Dynamic Active Constraints for Surgical Robots using Vector Field Inequalities
This work addresses safety challenges in deep brain and endonasal surgeries by enabling real-time collision avoidance for surgical robots, though it is incremental as it extends an existing method.
The authors tackled the problem of collision risk in constrained surgical workspaces by extending a vector-field-inequalities method to provide dynamic active constraints for multiple robots and moving objects, resulting in optimal trajectory error in simulations and successful autonomous collision prevention in real robotic experiments.
Robotic assistance allows surgeons to perform dexterous and tremor-free procedures, but robotic aid is still underrepresented in procedures with constrained workspaces, such as deep brain neurosurgery and endonasal surgery. In these procedures, surgeons have restricted vision to areas near the surgical tooltips, which increases the risk of unexpected collisions between the shafts of the instruments and their surroundings. In this work, our vector-field-inequalities method is extended to provide dynamic active-constraints to any number of robots and moving objects sharing the same workspace. The method is evaluated with experiments and simulations in which robot tools have to avoid collisions autonomously and in real-time, in a constrained endonasal surgical environment. Simulations show that with our method the combined trajectory error of two robotic systems is optimal. Experiments using a real robotic system show that the method can autonomously prevent collisions between the moving robots themselves and between the robots and the environment. Moreover, the framework is also successfully verified under teleoperation with tool-tissue interactions.