Text this: Towards Robust Speech Models: Mitigating Backdoor Attacks via Audio Signal Enhancement and Fine-Pruning Techniques