June 8, 2021 – KIRKLAND, WA — Spectrum Effect® today announced the Deep Convolutional Neural Network (CNN) models contained in the latest release of Spectrum-NET are achieving the highest accuracy ever measured for classifying and aggregating RF interference.
Spectrum-NET performs automated, machine learning-driven analysis of RF interference, which enables operators to rapidly and effectively mitigate RF interference. Spectrum-NET contains an RF interference classification library of Deep CNN models that have been trained with labeled measurement examples from mobile operator networks across the globe. The Spectrum-NET CNN models classify various interference sources such as PIM, CATV egress, WiMAX, DECT, BDA’s and TDD self-interference.
The measured accuracy of the latest versions of the Spectrum-NET CNN models is greater than 92% (top 1) and 98% (top 2).
“The cost of RF interference to mobile operators is substantial today and these costs are increasing significantly as operators deploy 5G NR”, commented Spectrum Effect CEO, Charles Immendorf. “Our latest CNN models are another major step in the advancement of Spectrum-NET and will enable operators to save more OPEX while driving further network performance improvements.”
About Spectrum Effect
Spectrum Effect’s mission is to solve the most challenging and costly problems in the wireless industry through innovation and automation. With R&D located in Kirkland, Washington, USA, and Monterrey, Mexico, the Spectrum Effect team is passionate about creating disruptive technologies, engineering excellence, and enhancing the experience of mobile operators. Spectrum Effect has created Spectrum-NET, the industry’s leading solution for the automated analysis of RF interference. Protected by 27 issued patents, Spectrum-NET performs machine learning-driven analysis of RF interference across all RF bands and operates seamlessly across multi-vendor 5G NR, LTE and UMTS cells within mobile networks continually. With Spectrum-NET, operators across the globe are rapidly addressing RF interference, improving network KPIs, surgically deploying their field assets, gaining insights into spectral efficiency, and saving significant OPEX and CAPEX. www.spectrumeffect.com.
Contact:
Cara Mormino
+1 847 243 3000
cara@spectrumeffect.com
