Research Article
Improved Generalized Sparsity Adaptive Matching Pursuit Algorithm Based on Compressive Sensing
| Input: | | Sensing matrix | | Observation vector | | Initial step size | | Constant parameter | | Stop threshold | | Initialization parameter: | | {initialize signal approximation} | | {loop index} | | {initial sparsity estimate} | | {empty preliminary index set} | | {empty candidate index set} | | {empty support index set} | | {residual vector} | | {while loop flag} | | While (∼done) | (1) | Compute the projective set | | | (2) | Compute the estimated sparsity | | If | | Then return on step (1) | | Else , , , return on step (3) | (3) | Compute a new projective set | | | (4) | Merge to update the candidate index set | | | (5) | Get the estimate signal value by least squares algorithm: | | | (6) | Prune to obtain current support index set | | | (7) | Update signal final estimate | | | | residual error: | (8) | check the iteration condition | | If | | quit iteration | | else if | | | | | | | | else | | | | | | end | | end | | Output: ( s-sparse approximation of signal ) |
|