Open to new roles
04 · Mechatronics · closed-loop automation
Senior project · 2023
An end-to-end senior project combining embedded control, sensing, automation logic, and a physical robot system.
Role
Senior project · 2023
What I built
Programmed remote operation, scheduling, automatic tool changing, and load-cell-based mating logic.
Selected evidence
IEEE publication
The problem
The system automated a delicate coaxial connector-mating workflow, where repeatability depended on motion, load feedback, and the practical realities of the fixture.
Technical ownership
System architecture
Raspberry Pi scheduler + ESP32 control → HX711 load feedback → Cartesian robot and tool changer.
Evidence + outcome
Tools + systems
Case study
The mechanism had to do more than reach the connector. It needed to detect contact, change tools, schedule tests, and leave behind force data that could be inspected afterward.
Senior engineers at Teradyne described a testing process that still depended on careful manual connector mating and repeatable data collection. That conversation became the brief for the robot.
Mating cycle
I programmed the Raspberry Pi and ESP32 workflow that combined robot motion with HX711 load-cell readings. Closed-loop load detection let the system respond to the connector instead of relying on position alone.
Automatic tool changer
The tool changer used a Maxwell-style kinematic coupling to locate the toolhead repeatably. A servo rotated the locking shaft and compressed a preload spring, while the cable toolhead waited on a rack between operations. I implemented the automatic tool-change logic used by the scheduled test workflow.
Automation architecture
The Raspberry Pi ran the operator-facing workflow. Tool changes and connector moves became scheduled robot actions, while the load path stayed measurable throughout the test.
The Raspberry Pi hosted remote operation, test scheduling, manual controls, and the data viewer.
Python workflow with remote accessThe robot moved to the rack, coupled to the toolhead, and actuated the servo-driven locking shaft before continuing the test.
Automatic pickup and release logicThe HX711 sampled the load cell, the ESP32 forwarded force readings, and the Raspberry Pi used and stored that feedback during mating.
Force-aware motion and recorded evidenceTest evidence
The force trace made the mating sequence inspectable across multiple connectors. Together with 0.01 mm travel precision and a measured 0.5 g load standard deviation, it gave the project evidence beyond a working demonstration.
