An Active Steering Hand-held Robotic System for Minimally Invasive Orthopaedic Surgery Using a Continuum Manipulator

Abstract

This paper presents the development and experimental evaluation of an active steering hand-held robotic system for milling and curved drilling in minimally invasive orthopaedic interventions. The system comprises a cable-driven continuum dexterous manipulator (CDM), an actuation unit with a handpiece, and a flexible, rotary cutting tool. Compared to conventional rigid drills, the proposed system enhances dexterity and reach in confined spaces in surgery, while providing direct control to the surgeon with sufficient stability while cutting/milling hard tissue. Of note, for cases that require precise motion, the system is able to be mounted on a positioning robot for additional controllability. A proportional-derivative (PD) controller for regulating drive cable tension is proposed for the stable steering of the CDM during cutting operations. The robotic system is characterized and tested with various tool rotational speeds and cable tensions, demonstrating successful cutting of three-dimensional and curvilinear tool paths in simulated cancellous bone and bone phantom. Material removal rates (MRRs) of up to 571 mm³/s are achieved for stable cutting, demonstrating great improvement over previous related works.

Keywords: Compliant Joints and Mechanisms; Medical Robots and Systems.

Fig. 1.

a) Continuum dexterous manipulator used for experiments, b) flexible cutting instrument developed for cutting of bone.

Fig. 2.

a) Robotic system assembled in its hand-held configuration, with the CDM, FCI, and handpiece installed. DOF of the actuation unit are highlighted in blue and b) Hand-held system milling a Sawbone sample.

Fig. 3.

Block diagram of the control system showing input and output.

Fig. 4.

Experimental setup for the testbed configuration, with the CDM, FCI, and linear stage. The blue arrow represents the direction of the feed velocity while the green arrow represents the direction of micrometer feed.

Fig. 5.

a) Experimental setup for system identification and b) experimental setup for milling, showing the starting position of the CDM relative to the Sawbone sample.

Fig. 6.

a) Experimental setup for testbed curved drilling and b) experimental setup for hand-held curved drilling, simulating core decompression of the femoral head.

Fig. 7.

a) Time-domain system response to sinestream cable tension setpoint with untuned PD controller. b) Step response of the system with a tuned PD controller, comparing experimental and simulated results. c) Cable tension variability data for different cable tension setpoints and FCI rotational velocities. Points represent the average cable tension for each trial, and error bars represent the variability of the time-domain data. d) Time-domain cable tension data for bending of the CDM with different cable displacement velocities.

Fig. 8.

a) Material removal rate for each combination of FCI rotational velocity, Sawbone density, and cable tension setpoint using a PD controller for milling. b) Time-domain cable tension data (unfiltered) for milling of 10 PCF Sawbone using a PD controller. c) Noise-to-signal ratio for each combination of FCI rotational velocity, Sawbone density, and cable tension setpoint for both PD control and velocity control milling experiments.

Fig. 9.

Cross-sectional view of the drilling experiment samples.

Fig. 10.

Demonstrations from the hand-held experiments, consisting of a a) overhead view of the block milling result, b) cross-sectional view of the block drilling result, c) overhead view of the block drilling trial, d) cross-sectional view of the femur phantom result, e) piecewise bending with two drive cables, and f) S-shape bending with two drive cables.

Fig. 11.

Time-domain cable tension data (filtered) for the hand-held cutting experiments.