Increasing precision of the Raven-II surgical robot by applying cascade control

Automated execution of surgical procedures using a robot has the potential to facilitate training of novice surgeons and reduce workload. However, many current surgical robots are not built for autonomous task execution, but manual teleoperation, which means that they are often imprecise. This also applies to the Raven-II surgical robot.

This paper explores a cascade control scheme applied to the Raven-II. This scheme is beneficial in the face of the significant dynamics of the robot. The hypothesis is that this alternative control scheme, compared to the previous method, will reduce the trajectory-following error and improve robot precision.

To compare controller performances, an initial qualitative experiment is carried out with the three positioning joints of the robot, where individual joints are each commanded to move along smooth reference trajectories in joint-space. A more quantitative measure is found in a second experiment in which the control methods are compared by looking at their ability to follow the same predefined path between 500 randomly generated tool-space configurations.

The cascade control scheme is shown to better achieve tight control, reduce trajectory-following error and improve robot precision even with less-than-optimally tuned gain parameters. The improvement is modest when compared to the error in the robot state estimate.