Remcom

We are very excited today to release 𝗗𝗲𝗲𝗽𝗠𝗜𝗠𝗢 𝘃𝟰 — a new vision for the wireless ray-tracing dataset framework!   Over the years, DeepMIMO has grown mainly as a dataset for deep learning and MIMO research. With DeepMIMO v4, we're reimagining it as a full 𝘁𝗼𝗼𝗹𝗰𝗵𝗮𝗶𝗻 𝗮𝗻𝗱 𝗱𝗮𝘁𝗮𝗯𝗮𝘀𝗲 𝗳𝗼𝗿 𝗿𝗮𝘆-𝘁𝗿𝗮𝗰𝗶𝗻𝗴 𝗱𝗮𝘁𝗮𝘀𝗲𝘁𝘀, built to accelerate wireless AI research for 5G, 6G, and beyond.   Main highlights: - 𝗢𝗽𝗲𝗻-𝘀𝗼𝘂𝗿𝗰𝗲 𝗰𝗼𝗻𝘃𝗲𝗿𝘁𝗲𝗿𝘀. Any researcher can now convert their own ray-tracing simulation into a standardized DeepMIMO scenario with a single line of code. - 𝗦𝘂𝗽𝗽𝗼𝗿𝘁𝗶𝗻𝗴 𝘁𝗼𝗽 𝗿𝗮𝘆 𝘁𝗿𝗮𝗰���𝗿𝘀. DeepMIMO converters now support Remcom Wireless InSite, NVIDIA Sionna RT, and AODT, with more coming soon. - 𝗔 𝗳𝘂𝗹𝗹 𝘁𝗼𝗼𝗹𝗰𝗵𝗮𝗶𝗻. DeepMIMO v4 bridges the gap between ray tracers and modern simulation toolboxes like Sionna PHY/SYS, MATLAB 5G Toolbox, and NeoRadium through a clean, standardized format. - 𝗦𝗵𝗮𝗿𝗲𝗮𝗯𝗹𝗲, 𝘃𝗲𝗿𝘀𝗶𝗼𝗻𝗲𝗱 𝗱𝗮𝘁𝗮𝘀𝗲𝘁𝘀. Researchers can now publish their DeepMIMO datasets, and anyone can reproduce the exact same data with a few lines of code using the DeepMIMO APIs, enabling benchmarking and reproducibility. - 𝗔 𝗴𝗿𝗼𝘄𝗶𝗻𝗴 𝘀𝗰𝗲𝗻𝗮𝗿𝗶𝗼𝘀 𝗱𝗮𝘁𝗮𝗯𝗮𝘀𝗲. Hundreds of datasets ready to use, with a publications database so you can search benchmarks by topic and application. - 𝗢𝗽𝗲𝗻 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 & 𝗰𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝘆. Improved documentation, new tutorials, a community forum, and a public roadmap. A huge thank you to the 𝟮,𝟴𝟬𝟬+ 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵𝗲𝗿𝘀, 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝘀, 𝗮𝗻𝗱 𝘀𝘁𝘂𝗱𝗲𝗻𝘁𝘀 from 400+ institutions across 30+ countries who accessed the beta over the last year; many provided valuable feedback and contributed to the code. The development of DeepMIMO v4 is mainly led by my former PhD student, now with NVIDIA, João Morais, web platform led by Soham Daga, with key contributions from Sadjad Alikhani, Namhyun Kim, and Kengmin Lin. Also thanks to Umut Demirhan and Abdelrahman Taha who led the development and maintenance of DeepMIMO v2 and v3 over more than 5 years! And thanks to Remcom and Tarun Chawla for the continued support to DeepMIMO! Website: https://www.deepmimo.net/ GitHub: https://lnkd.in/gYK8z2RM DeepMIMO v4 vs v3 at a glance: https://lnkd.in/gxAXUFMM #DeepMIMO #WirelessAI #6G #MIMO #RayTracing #DeepLearning #OpenSource #ASU

The dielectric inside your microstrip stack-up is your FRENEMY. The dielectric sandwiched between the signal and ground conductors stores electric field energy as the electromagnetic wave propagates down the transmission line. At the same time, due to non-idealities within the dielectric - primarily its loss tangent - part of that stored energy is dissipated as heat during every RF cycle as the material’s molecules are polarized by the electric field. In other words, the dielectric both guides the wave and slowly consumes its energy. In the animation shown, we visualize the density of dissipated power from multiple angles at 18 GHz using the same canonical microstrip structure used in several previous posts: • 20 mil RO4350B • 50 Ω microstrip line Isometric View This view shows the orientation of the moving 2D slice as it sweeps from the input connector to the output connector. The dielectric is made transparent so we can directly observe where the dissipated power density is highest: primarily inside the dielectric region beneath the trace. This view mainly serves as a reference for the side view. Top View (Z-Axis Sweep) Here, a slice moves upward through the substrate thickness. This helps illustrate how the dissipated power is distributed through the dielectric as a function of height. Notice that the loss is concentrated directly below the trace and near its edges where the electric field magnitude is strongest. Side View (Right-Most Connector) This is the key view. Most of the dissipated power occurs within the dielectric, particularly near the edges of the microstrip trace. This happens because the electric field becomes strongly concentrated at conductor edges due to boundary conditions, increasing the electric field magnitude locally. Since dielectric loss scales with frequency and the magnitude squared of the electric field, these high-field regions become the dominant contributors to heating at 18 GHz. ------------------ At MovaMicrowave, we design broadband DC-18 GHz RF/Microwave power amplifier modules structured around milestone-based SOW programs. If you have defined power, bandwidth, form factor, and DC constraints, we can architect and sustain the solution. Consult us for a complimentary session to better understand your mission and requirements. Feel free to reach out on LinkedIn, email, or directly from our website: Email: contact@movamicrowave.com Website: https://lnkd.in/gVgVNEgU LinkedIn Page: MovaMicrowave LLC #RF #Microwave #PowerAmplifier #Nonlinear #GaN #GaNonSiC #MicrowaveOffice #Educational #Contractor #Consultant Remcom and all other trademarks and logos for the company’s products and services are the exclusive property of Remcom Inc.