Program Information
A 6DOF Optimal Control Algorithm for Frameless Maskless Robotic SRS Systems That Is Robust Against Setup Errors
X Liu*, AH Belcher, R Wiersma, The University of Chicago, Chicago, IL
Presentations
TH-EF-FS1-1 (Thursday, August 3, 2017) 1:00 PM - 3:00 PM Room: Four Seasons 1
Purpose: For automated robotic correction of patient motion during radiation treatment it is critical that the robot moves the target following a shortest path trajectory in 6DOF space. Failure to do so will lead to unnecessary exposure of healthy tissues to high doses and poor tumor conformity. In this work we developed an optimal control law that performs shortest path real-time 6DOF corrections and is robust against both user setup errors and robot accuracy limitations.
Methods: The motion control algorithm was designed based on the head position in the end effector frame, which was updated according to the real-time measured head position in LINAC frame. An optimal end effector trajectory was selected so that the head moved back to the desired position with minimum overall target position errors. To evaluate the robust performance of the control algorithm a computer model was built. The model included uncertainties of the hexapod setup translational/rotational errors and the manufacturer errors in the joints position and leg lengths. Prior record real-time 6DOF head motion data from an IRB volunteer study was used as input.
Results: Without robotic head stabilization it was found that over a simulated SRS 15 minute treatment time the volunteer was only within 0.2mm/0.05degree position tolerance in 1% of time. However, with robotic stabilization and no setup/manufacturer errors this was greatly improved, the head position was within tolerance in 99.3% of the time. Surprisingly, in the case of large hexapod setup position errors 20mm/5degree and hexapod joint position/leg length errors 1mm, the corrected head position remained within tolerance in 98.7-99.3% (average 99.2%) of the time.
Conclusion: A 6DOF optimal control law for robotic SRS system was reported, and it was showed that the control performance was insensitive to large setup/manufacturer errors.
Contact Email: