Human-robot collaboration state monitoring method, device performing the same, and computer program

Human-Robot Collaboration State Monitoring Technology Using Low-Pass and High-Pass Filter Frequency Analysis

This technology is a monitoring method that determines in real-time whether a human-robot collaboration state is safe within a specific frequency band by passing multi-degree-of-freedom force signals through low-pass and high-pass filters and comparing the Euclidean norm values of each output signal.

Existing DFT-based frequency analysis techniques require large amounts of sampling data to achieve low frequency resolution, making it impossible to recognize collaboration states quickly within 0.5 seconds, which can lead to safety issues such as skin plastic deformation during collisions.

Instead of DFT, this technology separates frequency components using a 2nd-order IIR Butterworth filter and calculates the collaboration state value through median calculation using the ratio between filter outputs, derivative filter smoothing, and saturation processing. It can be applied to the safety control of collaborative robots and wearable robots, dramatically increasing operator safety through immediate risk detection within 0.5 seconds.

‍

‍

Key Features:

• A method for monitoring collaboration states through frequency analysis in environments where humans and robots physically collaborate
• A step of acquiring multi-degree-of-freedom force signals to distinguish between intended and unintended human movements
• A step of acquiring collaboration state values based on multi-degree-of-freedom force signals using low-pass and high-pass filters
• A step of acquiring intermediate values used for obtaining collaboration state values based on the multi-degree-of-freedom force signals that have passed through each filter