimized feature-matching kernels that reduce SLAM latency by 60%. This work bridged software SLAM research and hardware deployment for mobile robotics. Published in IEEE Access (IF: 3.6).

A fusion-based framework for emotion recognition using wearable sensors. EmoHEAL integrates data from multiple sensor modalities to accurately detect and classify emotional states, with applications in mental health monitoring and human-computer interaction.

A federated learning approach for privacy-preserving inertial odometry. FedNav enables collaborative model training across multiple devices without sharing sensitive location data, balancing privacy requirements with navigation accuracy.

A context-aware framework for indoor-outdoor detection using smartphone sensors. DeepIOD uses deep learning to accurately classify environments, enabling adaptive behavior in location-based applications without privacy-invasive data collection.