1. DESCRIPTIVE TITLE:Transportable Manned and Robotic Digital Geophysical Mapping (DGM) Tow Vehicle

2. ESTCP THRUST AREA:1) Unexploded Ordnance (UXO) Detection, Discrimination, and Remediation

3. LEAD ORGANIZATION: U.S. Army Corps of Engineers Engineering & Support Center-Huntsville (USAESCH), PO Box 1600, Huntsville, Alabama 35807-4301; Scott Millhouse, PE, ED-CS-D phone 256-895-1607 fax 256-895-1602 email:

4.Abstract:

a.) Objective: This project will integrate an innovative robotic tow vehicle with industry standard Digital Geophysical Mapping (DGN) sensors and advanced geo-location positioning equipment to autonomously map target demonstration areas. Phase I will focus on integration and path following to precisely replicate target coverage for multiple runs with multiple sensors. Phase II will continue on phase I developments and focus on more challenging site conditions that require obstacle recognition and avoidance and a secondary positioning system to maintain path when the primary positioning system is shadowed.

b.) Technology Description: The principle DGM tow vehicle is the Segway Robotic Mobility Platform (RMP) in both a two wheel RMP 200 and 4 wheel RMP 400 ATV configurations. This vehicle is battery powered and can be quickly broken down for express package shipment like the typical DGM sensors. Phase I positioning is provided to sub-centimeter accuracy by the ArcSecond Laser “Indoor GPS” triad package. Phase II as envisioned will utilize DGPS with INS, electronic magnetic compass, and dead reckoning solutions for positioning and a form of synthetic vision using 2-D LIDAR for obstacle recognition. Phase II will be demonstrated in autonomous, semi-autonomous and tele-operated modes.

c.) Expected Benefits: The robotic solution will permit more precise path following than by man towed equipment. The immediate benefit for sensor development and demonstration/prove out sites is to cover the exact same pathway multiple times with changed parameters to the same or with alternate sensors. For sites with explosive or safety concerns the robotic DGM vehicle will eliminate personnel risk. Operation will be monitored remotely. The geophysicist will be no longer be a “beast of burden” towing equipment. He can focus his efforts to monitor and acquire the highest possible quality data. It is expected that not only will quality dramatically improve but also so will productivity.

5.Problem Statement:

Unexploded Ordnance (UXO) poses a threat to both human life and the environment. Millions of UXO are located in the U.S. on active test and training ranges and Formerly Utilized Defense Sites (FUDS). In addition to the millions of UXO, there are many times more cultural and debris anomalies. Digital Geophysical Mapping (DGM) is used to map the areas and to locate, identify and select the items for sampling and removal. Many modes of DGM are utilized that include airborne and ground based platforms such as large towed arrays and man-portable equipment. All sites will need some amount of ground based mapping by man-portable or narrow width towed platforms depending upon terrain, ground cover and UXO objective. Man portable equipment essentially uses the operator as a “beast of burden” to carry the electronics and batteries and tow or swing the sensors in addition to monitoring the equipment and maintaining track. This can lead to reduced data quality due to deviation from pathway and inadequate sensor and position monitoring. Fatigue can cause reduced production and inattention to safety concerns. This proposal will initially test and develop an easily transportable battery powered tow vehicle that has been measured and documented for its effect on typical geophysical sensors.

New sensors, analysis techniques and field methodology are normally selected by a geophysical prove out process. Typically an area is seeded with the full range of expected Ordnance and Explosive items and clutter at a range of depths as a site specific evaluation or as part of a technology validation such as at the Yuma or Aberdeen Proving Grounds Standard UXO Technology Demonstration Sites. For the best comparisons the individual sensors should cover the same pathways as close as possible. Analysis algorithms are predicted to require a position accuracy of better than +/- 2 cm in x, y and z position to be effective. The next step is a robotic enhanced tow vehicle that will utilize the ArcSecond Indoor GPS positioning system to maintain the position accuracy and accurately maintain pathway track. All sensor suites being considered in evaluations will traverse the site by being robotically towed to the same pathway. This will facilitate analysis algorithm development and equipment performance validations.

Typical field mapping areas have varied terrain, obstacles, line-of-sight and horizon visibility. This adds challenges to maintaining a pathway and for maintaining sensor positioning. Semi-autonomous, autonomous and tele-operated mapping operations are desired for enhanced quality, personnel safety and production. Travel pathways can be designed to minimize interferences and for maximum production with the objective to survey all areas that are currently being covered by man portable towed sensors. For the most flexibility and production, autonomous obstacle recognition and avoidance and a methodology for maintaining track and positioning will be required for areas with the primary geo-location positioning system is shadowed.

6. Technology Description

a) Technology objectives. We propose to develop the Segway® Human Transporter (HT), Cross-Terrain Transporter (XT) and the two wheeled and four wheeled versions of the Robotic Mobility Platform (RMP),RMP 200 and 400, as a man powered DGM and robotic narrow towed array DGM tow vehicle. The equipment is planned for full performance testing as a manned tow vehicle to define capabilities and limitations and then augmented for two and four wheeled operation for tele-operated, semi-autonomous and autonomous robotic operation. Our objective is a small, economic, environmentally clean easily shippable tow vehicle. Phase II will add capability for more autonomous operation for larger less controlled locations by including obstacle recognition and avoidance and positioning and path following for when the primary positioning system is shadowed. Phase II will provide the capability to survey all areas that are typically surveyed by man portable towed sensors.

b) Technology description.The Segway® Cross-Terrain Transporter (XT) is the latest self-balancing human transporter that provides enhanced performance on a variety of terrain with minimal environmental impact. Featuring all-terrain tires, new extended-range lithium-ion batteries and specially tuned software, this rugged Segway XT will go practically anywhere you want to go.

The Segway XT has been specifically developed for stability, comfort and performance on uneven and tough terrain. The Segway XT's low-pressure tires smooth the ride on bumpy surfaces and minimize trail impact, while the wider track increases stability on uneven ground. The Segway XT's software has been modified to support the new tire size and to supply improved control and performance. The increased energy capacity of the standard lithium-ion batteries support the demands of the Segway XT, while still providing a 10 mile off-pavement range, depending on terrain, riding style and payload.

Segway® Human Transporter (HT) can self-balance because of a technology calleddynamic stabilization. Dynamic stabilization works in much the same way your own sense of balance does. Where you have an inner ear, eyes, muscles, and a brain to keep you balanced, the Segway HT has solid-state gyroscopes, tilt sensors, high-speed microprocessors, and powerful electric motors performing to keep it balanced. Working in concert, these extensively tested, redundant systems sense your center of gravity, instantaneously assess the information, and make minute adjustments one hundred times a second. Segway HT balances whetheryou're traveling at 10 mph, carrying a heavy load, slowly maneuvering in tight spaces, or standing perfectly still. See Appendix B for additional details on the technology behind the Segway.

The Robotic Mobility Platform (RMP) version of the HT is available in the typical 2 wheel mode as well as a dual unit with 4 wheels. The 4 wheel version provides the additional stability and traction for rough terrain applications of 4 individually powered wheels. This platform is based upon the mainstream commercial product except that it provides output to a generic PC external command control computer and takes instruction from that computer for maneuvering instead of relying upon the commercial product’s manual control input. Instead of having a balancing rider providing direction by leaning or turn commands by the operation of a twist grip, the command computer communicates with the HT by the CANbus interface at 100Hz. The unit comes with a standard joystick mode for remote operation as well as a command set for the unit and the decode for the state variables as sent by the RMP to the computer for robotic control.See Appendix B for additional details on the RMP.

The proposed Segway solutions meet our objectives for a small, economic, environmentally clean easily shippable tow vehicle for DGM.

c) Technology Maturity. The HT has been commercially available since March 2002 with many thousand sold and supported worldwide. The Robotic Mobility Platform (RMP) has been available since 2003 and is part of an active DARPA program with 12 University partners to develop to support military missions. The RMP is sold primarily to Labs, Universities and Government for robotic development with over 45 currently fielded.

d) Technical Approach. We propose the following program:

I. Pre-project- Segway HT testing (Self performed by the HuntsvilleCenter at no cost to ESTCP -value $20k)

  • A current-off-the-shelf used refurnished HT was purchased by USAESCH and tested for the static and the dynamic effect of the ferrous and non-ferrous components and for production of Electromagnetic (EM) fields.

  • First an area was established that was clear of subsurface anomalies. Then an EM-61 was set up and zeroed to background with a non-metallic tape stretched out from the center of the coil. The HT was then moved powered off towards the EM with EM measurements taken at 1’ intervals from 20’ to 2’ from the coil center. The test was done several times with the equipment powered off. The test was then performed several times once again with the HT manned and powered up and in balanced mode. Measurements were observed during movement towards the geophysical sensor with readings recorded at the 1’ intervals. Runs were consistent with the surprise of no difference between the HT powered on or off. As shown on the plot the unit did not affect the EM at approximately 2 meters from the center of the coil.

  • The test was repeated in similar fashion using the G-858 magnetic sensor in gradiometer mode. For these tests the HT strongly affected the gradiometer at approximately 1 meter and tapered to background at approximately 2 meters in the manned dynamic mode.

  • Based upon these results a tow bar length to get a minimum of 2 meters separation from the rear of the HT to the center of a sensor is required. The tow bar yoke is designed to be fully articulating so that no differential movement is transmitted between the senor and tow vehicle. A non-metallic wheel is provided for terrain following and balancing with tow bar connection allowing vertical and horizontal articulation. All components used to build the tow system are wood, composite or plastic so they do not affect any towed sensor.

  • This device was tested for suitability as a tow vehicle with typical EM-61 equipment at the USAESCH Redstone Arsenal McKinley Test Site. This site is a seeded open grassed field. The test site had not been mowed in months so the grass was approximately a foot high. In addition the unit’s battery was slightly depleted from being off charge for several days. Regardless it still provided approximately 3 hours of operation. The integrated system included the EM-61 and tow vehicle but not a primary positioning system due to non-availability (all at APG). Fiducial measurements were established and the array traversed typical DGM mapping lanes. It was easily controlled, highly maneuverable with adequate power and traction in grass even with the standard smooth pavement tires.The smooth tires were inadequate for the sand pit test area where the loss of traction caused the Segway to immediately shut down. Visualization and analysis of the acquired data set has shown no difference between the results from a man towed system even including the fiducial lag shown in the included plot. These efforts were performed at no cost to ESTCP by USAESCH.($20k contribution by USAESCH)

II. Phase I- Segway Robotic Mobility Platform Application

Overview:

  • Based upon the pre-project findings a Segway XT and RMP200 and 400 will be procured and modified to minimize the effect to the DGM sensors and to optimize it as a tow vehicle. The XT and RMPvehicles will be initially developed and tested as the two wheeled version in a similar fashion to the pre-project testing by operation in the remotely operated joy stick mode. The testing will then be repeated and the performance evaluated as configured in a four wheeled version. It is envisioned that the two wheel mode will be adequate for areas with smooth terrain and low vegetation with a light payload. The four wheel version may be required for the more challenging conditions. We will define the limits of terrain and payload for the two configurations.
  • The system will then be augmented by the addition of the ArcSecond Indoor GPS laser based positioning system and by a path following robotic control program on a PC control computer. For the geophysical sensor positioning will be provided by the ArcSecond Triad positioning system uses a rover sensor array that senses x-y-z position and multi-axis movements to sub-centimeter accuracy at a 40 Hz refresh rate for up to a 5 acre area. Since the tow vehicle and geophysical sensor’s relative position changes with terrain and maneuvering a separate position sensor will provide the tow vehicle location. The software will compare the tow vehicles and geophysical sensor’s positions to the desired path to compute the lever arm and then instruct the tow vehicle to maintain the geophysical sensor on track. This accurate position with a high refresh when integrated with the Segway 100Hz control system to position and maintain the DGM vehicle along the planned pathway will result in an unprecedented accuracy and adjustment speed for path following. The principle advantage with this concept is that a site (geophysical prove out or technology demonstration) can be traversed many times with multiple sensors configurations traveling at exactly the same design pathway and speed over the subsurface anomalies. This should provide laboratory test results quality in a real field applications.
  • The system will be repetitively tested at the McKinleyRange to validate performance and data quality using EM and magnetometer DGM sensors and the various tow vehicle configurations.
  • Upon successfully completing the McKinley testing, the system suite will be demonstrated at the APG Standard UXO Demonstration Site’s Calibration Grid using EM and magnetometer DGM sensors. The acquired data sets will be provided to researchers for in depth analysis using the newly developed predictive models. (Funded outside this project by existing and USAESCH funded projects)
  • Results will be documented into a Phase I report, presented at the IPR and the SERDP/ESTCP Environmental Conference for a Go/No go decision on Phase II.

Phase I Technical Details

  • The Phase I effort will be focused on demonstrating capability of both teleoperation of the Segway tow vehicle/DGM trailer system and autonomous path following in an obstacle-free environment. Work will be divided into two main tasks: a) video feedback/ manual teleoperation, followed by b) autonomous navigation and path following. The communication and control system hardware will include two laptop computers (one mounted on the Segway tow vehicle, the second remotely located and functioning as the user interface). The two computers will communicate wirelessly.

Video feedback and manual teleoperator control

  1. A low cost video camera (InsideOut Watchport V) mounted on a position controllable platform (Eagletron Trackerpod) will be installed on the tow vehicle. The Trackerpod platform and associated control software gives 160˚ pan and 110˚ tilt of the mounted camera, and connects to the tow vehicle computer's USB port (powered USB port is preferred). The camera is also connected to the computer through a USB port.
  1. The control commands for the Segway RMP will be entered through a graphical user interface (GUI) implemented on the laptop computer. The GUI will be implemented using Microsoft Visual Basic, and will run simultaneously with the video camera software on the Windows XP operating system.
  1. To achieve teleoperation, the laptop computer on the Segway will need to be remotely accessble. A JAVA-based remote access and computer desktop control service and software system will be used to connect the Segway computer and remote computer over the Internet (for example:

Communication system

The teleoperator system computers will communicate video and command signals over an 802.11b/g wireless network. The computer on the tow vehicle and the remote computer will be equipped with commercially available wireless network interface cards. If necessary, physical range can be extended by additional radio frequency amplifiers.