A Real-Time System for Detecting Couch Collisions During Radiotherapy Treatment Delivery
L Yuan1*, J Patrick2, E Klein1, Z Saleh1, (1) Brown University, Providence, RI, (2) London Regional Cancer Program, London, ON, CA
Presentations
TU-D1030-IePD-F6-3 (Tuesday, 7/12/2022) 10:30 AM - 11:00 AM [Eastern Time (GMT-4)]
Exhibit Hall | Forum 6
Purpose: The collision of the treatment couch and other nearby objects on the floor inside the vault can cause treatment interruption and may lead to expensive repairs. This study is aimed to develop a robust and affordable solution to detect such collisions during treatment delivery.
Methods: Two laser range scanners (PRLIDAR A2, SLAMTEC) were mounted horizontally at the distal corners of the treatment couch. The scanners have a 12-meter range with a distance resolution of 0.2 cm, and able to detect objects within 360 degrees. The angle and distance between the scanner and the object were collected and processed by an in-house Python script in real-time. The collision zone is simply defined as a space within 10cm from the couch edge. To exclude the couch itself, the viewing angle of the scanner was limited to a certain range based on the direction of the scanner. Objects that were detected within the collision zone would trigger an alarm sound immediately to alert the therapists.
Results: The laser range scanners were able to detect objects that appeared within the collision zone when the couch was stationary or in motion. The alarm sound was triggered until the object was cleared. The angle and the distance were recorded in a log file for further investigation. The variance of the measured distances and angles were ±3 cm / ±2° respectively.
Conclusion: We implemented a real-time collision detection system using 2D laser range scanners and an in-house Python script. The proposed system is cost-effective and can operate with minimal user intervention.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email