UTDD UTRA TDD Data Detector for TD-SCDMA and TD-CDMA 


Athena delivers the UTDD, a complete, scalable solution for TD- SCDMA and TD-CDMA. Leveraging Athena’s compact and scalable Complex Matrix Arithmetic (CMA) processor to achieve unprecedented code density and performance per unit area, the UTDD is a drop-in solution to realize TD-SCDMA or TD-CDMA processing in your next communications design.

The UTDD is available in configurations ranging from 1 to 8 cores and supports up to 8 antennas. Depending on the application, performance and area can be scaled for implementations in single-antenna handsets to multi-antenna, multi-cell base stations.

Product Description

Athena’s UTDD accepts slot data and internally performs multi-user joint data detection using zero-forcing block linear equalization (ZF-BLE) or minimum mean-square error block linear equalization (MMSE-BLE). ZF- BLE is implemented using either approximate Cholesky (ACZ) or block Fourier (BFZ) methods. MMSE-BLE is also implemented using either approximate Cholesky (ACM) or block Fourier (BFM) methods.

Data detection requires channel estimation, which can either be supplied externally or performed by the UTDD with an optional firmware module. For internal CIR estimation, the UTDD uses an efficient implementation of the Steiner estimator and optional post-processing with thresholds and CIR history.

Both single and dual estimated channel impulse responses (CIRs) per slot are supported in both external and internal channel estimation configurations. The dual CIR configuration provides increased signal detection performance with additional computational complexity compared to the single CIR configuration.

In addition to the standard UTDD firmware, Athena can work with customers to provide additional custom firmware functionality utilizing UTDD’s underlying CMA processor. With a full instruction set for high-level complex matrix operation such as matrix multiplication and convolution, the CMA can offload many advanced signal processing and communications algorithms.