Scout (Unitree Go2) - Four-legged robot

Focus: Context-Aware Risk Detection & Adaptive Risk Management

Problem & Motivation: Typically, risk assessment and safety development processes in robotics are highly conservative and rigid. However, the increasing use of mobile robots in human-centric environments requires that this risk assessment be versatile and context-dependent.
The Project Solution: The Scout project focuses on dynamic risk management. The robot dynamically adapts its safety parameters to its environment at runtime. In doing so, it actively distinguishes between people and other static obstacles to ensure safe and transparent robot safety management without compromising operational efficiency.

Context-Aware Safety: The system detects whether an object is a static obstacle or a person. Based on this, the robot dynamically adjusts safety distances and movement speeds depending on the proximity of people.

Context Awareness: The risk management strategy takes specific contextual factors into account—such as object classes and human attention—to trigger situation-appropriate warning signals.

Behavior Networks: The robot’s current safety status is continuously calculated and visualized in real time via an adaptive behavior network.