Optimal BER-Balanced Combining for Weighted Energy Detection of UWB OOK Signals

For the demodulation of ultra-wideband (UWB) on-off keying (OOK) signals, some weighted energy detection (WED) schemes have been proposed. However, the schemes available suffer from unbalanced bit error rates (BERs) for bits "0" and "1". Therefore, in this letter we will address the optimal BER-balanced WED. Specifically, we incorporate the BER balance requirement into the WED formulation and obtain the optimal weighting vector by solving a nonlinear convex optimization problem. To bypass the computational complexity required by the convex optimization, we derive a simple but near-optimal weighting vector. Simulations illustrate that the proposed BER-balanced WED schemes can achieve almost the same BER performance as the maximum likelihood (ML) demodulator, which is optimal in terms of BER performance but produces unbalanced BERs for different bits.