An efficient image pattern recognition system using an evolutionary search strategy
SMC'09 Proceedings of the 2009 IEEE international conference on Systems, Man and Cybernetics
Detection of protein conformation defects from fluorescence microscopy images
Engineering Applications of Artificial Intelligence
| Hi-index | 35.68 |
We introduce a computationally efficient detector of the space-time uncoded symbols for multiple-antenna systems. In a high-data-rate communication system, reliable real-time detection algorithm with low complexity is highly attractive. However, the space-time signal detection problem is mathematically equivalent to the combinatorial optimization problem, which is known to be NP hard. Essentially, most of the existing algorithms incorporate one or more of the following strategies: 1) sequential detection of signals in a lower dimensional space, 2) a posteriori probability density approximation, or 3) reduction of the signal searching space. The proposed detector explicitly incorporates all of these three strategies. The principal concept of the proposed detector is the exploration of the neighborhood for each transmitted symbol by using a signal decoupling technique and the expectation-maximization (EM) algorithm. The proposed detector provides a good balance between the computational complexity and the performance. Simulation results show that the symbol detection error rate of the proposed detector is close to the maximum-likelihood detector with reasonably low complexity