Room: Karl Dean Ballroom B1
Purpose: Robotic stabilization of the patient's head 6D position (both translational and rotational) during frameless maskless stereotactic radiosurgery requires positional deviations to be corrected in real-time. As the radiation beam is on during the correction process, it is mandatory that the robot trajectory is optimal both spatially and temporally in order to minimize unintentional exposure of healthy brain tissue. In this work we propose a novel 6D trajectory planning method for such control problem.
Methods: For a given measured target (tumor) position with a displacement away from the desired setpoint (isocenter), the required robot end-effector position for returning the target back to the setpoint is first computed. Then, the conventional method will send the end-effector back to the required position as fast as possible without regarding what the corresponding target trajectory is followed. The proposed trajectory planning, on the other hand, formulates the problem as real time optimization problem and find a trajectory with the steepest descent of target position error in each step, and thus minimize irradiation. Refer as target-D trajectory.
Results: The method was extensively tested using a simulated robotic SRS system with both synthetic motion and prior recorded volunteer head motion. In all cases it was found that target-D trajectory was optimal over conventional correction. For nine recorded volunteers head motion, the corrected target by using conventional method and proposed method held the target within the threshold (0.2mm/0.15deg) 94.2% and 99.1% of the time, respectively.
Conclusion: A general 6D target trajectory planning framework for robotic SRS was investigated. The motion control was formulated as an optimization problem and solved at real-time using the L-BFGS algorithm. The method was found to be flexible as it allows control over various performance requirements such as mechanical robot limits, velocities, acceleration, or other desired parameters.