No abstract

This paper aims to present the integrated design, development, and testing procedures for a state-of-the-art torsion-based series elastic actuator that could be reliably employed for long-term use in force-controlled robot applications. The main objective in designing the actuator was to meet weight and dimensional requirements whilst improving the long-term durability, ensuring high torque output, and containing its total weight. A four-fold design approach was implemented: (i) following recursive design-and-test procedures, an optimal torsional spring topology was unveiled with the help of SIMP (Solid Isotropic Material with Penalization) topology optimization method, (ii) the proposed spring was manufactured and multiple specimens were experimentally tested via a torsional test machine to validate linearity, loading rate response, and mechanical limits, (iii) the actuator’s thermal response was experimentally scrutinized to ensure that the generated heat was dissipated for long-term use, and (iv) the fatigue life of the spring was computed with the help of real-life experiment data. Having concluded the development and verification procedures, two different versions of the actuator were built, and preliminary torque control experiments were conducted. In conclusion, favorable torque tracking with a bandwidth of 19 Hz was achieved while peak-to-peak torque input was 20 Nm.

Robotic systems are being applied to medical interventions as they increase the operational accuracy. The proposed autonomous and ultrasound guided 5-DOF parallel robot can achieve such accuracy for needle biopsies, which particularly demand precise needle positioning and insertion. In this paper, the robot’s mechanical design, system identifications, and the design of its controller are explained. A torque computed controller with gravity compensation and friction models, yielding a 0.678[Formula: see text]mm RMS position error for the needle tip, was used. A novel method was used for 3D space calibration of the images for detecting the volume of interest in the biopsy procedure by a multipoint crosswire phantom with parallel threads. The calibration technique had a validation RMS error of 0.03[Formula: see text]mm.

No abstract

No abstract