NIST Researchers Automate Chamber Configuration, Quickly Replicating Industrial Settings for Wireless Systems Testing

NIST researchers developed an artificial intelligence protocol that can configure a test chamber to replicate the spatial characteristics of measured mmWave channels in industrial environments, thus allowing assessment of industrial wireless systems. Development of this protocol is described in “A Deep Reinforcement Learning Approach for Automated Chamber Configuration Replicating mmWave Directional Industrial Channel Behavior” which was presented by NIST researcher Mohamed Hany at the ARFTG-100th Microwave Measurement Symposium in January 2023.

The protocol meets a testing need for industrial wireless systems, which must be assessed for performance in industrial environments. Due to their highly reflective nature, industrial wireless channels differ from the characteristics of home and office channels. Thus, an industrial environment in which these wireless systems will operate must be replicated for industrial wireless equipment testing before deployment in “over-the-air” test chambers – which is often time-consuming.

NIST researchers developed a protocol that uses deep reinforcement learning – a subset of AI – that trains itself on a situation, getting rewards when it is right and costs when it is not. Researchers used the protocol to configure and tune the test chamber to the parameters needed for testing industrial wireless systems. Researchers then measured the performance of a specific mmWave system. The test method was validated by successfully comparing characteristics of channel measurements inside the test chamber, to measurements taken in a realistic environment.

This deep reinforcement learning protocol provides an automated approach to configuring a test chamber, thus allowing industrial wireless systems to be assessed for a wide range of environments, much faster and more efficiently than previously.

Comments are closed