Sawyer Research Robot

Sawyer
Status Active

Superusers

  • Ivan Penskiy (ipenskiy[at]umd[dot]edu)

Important documents and useful links

Overview

Sawyer from Rethink Robotics is a robot designed for research in human-safe robotic manufacturing and human-assistive robotics. It has one 7-axis arm, integrated cameras, torque sensors, and direct programming access via a standard ROS (Robot Operating System) interface.

Software setup

Sawyer is connected to the lab's LAN with a static IP. The controlling PC has Ubuntu 16.04 installed with ROS Kinetic running.

Location

Brendan Iribe Center 432, Room 0116.

Physical specifications

Maximum Reach 1260 mm
Gantry Reach 900 mm
Task Repeatability +-0.1 mm
Degrees of Freedom 7
Maximum Payload (Including End Effector) 4 kg

Reservation

Sawyer must be reserved before using. The reservation is done using the Google calendar (IRB-0116sa). You will be able to reserve the equipment personally after attending the RML lab orientation and then following the steps below to gain access to the equipment.

Video demonstration

The following videos demonstrate capabilities of Sawyer:

Sawyer: research and education Sawyer: examples from industry Sawyer: lecture on ROS and Sawyer in RRL

Note: these are promotional and instructional videos. They are included for demonstrational purposes only. The details of actual operational procedures in RML can differ!

Important reminders

  • While this is designed as a human-safe robot, it is still a big robot that can do a lot of damage. Make sure the red emergency stop button is accessible at all times.
  • Be generally careful when working with Sawyer.

Gaining access to equipment

  • Preparation
  • Scheduling
  • Instruction session
Preparation:

Sawyer has several well defined and structured tutorials, consequently RML expects users to be familiar with Sawyer robot specs and capabilities as well as its basics of operation. Specifically, users should have knowledge of the following topics:

  • ROS framework, one should go through the ROS tutorials for basic understanding.
  • Sawyer uses python classes for the control of the robot. These classes are built on top of the ROS API and understanding the python classes would be essential for working on Sawyer. The breakdown on how these classes are functioning could be found at the following link: Sawyer API References
  • Gazebo simulation and control using MoveIt.
Scheduling:

Once you are confident in the topics from the previous section, email the lab manager (Ivan Penskiy, ipenskiy[at]umd[dot]edu ) to schedule the instruction session (30-45 min).

Instruction session:

In the beginning of the session the user will be asked several questions to test the knowledge of Sawyer and the basics of ROS framework. The user also will have to demonstrate a Gazebo simulation of Sawyer. The following are the deliverables for the simulation:

  • Spawn Sawyer in a map in Gazebo (you can use standard maps from the tutorials).
  • Command Sawyer to perform a simple action.
  • Run Rviz in parallel to Gazebo and show the real-time visualization of the simulation in Rviz.

After confirming knowledge and skills required to operate Sawyer, the user will be instructed on rules and technical aspects of using Sawyer in RML (network connection, turning ON and OFF procedures, etc).

Workspace setup

Sawyer Workstation Setup

Sawyer Networking

Shutting down procedures

  • Bring arm in the neutral position and disable Sawyer
  • Make sure the end effector is not in the close vicinity to the table surface or other foreign objects
  • Turn off Sawyer using power button on the controller box