Robotics: The S1S2 Way
Extending the Classic Sense-Think-Act Loop for the Era of Foundation Models
At S1S2, we are not reinventing the wheel; we are upgrading the brain behind it.
Our mission is to bridge the gap between traditional robotics and modern AI by extending the classic Sense-Think-Act loop with a dual-process architecture: System 1 for high-speed execution and System 2 for deep reasoning.
Enhanced Sensing: Beyond Raw Data
Traditional “sensing” focuses on geometry. We will augment classic algorithms such as Kalman Filtering and SLAM with the semantic capabilities of Vision-Language Models.
Intelligent Thinking: Reasoning
High-frequency deterministic logic to handle immediate trajectory adjustments and Large Language Models (LLMs) to perform Hierarchical Task Deconstruction.
Robust Acting: The Verified Guardrail
The adaptive suggestions from VLA (Vision-Language-Action) models of System 2 are “proposed intent” with System 1 acting as the ultimate logical governor.