Virtual Boundaries and Haptic Feedback for Arm Collision Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current collision avoidance systems for medical devices with repositionable arms, especially those operated via teleoperation, face challenges in predicting collisions accurately, often resulting in poor operator feedback and potential damage to the device or sterile field due to delayed detection and overly conservative predictions.
Innovation Solution
The system employs high-fidelity CAD and kinematic models to create virtual boundaries around the arms' joints and links, using a physics engine to apply feedback forces that prevent collisions by mapping overlap detection to haptic feedback through input controls, allowing operators to avoid collisions before they occur.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If circumscribing volumes are used for collision detection, then collision detection capability is improved, but the volumes are overly large causing interference with arm operation and requiring larger separation distances
Solution Approach 1:
The patent segments the circumscribing volume into multiple smaller convex primitives (spheres, cylinders, boxes) that are attached to individual links and joints of the repositionable arm. This segmentation allows for more precise collision detection while reducing the overall volume occupied by the safety envelope, thereby improving both collision detection capability and arm operation freedom.
Solution Approach 2:
Different convex primitives are strategically placed at different locations (links and joints) with varying sizes and shapes according to the local geometry and collision risk. This local quality approach ensures that collision detection is most stringent where needed while allowing greater operational freedom in less critical areas, resolving the contradiction between detection reliability and operational ease.
2Reliability
If separate workspaces are provided for each repositionable arm, then collision avoidance is improved, but the workspace utilization efficiency deteriorates
Solution Approach 1:
The patent implements dynamic collision detection and avoidance where the virtual boundaries and convex primitives are continuously updated based on the real-time positions of the repositionable arms. This dynamic approach allows arms to operate in closer proximity than static separate workspaces would permit, improving workspace utilization while maintaining collision avoidance through active monitoring and adjustment of safety envelopes.
3Reliability
If large circumscribing volumes are used, then collision detection coverage is improved, but the precision of collision prediction deteriorates
Solution Approach 1:
By dividing the overall collision detection envelope into multiple smaller convex primitives distributed across different links and joints, the system achieves both comprehensive coverage and high precision. Each small primitive provides precise local collision prediction, while the collection of primitives together ensures complete coverage of the entire arm structure.
Solution Approach 2:
The patent employs three-dimensional convex primitives (spheres, cylinders, boxes) that accurately represent the spatial extent of each link and joint. This dimensional approach allows for precise volumetric collision detection rather than relying on two-dimensional projections or simplified distance calculations, thereby improving both coverage and prediction precision simultaneously.
Data Source
AI summary
A system and method of collision avoidance includes determining a position and an orientation, the position and the orientation being of a repositionable arm or of an instrument, the repositionable arm being configured to support the instrument; determining, based on the position and the orientation, a plurality of first virtual boundaries around the repositionable arm or the instrument; determining a second virtual boundary around an object; determining a first overlap force on the repositionable arm due to a first overlap between the second virtual boundary and a virtual boundary of the plurality of first virtual boundaries; determining a tip force on a distal end of the instrument based on the first overlap force; and applying the tip force as a first feedback force on the instrument or the repositionable arm.