49 Views

MathWorks Altera Collaboration on AI for 5G and 6G

LinkedIn Facebook X
March 07, 2025

Get a Price Quote

MathWorks has partnered with Altera to bring AI-based autoencoders to FPGAs for the advancement of 5G and 6G wireless systems. These autoencoders play a crucial role in compressing Channel State Information (CSI) data, leading to a significant reduction in fronthaul traffic and bandwidth requirements. This collaboration allows engineers working on wireless communications systems to uphold user data integrity, while also ensuring reliability and performance standards are met, all while reducing costs.

One of the key aspects of this collaboration is the utilization of tensor units within the fabric of the Altera Agilex families of FPGAs. These units are specifically designed to accelerate AI operations, enabling faster and more efficient processing of data for wireless applications.

According to Mike Fitton, Vice President and GM of Vertical Markets at Altera, this collaboration opens up a wide range of possibilities for AI integration in wireless communications applications, ranging from 5G RAN to advanced driver-assistance systems (ADAS). By leveraging the FPGA AI suite and MathWorks software, developers can seamlessly transition from algorithm design to hardware implementation, ensuring that their AI-based wireless systems meet the demanding requirements of modern applications.

The MathWorks Deep Learning HDL Toolbox, in conjunction with the HDL Coder tool, empowers developers to customize, build, and deploy efficient Deep Learning Processor IP Cores. The addition of autoencoder support further enhances performance and flexibility in wireless applications by providing support for standard networks and layers.

By combining the Deep Learning Toolbox with the OpenVINO toolkit, developers can streamline the optimization of AI inference through the Altera Quartus Prime FPGA compiler. This integration offers a seamless pathway for developers to enhance AI applications and advance 5G and 6G wireless systems, ultimately leading to more efficient and high-performance wireless solutions.

Recent Stories