In this paper, a novel approach for automated suturing is introduced and experimental results are presented. Unlike other similar works, the proposed approach adopts a single arm to implement a suturing task with a standard laparoscopic needle holder and curved suture needle. 3D information is obtained from a clinical (single camera) endoscope through an elliptical/circular pose measurement algorithm, which dynamically tracks the suture needle and surface markers. This drives robotic needle steering through a set of surgeon-defined entry/exit points on a tissue pad phantom. Implementation results indicate good depth resolution (1.5mm) and task repeatability (85%) for a variety of consistency, lighting, and location variation scenarios.