The TurtleBot 2i Mobile Research Robot was introduced today by Interbotix Labs and the Open Source Robotics Foundation (OSRF). The TurtleBot 2i builds upon previous versions of the popular Robot Operating System (ROS) based robotics platform to provide advanced out-of-the-box capabilities to fast-track prototyping of next-gen robots.
The goal of the TurtleBot platform has always been to bring complex technologies, such as autonomous navigation and robotic manipulation, into the hands of innovators and developers in a more approachable format.
The TurtleBot 2i ($2,395.95) does a great job of just that with a redesigned chassis, native support for robotic arms and the Intel Joule 570x compute module.
TurtleBot 2i Offers Native Support of a Robot Arm
The TurtleBot 2i offers the Pincher MK3 5 DOF Robotic Arm as a standard option. This is the first time a TurtleBot robot has offered native support of a robotic arm.
TurtleBot 2i Robot Arm & Dual 3D Cameras
The MK3 allows the TurtleBot 2i to interact with small objects, buttons, and tools in its environment. The Arbotix-M Robocontroller provides an interface for the Pincher MK3 arm, which is implemented using MoveIt, an open-source solution that allows users to control the arm using only high-level commands.
TurtleBot 2i Powered by Intel Joule
Like the TurtleBot 3, the TurtleBot 2i uses the Intel Joule 570x compute module. Along with the powerful computing power, the Joule 570x allows the TurtleBot 2i to easily integrate with dual Intel RealSense 3D cameras. The TurtleBot 2i uses an Intel ZR300 RealSense camera for navigation and mapping, while the short-range SR300 RealSense camera is dedicated to the MK3 robot arm.
Autonomous Navigation & Point Cloud Mapping
“Today’s robotics developers have advanced requirements that are not met by current dev boards, limiting progress,” says Tom Seaman, product manager with Intel’s Modular Innovation Group. “Developers of tomorrow’s robotic solutions need a higher-performance, power-efficient compute solution to embed in their robots, capable of handling demanding workloads, including 3D simultaneous localization and mapping (3D SLAM), object recognition and tracking, face recognition, and natural language processing.”
The combination of the Joule 570x and RealSense also benefit from hardware GPU acceleration from the on-board Intel Iris graphics.
Previous versions of the TurtleBot used netbooks or laptops for computing power. Andrew Dresner, principal engineer at Interbotix Labs, says using the Joule 570x shaved about $500 off the computing cost of the TurtleBot 2i.
“The TurtleBot 2i is one of the cheaper mobile manipulator robots to ever hit the market,” he says. “We rebuilt the hardware on the TurtleBot 2i from the ground up, added a robotic arm as a standard option, with Intel Joule and Intel RealSense. This is an industry first at the price level.”
Map Zone Identification and Waypoints
“Before the joule came into play, we were looking at multiple DragonBoards – one to run a 3D camera and one to run the physical hardware,” Dresner continues. “Intel joule gave us enough computational power to not have to split the load over multiple boards, but it has all the features we needed.”
TurtleBot 2i Teleoperation
TurtleBot 2i Features
TurtleBot 2i ROS Software/Demo Features
- Autonomous Navigation & Point Cloud Mapping
- Map Zone Identification & Waypoints
- Robotic Arm Object Manipulation & Sorting
- Obstacle Avoidance & Path Planning
- Teleoperation Examples
- Self-Charging w/ Dock
TurtleBot 2i Hardware Specifications
- Intel Joule 570X
- Gumstix Nodana Carrier Board
- 4GB RAM
- 16GB eMMC Storage
- 802.11AC WiFi / Bluetooth 4.0
- Ubuntu 16.04 / ROS Kinetic
TurtleBot 2i Sensors
- SR300 RealSense 3D Camera
- ZR300 RealSense 3D Camera
- Edge Detection & Bumper Sensors
TurtleBot 2i Mobile Robot
- Kobuki Mobile Base
- Modular & Interchangeable Decks
- Pincher MK3 Robo Arm
- Arbotix-M Robocontroller
- Secondary 3S 4500mAh LiPo Battery
- Maximum translational velocity: 70 cm/s 13
- Maximum rotational velocity: 180 deg/s (>110 deg/s gyro performance will degrade)
- Payload: 2kg (without arm), 1kg (with arm)
- Cliff: will not drive off a cliff with a depth greater than 5cm
- Threshold Climbing: climbs thresholds of 12 mm or lower
- Rug Climbing: climbs rugs of 12 mm or lower
- Expected Operating Time: 4-6 hours (operating time varies depending on loadout)
- Expected Charging Time: 2-3 hours (charge time varies depending on loadout)
- Docking/Charging: automatic within a 2mx5m area in front of the docking station (sold separately)