Experimental unmanned machines could assume high-risk field duties.
By the latter part of this decade, a fleet of wheeled robots now evolving toward autonomy may perform many of the tasks handled by today’s front-line soldier. The U.S. Army is experimenting with a prototype of radio-operated vehicles capable of engaging in various kinds of reconnaissance and surveillance activities. Once fully integrated into the service, these unmanned units will enable the execution of important objectives while reducing the casualties and logistical complexities often associated with rapid reaction forces.
Since the inception of the Army’s demonstration unmanned ground vehicle (UGV) program in the early 1990s, research in fielding radio-controlled vehicles that can be assigned potentially dangerous mission-critical tasks has been ongoing. Two UGV demonstration programs relied largely on human intervention to achieve designated performance goals. The first demonstration involved a combination of technologies in the development of a feasible unmanned vehicle project. The second demonstration focused on modeling and simulation techniques to measure the compatibility of robotic vehicles and humans in a series of virtual exercises.
The third demonstration brought a renewed focus to the ideas of operational autonomy and sensory perception. Using the early prototype of an experimental unmanned vehicle (XUV), the program examined ways in which adaptable field reconnaissance robots could be employed with little or no human intervention. “Our primary focus with Demo III is to further the state-of-the-art in autonomous mobility. The development of sensory and environmental perception capabilities within the XUV is a key effort that is moving us toward full autonomy,” Jeffrey Jaczkowski states. He is an electrical engineer and technical representative for the Demo III UGV program at the Tactical Army Command, Fort Knox, Kentucky.
“With the integration of an unmanned platform into a battlefield scenario, many undetectable aspects of a landscape can be discovered before they become harmful to the manned portion of an operation,” Jaczkowski points out. “Both natural and man-made dangers such as ditches or explosives can be located and made known to advancing units ahead of time.”
Under guidelines set forth by the Army’s military scout mission, the program hardware centers around a reconnaissance and surveillance target acquisition (RSTA) suite mounted on each XUV. The RSTA serves as the sensory hub of each vehicle. As in manned scout operations where a designated individual performs target location duties, the suite would be the central system that controls targeting capabilities within the Demo III program architecture.
Testing began on two XUVs in August 1999 in coordination with a high mobility multipurpose wheeled vehicle (HMMWV). An operational control unit (OCU) was mounted within each HMMWV. This man-machine interface serves as the mobile command center for XUV operations. Using digitized data management, a communication loop is established between the XUV and the OCU.
Prior to the start of a simulated mission, the objective commander loads an operational plan, including terrain assessment, route mapping and weather conditions into the OCU. The operator chooses a single- or multiple-point vehicle route that is initially dependent on line-of-sight limitations such as weather patterns and ground contour. Once underway, the XUV uses onboard sensors to detect positive obstacles such as trees and rocks that were not part of the initial map input and then automatically adds them to the virtual landscape. In the case of negative obstacles such as holes or ditches, sensors cannot detect subtle variations in the terrain surface. The lack of incoming data alerts the vehicle to avoid these areas.
“The problem with negative obstacles is that it is more difficult for the XUV to pick up the hidden surface of a drop in elevation,” Jaczkowski remarks. “In an effort to increase our capability to detect negative obstacles, we have moved the sensors up and back, increasing the angle of trajectory with which the vehicle sees.”
Equipped with both standard and laser detecting and ranging radar (LADAR) capabilities, each vehicle can see objects at an unobstructed range of approximately 15 kilometers. If visibility is decreased by thick foliage, LADAR allows the XUV to pinpoint designated targets individually, helping the vehicle avoid otherwise hidden threats. A forward looking infrared (FLIR) targeting system allows the OCU operators to achieve extended vision during night operations. The development of a foliage penetrating radar that will operate using a wideband ultrahigh frequency signal and enable target acquisition in heavily vegetated areas is currently underway.
“Aside from detecting enemy locations, the XUVs can relay data back to the OCU vehicle about the best pathway of approach for manned operations,” Jaczkowski points out. “This capability will save an advancing unit time and possible casualties by giving a heads up to a tactical command on the conditions in front of them.”
Communications between the XUV and the OCU operator are carried by a standard near term data radio (NTDR). Data can be transferred from the OCU unit to the XUV at 25 kilobits per second over a distance of up to 15 kilometers. Very high frequency (VHF) digital streaming ensures the availability of a sufficient amount of bandwidth for this type of short-range data transmission.
“Right now, the XUV/OCU vehicle range is 4 kilometers depending on terrain conditions for operator control purposes,” Jaczkowski indicates. “The capability of switching to a high frequency satellite communications link is there, but so far the vehicles have only been tested for close-in VHF operations.” NTDR scalability will allow for future modifications to the radio platform to accommodate additional forms of connectivity.
In the programming phase of a mission, the OCU operator inputs the coordinates that will place the XUV on the proper course without considering the obstacles that the vehicle can encounter. Any obstructions not accounted for on the map are added as they are encountered. The number of destinations entered can vary greatly and range from a single to a multiple-stop command.
Laptop-based visualization of images collected by the XUV sensors during scouting missions is in the form of four-dimensional, real-time control system messaging that continually changes as new data is received. Highly modular digital operability enables the OCU operator to maintain control of the vehicle so that the best decisions can be made regarding the speed and direction of travel in varying terrain and weather conditions. A pull-down menu on the laptop screen offers mapping options that the operator uses to determine the optimal route for a mission.
“As long as there is a feasible way to arrive at a location, the XUV will determine the optimal path,” Jaczkowski notes. “Once the mission route is entered into the planner, the XUV uses its map data to generate a high-level vehicle plan. The XUV then attempts to follow the initial plan using its onboard mobility sensors in making appropriate adjustments to the original planned route.”
Speed requirements for Demo III XUVs are important for the effective operation of the vehicles. Current limits have been set at 32 kilometers per hour (20 miles per hour) for daytime use and 16 kilometers per hour (10 miles per hour) after dark. These limitations are subject to changes based on the terrain, weather and mission type. The pace at which an XUV is moving directly affects the RSTA’s ability to identify specific objects in the vehicle’s path. As speed increases, the system has less time to process sensory data on particular targets before overtaking them. A goal of 40 kilometers per hour (25 miles per hour) has been set for the fall of 2001.
During a deployment, as the RSTA suite begins targeting certain contacts, icons that correlate to these objects pop up on the OCU display map. By selecting a given icon, the operator can magnify the resolution of the image for a better view of the potential target. Once chosen, the icon sends a “chip” to the screen, giving coded information based on the assessment that the RSTA sensors have made about the target. The OCU operator can then call the tactical command center and relay the data received. The tactical action officer can then determine the course of action that should be taken.
“The RSTA suite scans an area in a low degree of resolution so that it can cover a maximum amount of area,” Jaczkowski explains. “The ‘chip’ obtained by the operator is also in low resolution; however, this can be increased by box magnification of a selected piece of the image.” Once locked in, each target is given a coded designation for categorization and prioritization purposes.
The Army Research Laboratory in Aberdeen, Maryland, is pursuing enhancements in the autonomous target recognition (ATR) capabilities of RSTA sensors. By matching collected information to a database of stored images, the RSTA is able to characterize certain targets. This is accomplished through the use of specific ATR algorithms that enable the system’s sensory mechanisms to identify particular objects.
The XUV’s capabilities in stationary targeting permit most objects to be identified from 2 kilometers away. During movements, pure recognition requires distances of approximately 1.5 kilometers. A freeze-frame capability enables the OCU operator to examine a contact even if it is moving at a high speed. Work is being done that will ultimately allow highly specific target recognition, down to the type or model of a piece of equipment.
With target acquisition and identification comes the effective dissemination of gathered data to units within a task force. In addition to NTDR communications, each OCU vehicle has network-based connectivity with the other operator units in a mission. In the future, highly collaborative operations involving joint task forces will be able to constantly update both field and command units in real-time regarding the progress of XUV scout robots on the battlefield.
“A significantly smaller footprint than the standard HMMWV without sacrificing durability was important to increasing vehicle survivability,” Jaczkowski notes. “A scout mission’s success is dependent on not being detected, and decreased platform size is key to this objective.” If environmental conditions are not conducive to larger units––such as in densely wooded areas or narrow mountainous landscapes—performing certain maneuvers can depend on the equipment’s physical dimensions. Current model XUVs are 9 feet long and 5 feet wide, with as much as a 10-foot RSTA antenna clearance. The unloaded weight of each vehicle is about 3,400 pounds.
Operability during adverse weather conditions is another factor in actual battlefield dynamics. Suited for any climate or temperature zone, the Demo III XUV can scout missions that would otherwise jeopardize lives of deployed troops. The XUV can replace humans in performing reconnaissance duties under extreme conditions when an enemy or the weather is dangerously unpredictable. Retrieving important equipment under such circumstances is also easier with an XUV.
Within the arena of logistics, developments in UGV technology are paving the way for a self-sufficient fleet of supply-capable robotic vehicles. Proposed modifications to existing XUV models such as the mobile detection assessment and response system (MDARS) would increase the size of the platform to accommodate heavier load tasking requirements. New prototypes could climb 60-degree grades, maneuver on 40-degree slopes, and reach speeds of 64 kilometers per hour on paved roads. A future initiative of the UGV program includes the addition of robotic arm capabilities for use in equipment retrieval and material detection such as toxic chemicals or mines.
Zafar Hassan, test director for the Demo III UGV program, Army Research Laboratories, states, “Aside from enhancing the autonomous capabilities of the vehicles, we are trying to increase the general robustness of the robotics technology itself. By incorporating modularity into the design, we raise the scalability limitations and make the equipment more receptive to future upgrading.” Current experimentation and simulation efforts are being conducted at Aberdeen and Fort Knox, respectively.
“Preparing the ordinary soldier for the gradual integration of these vehicles into service can only be done by making the technology as real as possible,” Hassan comments. “Until the machines are part of an actual deployment, simulating their presence is the best and only way we have to get our troops ready for the future of warfare.”
A 6-month user appraisal of the field scouting robustness of the Demo III XUV is scheduled for fall 2002.