Skip to content

Latest commit

 

History

History
95 lines (71 loc) · 4.44 KB

README.md

File metadata and controls

95 lines (71 loc) · 4.44 KB
Raw

cwru_davinci_grasp

This package is the implementation of needle grasping for the Patient-side manipulators (PSMs) of the da Vinci® surgical robot. The needle grasp manipulation path is computed by class pick_place::PickPlan in MoveIt! library.

The current implementation allows user to either control the robot to grasp a needle with a set of defined grasping parameters or help user choose a valid grasp pose within a list of candidates grasp.

This package includes:

  • Simple needle grasper (library and main function) implemented by class pick_place::PickPlan in MoveIt! library.
  • Simple pose-based grasp generator implemented by optimal needle grasping algorithms.
  • A needle grasping data processing program.
  • A .yaml needle grasp config file, inside of which the grasping parameters are defined
  • Test code

Developed by Su Lu at the MeRCIS Lab, Case Western Reserve University.

Install

Ubuntu

Kinetic:

git clone https://github.com/cwru-davinci/cwru_davinci_grasp

Dependent packages

git clone https://github.com/JenniferBuehler/convenience-pkgs.git
git clone https://github.com/cwru-davinci/cwru_davinci_moveit.git

How to use

Before use, make sure the pose information of the needle is being published to the topic "/updated_needle_pose" in the form of "geometry_msgs/PoseStamped" ROS message.

To use main function

To use davinci_simple_needle_grasper_main function, user needs to specify which PSMs("psm_one" or "psm_two") is going to be used to grasp needle, the needle's name and to pick or place. The default setting is to use "psm_one" to pick the "needle_r". Then the main function will ask user to use which grasp mode to grasp the needle.

To pick up the needle:

roslaunch cwru_davinci_grasp davinci_simple_needle_grasp.launch which_arm:=psm_one needle_name:=needle_r movement:=pick

To place the same needle (Implementation In Progress):

roslaunch cwru_davinci_grasp davinci_simple_needle_grasp.launch which_arm:=psm_one needle_name:=needle_r movement:=place

To use code in user's program

Within your ROS project package, add this package to your package.xml, CMakeLists.txt. Then in the C++ implementation file add this to your includes:

#include <cwru_davinci_grasp/davinci_simple_needle_grasper.h>

Add to your class's member variables and properly initialize it correspondingly:

// plain version
DavinciSimpleNeedleGrasper simpleNeedleGrasper_;
// or smart pointer version in boost::shared_ptr
DavinciSimpleNeedleGrasperPtr simpleNeedleGrasperPtr_;

In your member initialization list of your class constructor:

simpleNeedleGrasper_(node_handle, private_node_handle, "psm_one" or "psm_two", "needle_name");

To have a private_node_handle, create a ros::NodeHandle initialized with "~", such as:

ros::NodeHandle private_node_handle("~");

Modes to grasp the needle

To pick up needle, you can either control the robot to grasp a needle with a set of defined grasping parameters or have the program find a valid grasp pose for you. There are four grasp modes to choose from, "DEFINED", "SUBOPTIMAL", "FINDGOOD" and "RANDOM".

To change the defined grasping parameters, change theta_normal value in file dvrk_psm_grasp_needle_data.yaml

// pick up needle by defined grasping parameters
simpleNeedleGrasper_.pickNeedle("needle_name", NeedlePickMode::DEFINED);

// or pick up needle by randomly searching through the list of candidates grasp
simpleNeedleGrasper_.pickNeedle("needle_name", NeedlePickMode::RANDOM);

// or pick up needle by sequentially finding the first valid grasp from the pre-sorted list of candidates grasp
simpleNeedleGrasper_.pickNeedle("needle_name", NeedlePickMode::SUBOPTIMAL);

// or pick up needle by sequentially finding the first valid grasp from the unsorted list of candidates grasp
simpleNeedleGrasper_.pickNeedle("needle_name", NeedlePickMode::FINDGOOD);

TODO

Future features to be added to this project:

  • Provides grasp visualization and GUI for user to explicitly select preferred grasp.
  • Considers the interaction between the gripper tip and soft tissue to minimize intrusion.
  • Finishes the place functionality.