The Polyglotbot is a 7-DOF Franka arm designed to recognize, translate, and transcribe handwritten text across diverse languages. The primary objective is to utilize a custom python MoveIt2 API to facilitate real-time linguistic conversion and writing, leveraging optical and computational techniques.
Key features include voice recognition, verbal translation, and automated pen management.

Utilizes an Intel RealSense Camera and YOLO technology for real-time detection and reading of handwritten text. While a user is writing a word to be translated on the whiteboard, the robot will detect there is a human in the frame and remain idle. After the robot detects that the user has cleared the area, it will use Optical Character Recognition (OCR) to read the handwritten text and begin the translation sequence

Integrates with the Google Translate API for accurate translation from the detected language to a specified language, followed by the robot speaking the word in both the original and translated languages.

Converts the translated text into waypoints using a custom 'string-to-waypoints' ROS2 package. These waypoints are converted into robot movement using a custom python MoveIt2 API and a custom 'write-letters' ROS2 package.

Employs three April tags and a second Intel RealSense camera to determine the optimal position for writing.

Supports voice commands for word submissions in any language, translating to English.

Automated handling of the marker, retrieving it from and returning it to a designated location at startup and shutdown.

1. Robot Arm: 7-DOF Franka Robot Arm
2. Cameras: 2 Intel RealSense Cameras for OCR and positioning with AprilTags
3. Software: Utilizes ROS2,YOLO, Google Translate API, custom python MoveIt2 API, custom 'string-to-waypoint' and 'write-letters' ROS2 packages for robot movement