Attrac Lab develops LLM-based robot control and wireless charging systems

Attrac Lab has announced the development of two new systems designed to enhance the autonomy and operational efficiency of field robots, including drones, unmanned vehicles, and vessels.

The first system, AT-SYNAPSE, is an autonomous control system that utilizes Large Language Models (LLMs) as a robot’s brain. Compatible with the Model Context Protocol (MCP), it allows robots to use various AI models such as Claude, ChatGPT, or Gemini. The system interprets natural language instructions via voice or text and translates them into actions using six core tools: vision, speech, facial expressions, motion, navigation, and safety. It utilizes an open ROS 2 architecture to support diverse applications like patrolling, delivery, and research.

The second system, AT-Dock, is a wireless charging solution integrated with ROS 2. Utilizing technology from WiBotic, AT-Dock enables robots to manage the entire charging process autonomously—from returning to the station and adjusting position for optimal coil alignment to monitoring power levels and resuming operations. The system supports power outputs ranging from 150W to 1kW and features IP67-rated dust and water resistance. By combining AT-SYNAPSE and AT-Dock, the company aims to facilitate long-term, unmanned robot operations in various environments.