Development of 5G and 6G
Mathworks and Altera cooperate
Mathworks and Altera cooperate to advance the development of wireless systems for Altera FPGAs.
Mathworks, a developer of software for mathematical computing, and Altera, a subsidiary of Intel, have joined forces to accelerate the development of wireless systems for Altera FPGAs. The use of AI-based autoencoders enables wireless system developers to compress channel state information (CSI) data and significantly reduce fronthaul traffic and bandwidth requirements. This is intended to ensure the security of user data, maintain the reliability and performance standards of wireless communication and reduce costs at the same time.
Mathworks offers a comprehensive tool suite that optimizes AI and wireless development - especially for Altera FPGAs. The 'Deep Learning HDL Toolbox' is specifically tailored to the needs of engineers looking to implement deep learning networks on FPGA hardware. Using the 'HDL Coder', the toolbox enables users to customize, create and deploy an IP core for deep learning processors. Support for standard networks and layers significantly increases the performance and flexibility of wireless applications.
The 'FPGA AI Suite' is characterized by the IP generation of user-defined AI inference accelerators on Altera FPGAs available at the push of a button. This is done using the 'OpenVino' toolkit, which is based on pre-trained AI models from widely used industry frameworks. In addition, FPGA developers can seamlessly integrate AI inference accelerator IP into the FPGA design using the industry-leading Quartus Prime Software FPGA flows. The combination of the Deep Learning Toolbox and the OpenVino toolkit provides developers with an efficient way to optimize AI inference on Altera FPGAs.










