TU Delft EMI Lab: AI-native digital front-end, neural DPD and RF impairment compensation for energy-efficient FR3 6G FEM
TU Delft EMI Lab: AI-native digital front-end, neural DPD and RF impairment compensation for energy-efficient FR3 6G FEM
TU Delft EMI Lab offers expertise in AI-native digital front-end techniques for energy-efficient 6G front-end modules. We contribute neural digital predistortion (DPD), PA/FEM behavioural modelling, adaptive RF impairment compensation, low-complexity AI implementation, and measurement-validated benchmarking for wideband RF systems.
Our group has developed OpenDPD (https://github.com/Lab-EMI/OpenDPD), an open-source end-to-end learning and benchmarking framework for wideband PA modelling and DPD, together with follow-up work on mixed-precision, temporal-sparsity-aware, TCN/sparse and accelerator-oriented neural DPD. We also bring expertise in hardware-software co-design for edge AI, FPGA/ASIC-oriented implementation, quantisation/sparsity, and energy-efficient recurrent neural network accelerators.
We are interested in joining an industry-led SNS JU 2026 FEM consortium as an academic partner, potentially contributing to or leading a task/work package on AI-assisted digital front-end design, energy-efficient DPD, adaptive FEM control, hardware-aware implementation, and measurement validation for FR3/wideband 6G scenarios. We are particularly interested in collaboration with partners providing FEM/PA/array/packaging test vehicles, RF measurement infrastructure, system requirements, and industrialisation pathways.
- STREAM-B-01: Collection, Generation and Validation of Datasets suitable for training AI Models for 6G Networks and for AIaaS
- STREAM-C-01: SNS experimental infrastructure
- STREAM-CSA-02: 6G Devices
- 6G Access Networks
- Air Interface Innovations
- Experimental Infrastructure
- Innovative 6G Technologies
- Multi-Antenna
- Novel Radio System Architecture
- Spectrum
- Sustainability
- Wireless Innovations
Contact
Company: Delft University of Technology (TU Delft)
Type of Organisation: University
Country: Netherlands
Web: https://www.tudemi.com
Telephone:
Brief description of my Organisation
Delft University of Technology (TU Delft) is a leading technical university in the Netherlands. The Lab of Efficient Machine Intelligence (EMI Lab) in the Department of Microelectronics develops hardware-software co-design methods for energy-efficient AI and RF/mixed-signal systems. The group works on neural DPD, AI-assisted RF and mixed-signal correction, edge AI accelerators, FPGA/ASIC-oriented implementation, and measurement-driven benchmarking. The lab is led by Dr. Chang Gao in the Faculty of Electrical Engineering, Mathematics, and Computer Science.
Please complete the form to contact this Expertise Offer
Expertise Offering
