Voice Controlled 4 DOF Robotic Arm
CompleteWhy
Voice interfaces lower the barrier to operating robotic hardware without a direct controller. This project investigates how natural language commands can map to precise servo movements in a 4 DOF arm, all while keeping the entire build under $100, which was the biggest constraint when procuring parts.
How
Modeled the full arm assembly in Onshape and printed the components in ABS, then reworked the design after SimScale FEA simulations to minimize high-stress points in the servo enclosures. Wired the system from a schematic made in Cirkit Designer, soldering the servo driver connections to avoid loose wires. On the software side, I wrote inverse kinematics control firmware in C++ on an Arduino and built a Python speech-recognition layer that parses spoken commands and sends serial instructions to the microcontroller. Once the core code worked, I used Claude Code and Cursor to build a "J.A.R.V.I.S." website for sending voice commands, adding another layer to the pipeline.
Results
Achieved reliable control through voice alone with sub-second command latency, and developed a "J.A.R.V.I.S." style interface that can execute multiple commands. The project was presented at the University of Redlands Engineering open house to faculty and students. Next, I plan to build a larger arm with cycloidal gearboxes at each joint for smoother control and higher load capacity.