Can you define minimum separation distance to obstacles for robotics toolbox Rapidly-exploring Random Tree (RRT) path planner?
18 visualizaciones (últimos 30 días)
Mostrar comentarios más antiguos
Connor Blake
el 1 de Sept. de 2022
Comentada: yovel atia
el 11 de Abr. de 2023
I am using the robotics toolbox and working with a obstacle-heavy environment for path planning. For safety reasons, I would like to specify the minimum separation distance from any object in a valid path of the robotic arm to be some value around 10 cm. I know I can plan a collision free path with the rapidly-exploring random tree (RRT) and then later validate using checkCollision that a given path has a minimum separation distance with all obstacles, but is there a way to enforce this constraint in the RRT algorithm? Checking hundreds of RRT paths and then rejecting them to be run again if they violate the separation condition takes too long to be viable. Should I just inflate the STL scales so that this condition is incorporated into the RRT algorithm by redefining what it means to collide?
0 comentarios
Respuesta aceptada
Karsh Tharyani
el 7 de Sept. de 2022
Editada: Karsh Tharyani
el 7 de Sept. de 2022
Hi Connor,
You can try overriding the "isStateValid" routine of a "manipulatorCollisionBodyValidator" by subclassing from it. Your overriden "isStateValid" in your subclass could perform "collisionChecking", evaluate the sepration distance, and if within acceptable threshold, accept or reject a state.
Refer to a pattern, which overrides the "interpolate" for a "manipulatorStateSpace", in this example: https://www.mathworks.com/help/robotics/ug/plan-paths-with-end-effector-constraints.html
I hope this helps.
Karsh
1 comentario
yovel atia
el 11 de Abr. de 2023
Hi, did you manage to define a minimum distance from obstacles? If so, can you share how?
Más respuestas (0)
Ver también
Categorías
Más información sobre Robotics en Help Center y File Exchange.
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!