Tony Currà

Portable Gas Detection Sensor for Police Applications

I redesigned an initial sensor system into a compact, portable prototype for police use in detecting laughing gas in drivers. The project—completed in roughly 16 days—involved:

Robotic ape for olive harvesting. From designing and prototyping mechanisms, putting together motors and power electronics for a portable robotics architecture able to move in agricultural terrain like an ape. Currently training the robot in physics engine (mujoco) with reinforcement learning and designing the software architecture for bioinspired perception and autonomous learning.

Led the design and development of a smart sensing add-on for hand rehabilitation devices, enabling remote patient monitoring through embedded sensing, AI, and web-based analytics. Took full responsibility for system architecture, prototyping, and field validation.

Key Contributions

- Research & System Architecture: Evaluated sensing technologies and designed a scalable roadmap from concept to clinical prototype.

- Embedded Electronics: Selected and integrated IMUs (6–9 DOF), environmental sensors, and microcontrollers (ESP32, nRF52840). Designed custom PCBs and power systems with LiPo batteries, BLE mesh networking, and low-power firmware (deep sleep, optimized sampling).

- Firmware & ML on Edge: Developed C++ firmware for sensor fusion, algorithms, BLE telemetry, and motion classification using lightweight ML models. Achieved <100 ms latency and multi-day battery life.

- Mechanical & Product Design: Modeled and 3D-printed ergonomic, biocompatible casings for wearable use. Iterated multiple prototypes toward miniaturized, battery-powered wireless nodes.

- Web & Cloud Integration: Built a BLE-enabled web dashboard (HTML/CSS/JS) for live visualization, data logging, and performance analysis. Deployed via GitHub Pages with database connectivity for clinical trials.

- Computer Vision Integration: Deployed MediaPipe-based hand landmark detection to compare camera-based vs. IMU-based tracking accuracy. Supervised an intern benchmarking both modalities for precision and latency.

- Testing & Validation: Conducted trials with physiotherapists and patients to collect motion data, validate algorithms, and refine UX and data visualization.

- Commercialization & Partner Collaboration: Engaged with an manufacturing partner to translate prototypes into scalable, production-ready designs.

Impact

Delivered a fully working wearable platform for motion tracking and rehabilitation analytics — bridging embedded systems, AI, and healthcare usability in a single, portable device.