The main goals of this research are to enhance a commercial off-the-shelf (COTS) software platform to support unmanned ground vehicles (UGVs) exploring the complex environment of tunnels, to test the platform within a simulation environment, and to validate the architecture through field- testing.
In this research, software was designed and tested for a UGV in order to give it the capability to localize and map an unknown area while navigating. The UGV was equipped with a camera, acoustic sensor, and a laser range finder that provides the UGV with readings to determine its pose and landmarks within the unknown area to create a map of the unknown research.
The developed system was validated using simulation and field-testing to determine its ability to navigate the unknown environment, detect landmarks, and build a map of the navigated environment. Three scenarios were used for simulation testing. The simulated test environment was the STAGE simulator and the test sites were in facilities at the ERDC in Vicksburg, MS.
The following UGV platforms received consideration as possible platforms for this research: (1) the Adept Pioneer 3-AT, (2) the Coroware Explorer, and (3) Superdroid Robots HD2 Treaded ATR Tank Robot Kit. Each of these platforms is a programmable robot, easily adaptable to varied sensors.
The selection of the robot was based on several requirements.
It may be tracked or wheeled.
It must be capable of crossing objects of various sizes such as rocks, water, and gravel.
It must support a range of sensors.
The robot ultimately selected was the Coroware Explorer. The Explorer has 6-inches of ground clearance or more. It is equipped with two sensors: a two-megapixel color webcam to capture scenes and a laser range finder for detection of landmarks. It has a 2.0 GHz CPU, 1 GB of RAM, a 2.0 GHz CPU, a 4hr battery life, a dual-boot operating system (Ubuntu Linux or Windows), Wi-Fi, and a CUDA-capable main board. Its dimensions are 23 in. long, 21 in. wide, and 16 in. high with a weight of about 20 lb. The robotic software platform is the Robotic Operating System.
The Robotic Operating System (ROS), an Open Source UGV software, was selected for use with this project. ROS provides libraries and tools to help software developers create robot applications. ROS is not an operating system in the traditional sense of process management and scheduling; rather it provides a structured communication layer above the host operating system of a heterogeneous compute cluster.
ROS was designed to meet a specific set of challenges encountered when developing large-scale service robots as part of the STAIR project at Stanford University and the Personal Robots Program at Willow Garage, but the resulting architecture is far more general than service-robot and mobile-manipulation domains. The philosophical goals of ROS can be summarized as (1) peer-to-peer, (2) tools-based, (3) multi-lingual, (4) thin, and (5) free and open-source.
ROS provides services expected of an operating system, including hardware abstraction, low-level device control, implementation of commonly used functionality, message-passing between processes, and package management. It also provides tools and libraries for obtaining, building, writing, and running code across multiple computers. ROS currently only runs on Unix-based platforms. Software for ROS is primarily tested on Ubuntu and Mac OS X systems, although the ROS community has been contributing support for Fedora, Gentoo, Arch Linux, and other Linux platforms.
This work was done by Doris M. Turnage for the Army Engineer Research and Development Center. ERDC-0006
This Brief includes a Technical Support Package (TSP).
Localization and Mapping of Unknown Locations with Unmanned Ground Vehicles
(reference ERDC-0006) is currently available for download from the TSP library.
Don't have an account? Sign up here.