Decoupling Path Following and Velocity Profile in Vision-Guided Laser Steering

Abstract

Laser surgery requires accurate following of a path defined by the surgeon, while the velocity on this path is dependent on the laser-tissue interaction. Therefore, path following and velocity profile control must be decoupled. In this paper, nonholonomic control of the unicycle model is used to implement velocityindependent visual path following for laser surgery. The proposed controller was tested in simulation, as well as experimentally in several conditions of use: different initial velocities (step input, successive step inputs, sinusoidal inputs), optimized/nonoptimized gains, time-varying path (simulating a patient breathing), and complex curves with curvatures. Thereby, experiments at 587 Hz (frames/s) show an average accuracy lower than 0.22 pixels (∼10 μm) with a standard deviation of 0.55 pixels (∼25 μm) path following, and a relative velocity distortion of less than 10 −6 %.

Automatic Noncontact Extraction and Independent Manipulation of Magnetic Particles Using Electromagnetic Needle

“…The desired velocity vector is planned using the control law presented in [38]. The kinematics of the particle can be expressed in a mobile frame attached to the orthogonal projection, p s , of the particle on the path (Fig.…”

“…where ω =θ p is the rotational velocity of the particle motion vector, s the curvilinear abscissa on the path, C(s) the curvature of the path, d the distance with the path and θ e is the angle between the particle velocity vector and the tangent vector to the path in p s . This system can be transformed into a chained system using a change of variables (the complete derivation can be found in [38]). Then, proportional state feedback which ensures convergence of the distance and the orientation errors can be achieved:…”

“…The control variable u is linked to the rotational velocity ω, which updates the velocity vector at each iteration [38]:…”

Some Examples of Path Following in Microrobotics

“…A similar problem was already solved in the literature for automated guided vehicles, where an additional non-holonomic constraint is at stake: the vehicle can not move sideways. In order to have smooth motion of the laser (and to avoid zig-zaging scars), we deliberately applied the same nonholonomic constraint to the laser, although it was steered by a mirror and was physically not subject to such a constraint [3].…”

Eye-in-Hand Visual Servoing of Concentric Tube Robots

“…To achieve medical tasks, CTR can be teleoperated through a surgeon-robot interface or closed-loop controlled using exteroceptive sensors as stereo camera [7], magnetic resonnance imaging [8], computed tomography [9], ultrasounds [10], or more recently Optical Coherence Tomography (OCT) [11] in an eye-to-hand configuration (when the camera views simultaneously both the robot and the scene). Some tasks require visual feedback to perform an automatic repetitive biopsy to follow-up cell evolution [12], achieve path following in the image [13], and keep a tool in the field-of-view [10], [14] during interventions. In addition, the development of surgical robotic systems with two CR is increasing [15], [16].…”