This PDF file contains the front matter associated with SPIE Proceedings Volume 6962, including the Title Page, Copyright information, Table of Contents, Introduction, and the Conference Committee listing.

Detecting people around unmanned ground vehicles (UGVs) to facilitate safe operation of UGVs is one of the highest priority issues in the development of perception technology for autonomous navigation. Research to date has not achieved the detection ranges or reliability needed in deployed systems to detect upright pedestrians in flat, relatively uncluttered terrain, let alone in more complex environments and with people in postures that are more difficult to detect. Range data is essential to solve this problem. Combining range data with high resolution imagery may enable higher performance than range data alone because image appearance can complement shape information in range data and because cameras may offer higher angular resolution than typical range sensors. This makes stereo vision a promising approach for several reasons: image resolution is high and will continue to increase, the physical size and power dissipation of the cameras and computers will continue to decrease, and stereo cameras provide range data and imagery that are automatically spatially and temporally registered. We describe a stereo vision-based pedestrian detection system, focusing on recent improvements to a shape-based classifier applied to the range data, and present frame-level performance results that show great promise for the overall approach.

For the TARDEC-funded Daredevil Project, iRobot Corporation is developing capabilities that will allow small UGVs to navigate autonomously in adverse weather and in foliage. Our system will fuse sensor data from ultra wideband (UWB) radar, LIDAR, stereo vision, GPS, and INS to build maps of the environment showing which areas are passable (e.g. covered by tall grass) and which areas must be avoided (i.e. solid obstacles). In Phase I of this project, we demonstrated that UWB radar sensors can see through precipitation, smoke/fog, and foliage and detect solid obstacles. In Phase II, we are integrating all of these sensors with an iRobot PackBot. By the end of Phase II, we will demonstrate a fully-autonomous Daredevil PackBot that can avoid obstacles, build maps, and navigate to waypoints in all-weather conditions and through foliage.

The vision system of a mobile robot for checkpoint and perimeter security inspection performs multiple functions: providing surveillance video, providing high resolution still images, and providing video for semi-autonomous visual navigation. Mid-priced commercial digital cameras support the primary inspection functions. Semi-autonomous visual navigation is a tertiary function whose purpose is to reduce the burden of teleoperation and free the security personnel for their primary functions. Approaches to robot visual navigation require some form of depth perception for speed control to prevent the robot from colliding with objects. In this paper present the initial results of an exploration of the capabilities and limitations of using a single monocular commercial digital camera for depth perception. Our approach combines complementary methods in alternating stationary and moving behaviors. When the platform is stationary, it computes a range image from differential blur in the image stack collected at multiple focus settings. When the robot is moving, it extracts an estimate of range from the camera auto-focus function, and combines this with an estimate derived from angular expansion of a constellation of visual tracking points.

Maintaining real-time situational awareness of military combat vehicles (manned/unmanned) with onboard vision sensors for either autonomous mobility or reconnaissance missions such as moving target indication (MTI) and automatic target recognition (ATR) while the vehicle is on the move has been technically and operationally challenging. In this paper, we investigate and present a practical implementation of a robust real-time structure from motion technique that allows moving robotic vehicles to be able to reconstruct 3D models from observed 2D features with dynamically adjusting motion parameters. We also demonstrate applications that locate and track moving targets within the structured environment built by the SFM and recognize the targets such as vehicles and humans through a hierarchical shape model.

Robot navigation in complex, dynamic and unstructured environments demands robust mapping and localization solutions. One of the most popular methods in recent years has been the use of scan-matching schemes where temporally correlated sensor data sets are registered for obtaining a Simultaneous Localization and Mapping (SLAM) navigation solution. The primary bottleneck of such scan-matching schemes is correspondence determination, i.e. associating a feature (structure) in one dataset to its counterpart in the other. Outliers, occlusions, and sensor noise complicate the determination of reliable correspondences. This paper describes testing scenarios being developed at NIST to analyze the performance of scan-matching algorithms. This analysis is critical for the development of practical SLAM algorithms in various application domains where sensor payload, wheel slippage, and power constraints impose severe restrictions. We will present results using a high-fidelity simulation testbed, the Unified System for Automation and Robot Simulation (USARSim).

In September 2007 the Army Research Laboratory (ARL) Robotics Collaborative Technology Alliance (CTA) conducted an assessment of multiple pedestrian detection algorithms based upon LADAR or video sensor data. Eight detection algorithms developed by the Robotics CTA member organizations, including ARL, were assessed in an experiment conducted by the National Institute of Science & Technology (NIST) and ARL to determine the probability of detection/misclassification and false alarm rate as a function of vehicle speed, degree of environmental clutter, and pedestrian speeds. The study is part of an ongoing investigation of safe operations for unmanned ground vehicles. This assessment marked the first time in this program that human movers acted as targets for detection from a moving vehicle. A focus of the study was to choreograph repeatable human movement scenarios relative to the movement of the vehicle. The resulting data is intended to support comparative analysis across treatment conditions and to allow developers to examine performance with respect to specific detection and tracking events. Events include humans advancing and retreating from the vehicle at different angles, humans crossing paths in close proximity and occlusion situations where sight to the mover from the sensor system is momentarily lost. A detailed operational procedure ensured repeatable human movement with independent ground truth supplied by a NIST ultra wideband wireless tracking system. Post processing and statistical analysis reconciled the tracking algorithm results with the NIST ground truth. We will discuss operational considerations and results.

An investigation of power and resolution for laser ranging sensors is performed in relation to sense and avoid requirements of miniature unmanned aircraft systems (UAS). Laser rangefinders can be useful if not essential complements to video or other sensing modalities in a sense and avoid sensor suite, particularly when applied to miniature UAS. However, previous studies addressing sensor performance requirements for sense and avoid on UAS have largely concentrated, either explicitly or implicitly, on passive imaging techniques. These requirements are then commonly provided in terms of an angular resolution defined by a detection threshold. By means of a simple geometric model, it is assumed that an imaging system cannot distinguish an object that subtends less than a minimum number of detector pixels. But for sensors based on active ranging, such as laser rangefinders and LADAR, detection probability is coupled to the optical power of the laser transmitter. This coupling enables the sensors to achieve sub-pixel resolution, or resolution better than the instantaneous field-of-view, and to compensate for insufficient angular resolution by increasing transmitter power. Consequently, when considering sense and avoid detection requirements for laser rangefinders or LADAR, a tradeoff emerges between resolution and power, which, owing to the inherent relationship of size and weight to system resolution, translates to a tradeoff between resolution and sensor size, weight, and power. In this presentation, we investigate the existence of an optimum compromise between sensor resolution and power, concentrating on platforms with particularly challenging payload limitations.

In this work we propose the use of digital watermarks to establish track history of multiple tracked objects. Target tracks extracted from a frame sequence can be represented and analyzed using spatio-temporal graphs. However, this approach requires extensive computation and data storage. It is possible to achieve the same result by digitally watermarking each tracked target. This self-contained video can then be searched to extract track history. Track history is obtained by computing a quantity called target adjacency matrix. Interpretation of the elements of this matrix, their location, magnitude and movement will establish desired track history. Given two randomly selected frames we can establish changes of spatial relationship among targets in the two frames as well as entry or exit of new targets in each frame.

In this paper we present our new method of automatic control of X-ray picture gray scale stretch, noise reduction and visual spatial resolution enhancement that improves the human visual picture analysis. The method is based on our set-theoretical model of the image using details clustering of a class of large details with dimensions more 4 pixels and a low dimension detail class. The last class is divided into two subclasses of distinguishable details and detectable details only. The large detail data histogram determines a pixels volume in dark area and a data value related to histogram maximum in the dark area. The received picture features are used as adaptation parameters for optimization of a picture global or local automatic gray scale stretch, noise reduction and visual spatial resolution enhancement improving object recognition. The small detail clustering into the two subclasses provides automatic visual resolution enhancement without noise visibility increase. The developed automatic control of X-ray image improvement was took training and was checked by processing series of objects: test patterns, various baggages, telephone sets, etc. The check results provided a fine tune of the developed automat improving object recognition. The experimental and practical results are discussed in the paper.

The use of unmanned vehicles in Reconnaissance, Surveillance, and Target Acquisition (RSTA) applications has received considerable attention recently. Cooperating land and air vehicles can support multiple sensor modalities providing pervasive and ubiquitous broad area sensor coverage. However coordination of multiple air and land vehicles serving different mission objectives in a dynamic and complex environment is a challenging problem. Swarm intelligence algorithms, inspired by the mechanisms used in natural systems to coordinate the activities of many entities provide a promising alternative to traditional command and control approaches. This paper describes recent advances in a fully distributed digital pheromone algorithm that has demonstrated its effectiveness in managing the complexity of swarming unmanned systems. The results of a recent demonstration at NASA's Wallops Island of multiple Aerosonde Unmanned Air Vehicles (UAVs) and Pioneer Unmanned Ground Vehicles (UGVs) cooperating in a coordinated RSTA application are discussed. The vehicles were autonomously controlled by the onboard digital pheromone responding to the needs of the automatic target recognition algorithms. UAVs and UGVs controlled by the same pheromone algorithm self-organized to perform total area surveillance, automatic target detection, sensor cueing, and automatic target recognition with no central processing or control and minimal operator input. Complete autonomy adds several safety and fault tolerance requirements which were integrated into the basic pheromone framework. The adaptive algorithms demonstrated the ability to handle some unplanned hardware failures during the demonstration without any human intervention. The paper describes lessons learned and the next steps for this promising technology.

The U.S. Army Research Laboratory's (ARL) Computational and Information Sciences Directorate (CISD) has long been involved in autonomous asset control, specifically as it relates to small robots. Over the past year, CISD has demonstrated the ability to control and view streaming video from an FCS-surrogate PackBot robotic system over multiple network types (Soldier Radio Waveform (SRW), 802.11), as well as tasking the robot to follow both manually (ARL DigitalInk) and autonomously planned (CERDEC C2ORE) GPS waypoint routes. These capabilities remove the "stand alone system" limitations of traditional small robot systems and allow any and all data produced by such platforms to be available to anyone on the network, while at the same time reducing the amount of operator intervention required to utilize a robot. However, assumptions were made about the paths the robot was to traverse, specifically that they would be free from major obstacles. To address these system limitations, CISD is implementing obstacle detection and avoidance (OD/OA) on the PackBot. The OD/OA utilizes COTS ranging sensors with indoor and/or outdoor capabilities, and leverages existing software algorithm components into the existing CISD robotic control architecture. These new capabilities are available in an integrated environment consisting of common command and control (C2) and network interfaces and on multiple platforms (ARL ATRV, LynchBot, PackBot, etc.) due to the modular and platform/network independent architecture that ARL employs. This paper will describe the current robotic control architecture employed by ARL and provide brief descriptions of existing capabilities. Further, the paper will discuss the small robot obstacle detection/avoidance integration effort performed by ARL, along with some preliminary results on its performance and benefits.

Unmanned ground vehicles (UGV) operating in outdoor environments must traverse unstructured terrain. This terrain is diverse in nature and contains natural obstacles such as rocks, brushes, berms, and low lying wet areas. Outdoor terrain is not static as it varies on a seasonal basis due to the life cycle associated with natural vegetation. Additionally, outdoor terrain may change appearance due to variations in lighting conditions that result from the Sun's relative position and from weather conditions such as clouds, fog or rain. This environmental diversity has long caused researchers considerable grief, as developing a classical terrain classification algorithm has proven to be a very difficult if not an impossible task. Researchers have skirted this problem by relying upon ranging sensors and constructing 2 ½D or, more recently, 3D world representations. Although geometrical representations have been used extensively, the low data rates associated with laser rangefinders, the unreliability of stereo vision, and the interaction between geometry and orientation estimation errors have limited the lookahead distance, thereby reducing the maximum attainable vehicle speeds. Learning from experience, in a more human like manner, promises to reduce or alleviate many of the issues posed by unstructured outdoor terrain. Defence R&D Canada (DRDC) "Learned Trafficability" program researches learning from experience. The paper presents DRDC's progress in extending a 2 ½D world representation using vision and learning from experience.

Unmanned ground vehicles (UGVs) will play an important role in the nation's next-generation ground force. Advances in sensing, control, and computing have enabled a new generation of technologies that bridge the gap between manual UGV teleoperation and full autonomy. In this paper, we present current research on a unique command and control system for UGVs named PointCom (Point-and-Go Command). PointCom is a semi-autonomous command system for one or multiple UGVs. The system, when complete, will be easy to operate and will enable significant reduction in operator workload by utilizing an intuitive image-based control framework for UGV navigation and allowing a single operator to command multiple UGVs. The project leverages new image processing algorithms for monocular visual servoing and odometry to yield a unique, high-performance fused navigation system. Human Computer Interface (HCI) techniques from the entertainment software industry are being used to develop video-game style interfaces that require little training and build upon the navigation capabilities. By combining an advanced navigation system with an intuitive interface, a semi-autonomous control and navigation system is being created that is robust, user friendly, and less burdensome than many current generation systems. mand).

This paper presents an exploration of methods for estimating terrain trafficability from visual appearance. Two different sets of data are used. The first set is extracted from video sequences and has a small number of different terrains. A fuzzy c-means clustering algorithm is used to predict terrain type. The second set is derived from high-resolution still images and has a large variety of terrains. A decision tree algorithm is used to provide a subjective assessment of trafficability. A variety of local features are explored, based on color and texture, as input to the learning algorithms.

In this paper, we present SCOUTS, an autonomous system for the allocation of multiple small UAVs equipped with video sensors specifically geared to tracking and reacquisition of vehicle targets in urban environments. To demonstrate SCOUTS performance, we test our system in a simulation incorporating the key challenges encountered by real-time UAV tracking scenarios, including unreliable reports from poor image quality, disruption latency in communication, and temporary gaps in persistent surveillance requiring target re-acquisition strategies. The results provided in this paper establish SCOUTS as a robust approach to this challenging problem.

Sliding scale autonomy has been suggested as a practical way of organizing robot controls. In this paper, we investigate practical problems in sliding scale autonomy. We take a constructionist approach of building systems with sliding scale autonomy and identifying the challenges involved. We implement robot systems capable of operating semi-autonomously with user guidance. We describe the techniques used to construct the system, the problems discovered along the way, and the improvements achieved. We discuss key parts of two robotic systems we have constructed. We describe our user interface, obstacle avoidance, and localization as implemented on the robot systems.

Small, agile rovers for harsh outdoor environments offer good potential to support rescue teams in emergency operations. For such exploration purposes the outdoor MERLIN rovers have been developed in tracked and wheeled versions for the weight class between 10 and 20 kg. Those vehicles can achieve velocities up to 50 km/h. Therefore the drive assistance system has to provide the functionalities to perform safe and efficient tele-operations in combination with autonomous reaction capabilities. The tele-operator can select appropriate levels of autonomy, ranging from warning signals to autonomous reactions of the vehicle's on-board data processing system, if an endangering situation is not anticipated. Implemented features include detection of obstacles in the path, as well as an adaptation of speed appropriate to terrain roughness and slope, but also to path curvature. An autonomous return to the initial position is to be realized, when the telecommunication contact to the tele-operator has been lost. This paper addresses the implemented sensor and data processing techniques to handle those tasks in a robust way. Results from extensive tests in various environments will be reported. In particular the results from the C-ELROB 2007 competition, the European Land Robotics trial, will be reported, where the Outdoor MERLIN was the winner of the urban terrain challenge.

Mobile manipulation in many respects represents the next generation of robot applications. An important part of design of such systems is the integration of techniques for navigation, recognition, control, and planning to achieve a robust solution. To study this problem three different approaches to mobile manipulation have been designed and implemented. A prototypical application that requires navigation and manipulation has been chosen as a target for the systems. In this paper we present the basic design of the three systems and draw some general lessons on design and implementation.

Robotic manipulators used on current EOD robotic platforms exhibit very few autonomous capabilities. This lack of autonomy forces the operator to completely control manipulator movements. With the increasing complexity of robotic manipulators, this can prove to be a very complex and tedious task. The development of autonomous capabilities for platform navigation are currently being extensively researched and applied to EOD robots. While autonomous manipulation has also been researched, this technology has yet to appear in fielded EOD robotic systems. As a result, there is a need for the exploration and development of manipulator automation within the scope of EOD robotics. In addition, due to the expendable nature of EOD robotic assets, the addition of this technology needs to add little to the overall cost of the robotic system. To directly address the need for a low-cost semi-autonomous manipulation capability for EOD robots, the Naval Explosive Ordnance Disposal Technology Division (NAVEODTECHDIV) proposes the Autonomous Robotic Manipulator (ARM). The ARM incorporates several semi-autonomous manipulation behaviors including point-and-click movement, user-defined distance movement, user-defined angle positioning, memory locations to save and recall manipulator positions, and macros to memorize and repeat multi-position repetitive manipulator movements. These semi-autonomous behaviors will decrease an EOD operator's time on target by reducing the manipulation workload in a user-friendly fashion. This conference paper will detail the background of the project, design of the prototype, algorithm development, implementation, results, and future work.

As mobile robots continue to gain acceptance across a variety of applications within the defense and civilian markets, the number of tasks that these robot platforms are expected to accomplish are expanding. Robot operators are asked to do more with the same platforms - from EOD missions to reconnaissance and inspection operations. Due to the fact that a majority of missions are dangerous in nature, it is critical that users are able to make remote adjustments to the systems to ensure that they are kept out of harm's way. An efficient way to expand the capabilities of existing robot platforms, improve the efficiency of robot missions, and to ultimately improve the operator's safety is to integrate JAUS-enabled Intelligent Modular Manipulation payloads. Intelligent Modular Manipulation payloads include both simple and dexterous manipulator arms with plug-and-play end-effector tools that can be changed based on the specific mission. End-effectors that can be swapped down-range provide an added benefit of decreased time-on-target. The intelligence in these systems comes from semi-autonomous mobile manipulation actions that enable the robot operator to perform manipulation task with the touch of a button on the OCU. RE2 is supporting Unmanned Systems Interoperability by utilizing the JAUS standard as the messaging protocol for all of its manipulation systems. Therefore, they can be easily adapted and integrated onto existing JAUS-enabled robot platforms.

This paper describes progress toward robotic arm system to collect swab samples for trace chemical analysis. Collecting a swab sample requires bringing the swab in contact with the target object, maintaining adequate pressure against the object while dragging the swab tip across the surface, conforming to the surface compliance, curvature and irregularities. It also requires detecting when the swipe motion is blocked, when it has reached the end or edge of the object, and when the normal swipe excursion has been completed. Remote or robotic swab sample collection is complicated by the fact that key physical properties of the target object, e.g., its location, surface contour, compliance, friction, are unknown or, at best, uncertain. We are pursuing a two-fold approach to computer assisted robotic examination swab sampling. We are developing a force-feedback master-slave puppet arm control system in which the operator manipulates and receives force feedback through a scale model of the remote arm. We are also developing adaptive motion behaviors for autonomous swab sample collection, in which the arm feels its way over the surface, adjusting its configuration to conform to the surface contour, while maintaining pressure and keeping the swab tip in the desired orientation with respect to the surface as it drags the swab across the target object. Experiments with the master-slave system will provide data on human operator adaptive motion behaviors, and provide a baseline for evaluation of the automatic system. This paper describes the force-feedback master-slave puppet arm control system, presents example teleoperated swab dynamics data, describes the emerging framework for analysis of adaptive motion behaviors in swab sample collection, and describes our approach to autonomous swab sampling adaptive behavior and control.

Autonomous Solutions has multiple projects involving mobile robot arms for EOD applications. This paper will describe 3 such projects. In the first, a 5 DOF robot arm of our own design is mounted on a 4-wheeled omni-directional platform. The arm is capable of lifting 100 pounds at close to full extension, and it can lift significantly more at shorter distances. The operator can specify either individual joint control or "fly the gripper" control, where the individual joint velocities are automatically calculated in response to commanded gripper velocities. In the second project, 3D data of a scene of interest is gathered and then presented to the EOD technician, along with a representation of the arm. The technician can then rotate the scene and arm to obtain the best possible view for the task. In the final project, a 3 DOF arm will be mounted to an omni-directional platform. The arm and platform will be treated as a single 6 DOF manipulator, and the operator will specify 6 DOF gripper velocity commands using a 6-axis input device, such as those used for solid modeling programs.

Manipulator control interface development and optimization is a well researched topic covering both hardware and software implementations. Studies of this nature are intended to better understand the relationship between an operator and a teleoperated robotic platform. Few studies, however have conducted a systematic and statistical study of an operator's level of performance piloting a mobile manipulator platform in a line-of-sight, (LOS) situation based on the relationship between the degrees of freedom present on a controller and the degrees of freedom present on the robotic platform. This paper presents the development and results of different hardware controller interfaces for the ODIS-T2 mobile platform and an attached 4 degree of freedom manipulator, using both established and novel methodologies. Statistical results from operators' performance for each controller interface during line-of-sight operation along with operators' preferences and perceived workloads are used to make a conclusion with regards to the optimal controller configuration for the ODIS-T2 mobile manipulator platform.

This paper discusses the hardware and software of a robot-assisted vehicle inspection system - describing, among other things, the inverse-kinematics and special trajectory and "corridor" control needs of the application, plus the coordination of mobile base and robotic arm control. Typical target vehicles are illustrated and task sequencing presented which drives a high level of surveillance and analysis coverage of the vehicle. There are many activities going on all at once during the surveillance sweep of a vehicle, all of which must be and are easily accessible for control and monitoring from a simple uncluttered operational console and dashboard.

As unmanned systems become more commonplace in military, police, and other security forces, they are tasked to perform missions that the original hardware was not designed for. Current military robots are built for rough outdoor conditions and have strong inflexible manipulators designed to handle a wide range of operations. However, these manipulators are not well suited for some essential indoor tasks, including opening doors. This is a complicated kinematic task that places prohibitively difficult control challenges on the robot and the operator. Honeybee and iRobot have designed a modular door-breaching manipulator that mechanically simplifies the demands upon operator and robot. The manipulator connects to the existing robotic arm of the iRobot PackBot EOD. The gripper is optimized for grasping a variety of door knobs, levers, and car-door handles. It works in conjunction with a compliant wrist and magnetic lock-out mechanism that allows the wrist to remain rigid until the gripper has a firm grasp of the handle and then bend with its rotation and the swing of the door. Once the door is unlatched, the operator simply drives the robot through the doorway while the wrist compensates for the complex, multiple degree-of-freedom motion of the door. Once in the doorway the operator releases the handle, the wrist pops back into place, and the robot is ready for the next door. The new manipulator dramatically improves a robot's ability to non-destructively breach doors and perform an inspection of a room's content, a capability that was previously out of reach of unmanned systems.

Current practice for the detection of chemical, biological and explosive (CBE) agent contamination on environmental surfaces requires a human to don protective gear, manually take a sample and then package it for subsequent laboratory analysis. Ground robotics now provides an operator-safe way to make these critical measurements. We describe the development of a robot-deployed surface detection system for CBE agents that does not require the use of antibodies or DNA primers. The detector is based on Raman spectroscopy, a reagentless technique that has the ability to simultaneously identify multiple chemical and biological hazards. Preliminary testing showed the ability to identify CBE simulants in 10 minutes or less. In an operator-blind study, this detector was able to correctly identify the presence of trace explosive on weathered automobile body panels. This detector was successfully integrated on a highly agile robot platform capable of both high speed and rough terrain operation. The detector is mounted to the end of five-axis arm that allows precise interrogation of the environmental surfaces. The robot, arm and Raman detector are JAUS compliant, and are controlled via a radio link from a single operator control unit. Results from the integration testing and from limited field trials are presented.

Early applications of mobile robots focused on reconnaissance, surveillance, and target acquisition (RSTA) or related activities. For such scenarios, a significant body of experience has been developed related to the use of camera feedback for non-line-of-sight (NLOS) tele-operation for basic robot motion control (or driving) and remote surveillance. Recently, there has been increased interest in enhancing the capabilities of mobile platforms through the addition of robotic arms and manipulators, thus allowing remote manipulation of payloads or other devices in the environment (e.g., disarming explosives in a military application). In this case, the demands on the remote vision system become very different, as there are various modes in which the user can gain awareness of the motion of the platform and the manipulator arm. This includes robot view (typical driving camera), the view from an arm-mounted camera (eye-in-hand), and possibly the view from a fixed camera or from a camera mounted on the robot that captures the complete arm motion (scene view). To date, there have been few systematic studies that compare and contrast these different approaches to operator visual interfaces for NLOS robot and arm control interfaces. In this paper we address this problem, presenting a systematic, controlled, statistical study involving multiple subjects performing multiple trials of a NLOS tele-operated mobile manipulation task. The results of the study can be used to guide the development of operator visual interfaces for mobile manipulators.

The US Army Future Combat System (FCS) will implement Unmanned Ground Vehicles (UGV) in numbers not previously seen before in military operations. Many of these vehicles will also be larger and faster than the small robots typically used today for explosive ordnance disposal and general improvised explosive device handling. More importantly, FCS will implement these UGV's in scenarios were they will be in much closer proximity to soldiers and other non-combatant personnel. This paper describes the plan for developing an appropriate match of technology for autonomous UGV maneuver with the emerging need for safety release verification for these systems prior to fielding. The plan is followed by descriptions of initial data collections with a UGV, that will form the starting point in this safety release process, and stimulate further use and refinement of this process for large UGV's in applications beyond FCS.

Military strategists are currently seeking methodologies to control large numbers of autonomous assets. Automated planners based upon the Nash equilibrium concept in non-zero sum games are one option. Because such planners inherently consider possible adversarial actions, assets are able to adapt to, and to some extent predict, potential enemy actions. However, these planners must function properly both in cases in which a pure Nash strategy does not exist and in scenarios possessing multiple Nash equilibria. Another issue that needs to be overcome is the scalability of the Nash equilibrium. That is, as the dimensionality of the problem increases, the Nash strategies become unfeasible to compute using traditional methodologies. In this paper we introduce the concept of near-Nash strategies as a mechanism to overcome these difficulties. We then illustrate this concept by deriving the near-Nash strategies and using these strategies as the basis for an intelligent battle plan for heterogeneous teams of autonomous combat air vehicles in the Multi-Team Dynamic Weapon Target Assignment model.

The agility and adaptability of biological systems are worthwhile goals for next-generation unmanned ground vehicles. Management of the requisite number of degrees of freedom, however, remains a challenge, as does the ability of an operator to transfer behavioral intent from human to robot. This paper reviews American Android research funded by NASA, DARPA, and the U.S. Army that attempts to address these issues. Limb coordination technology, an iterative form of inverse kinematics, provides a fundamental ability to control balance and posture independently in highly redundant systems. Goal positions and orientations of distal points of the robot skeleton, such as the hands and feet of a humanoid robot, become variable constraints, as does center-of-gravity position. Behaviors utilize these goals to synthesize full-body motion. Biped walking, crawling and grasping are illustrated, and behavior parameterization, layering and portability are discussed. Robotic skill acquisition enables a show-and-tell approach to behavior modification. Declarative rules built verbally by an operator in the field define nominal task plans, and neural networks trained with verbal, manual and visual signals provide additional behavior shaping. Anticipated benefits of the resultant adaptive collaborative controller for unmanned ground vehicles include increased robot autonomy, reduced operator workload and reduced operator training and skill requirements.

The use of alternative energy technology for vehicle propulsion and auxiliary power is becoming more important. Work is being performed at Michigan Technological University's Keweenaw Research Center on an Army Research Laboratory cooperative agreement to develop two unmanned ground vehicles for military applications. A wide range of alternative energy technologies were investigated. Hydrogen-powered proton exchange membrane fuel cells were identified as the most appropriate alternative energy source. This is due to some development and commercialization which makes the technology "drop-in plug-in" for immediate use. We have previously presented research work on a small unmanned ground vehicle demonstration platform where the fuel cell is the only power source. We now present research work on the integration of a fuel cell onto a larger skid steer platform. The dual-power capability of this vehicle can provide a modest level of propulsion in "engine-off mode" and may also be used to power directed energy devices which have applications in countermine and similar threat technologies.

Hands-free control of unmanned ground vehicles is essential for soldiers, bomb disposal squads, and first responders. Having their hands free for other equipment and tasks allows them to be safer and more mobile. Currently, the most successful hands-free control devices are speech-command based. However, these devices use external microphones, and in field environments, e.g., war zones and fire sites, their performance suffers because of loud ambient noise: typically above 90dBA. This paper describes the development of technology using the ear as an output source that can provide excellent command recognition accuracy even in noisy environments. Instead of picking up speech radiating from the mouth, this technology detects speech transmitted internally through the ear canal. Discreet tongue movements also create air pressure changes within the ear canal, and can be used for stealth control. A patented earpiece was developed with a microphone pointed into the ear canal that captures these signals generated by tongue movements and speech. The signals are transmitted from the earpiece to an Ultra-Mobile Personal Computer (UMPC) through a wired connection. The UMPC processes the signals and utilizes them for device control. The processing can include command recognition, ambient noise cancellation, acoustic echo cancellation, and speech equalization. Successful control of an iRobot PackBot has been demonstrated with both speech (13 discrete commands) and tongue (5 discrete commands) signals. In preliminary tests, command recognition accuracy was 95% with speech control and 85% with tongue control.

A low-cost robotic arm and controller system is presented. The controller is a desktop model of the robotic arm with the same degrees of freedom whose joints are equipped with sensors. Manipulating the controller by hand causes the robotic arm to mimic the movement in maser-slave fashion. The system takes advantage of the low cost and wide availability of hobby radio control components and uses a low-cost, easy-to-program microprocessor. The system is implemented with a video camera on the robotic arm, and the arm is mounted on an unmanned omnidirectional vehicle inspection robot. With a camera on the end of a robot arm, the vehicle inspection system can reach difficult to-access regions of the vehicle underbody. Learning to manipulate the robot arm with this controller is faster than learning with a traditional joystick. Limitations of the microcontroller are discussed, and suggestions for further development of the robot arm and control are made.

Unmanned aerial vehicles (UAVs) have proven themselves indispensable in providing intelligence, reconnaissance, and surveillance (ISR). We foresee a future where constellations of multi-purpose UAVs will be tasked to provide ISR in an unpredictable environment. Automated systems will process imagery and other sensor data gathered by the constellations to provide continuous situational awareness for the warfighter on the ground. In this paper, we present a tool that generates coordinated mission plans for constellations of UAVs with multiple goals and objectives. We call this tool Spatially Produced Airspace Routes from Tactical Evolved Networks, or SPARTEN. SPARTEN uses evolutionary algorithm (EA)-based, multi-objective optimization to generate coordinated sortie routes for constellations of UAVs. These sortie routes maximize sensor coverage, avoid conflicts between UAVs, minimize the latency of sensor data, and avoid areas of poor weather to provide valid route solutions. We use an Air Maneuver Network (AMN) based on terrain reasoning to constrain the solution space. We make two contributions to the field of UAV route planning. We develop a tool to optimize planning across multiple objectives for constellations of UAVs, and we explore the performance of this tool on a battlefield scenario.

Fielded military unmanned systems are currently extending the reach of the U.S. forces in surveillance and reconnaissance missions. Providing long-range eyes on enemy operations, unmanned aerial vehicles (UAVs), such as the AeroVironment Raven, have proven themselves indispensable without risking soldiers' lives. Meanwhile, unmanned ground vehicles (UGVs), such as the iRobot PackBot, are quickly joining ranks in Explosive Ordnance Disposal (EOD) missions to identify and dispose of ordnance or to clear roads and buildings. UAV-UGV collaboration and the benefit of force multiplication is increasingly more tangible. iRobot Corporation and CMU Robotics Institute are developing the capability to simultaneously control the Raven small UAV (SUAV) and PackBot UGV from a single operator control unit (OCU) via waypoint navigation. Techniques to support autonomous collaboration for pursuing and tracking a dismounted soldier will be developed and integrated on a Raven-PackBot team. The Raven will survey an area and geolocate an operator-selected target. The Raven will share this target location with the PackBot and together they will collaboratively pursue the target intelligently to maintain track on the target. We will accomplish this goal by implementing a decentralized control and data fusion software architecture. The PackBot will be equipped with on-board waypoint navigation algorithms, a Navigator Payload containing a stereo-vision system, GPS, and a high-accuracy IMU. The Raven will have two on-board cameras, a side-looking and a forward-looking optical camera. The Supervisor OCU will act as the central mission planner, allowing the operator to monitor mission events and override vehicle tasks.

Real-time monitoring the status of in-service structures such as unmanned vehicles can provide invaluable information to detect the damages to the structures on time. The unmanned vehicles can be maintained and repaired in time if such damages are found. One typical cause of damages of unmanned vehicles is from impacts caused by bumping into some obstacles or being hit by some objects such as hostile fire. This paper introduces a novel impact event sensing system that can detect the location of the impact events and the force-time history of the impact events. The system consists of the Piezo-electric sensor network, the hardware platform and the analysis software. The new customized battery-powered impact event sensing system supports up to 64-channel parallel data acquisition. It features an innovative low-power hardware trigger circuit that monitors 64 channels simultaneously. The system is in the sleep mode most of the time. When an impact event happens, the system will wake up in micro-seconds and detect the impact location and corresponding force-time history. The system can be combined with the SMART sensing system to further evaluate the impact damage severity.

Unmanned ground vehicles have a large number of scientific, military and commercial applications. A convoy of such vehicles can have collaboration and coordination. For the movement of such a convoy, it is important to predict the reliability of the system. A number of approaches are available in literature which describes the techniques for determining the reliability of the system. Graph theoretic approaches are popular in determining terminal reliability and system reliability. In this paper we propose to exploit Fuzzy and Neuro-Fuzzy approaches for predicting the node and branch reliability of the system while Boolean algebra approaches are used to determine terminal reliability and system reliability. Hence a combination of intelligent approaches like Fuzzy, Neuro-Fuzzy and Boolean approaches is used to predict the overall system reliability of a convoy of vehicles. The node reliabilities may correspond to the collaboration of vehicles while branch reliabilities will determine the terminal reliabilities between different nodes. An algorithm is proposed for determining the system reliabilities of a convoy of vehicles. The simulation of the overall system is proposed. Such simulation should be helpful to the commander to take an appropriate action depending on the predicted reliability in different terrain and environmental conditions. It is hoped that results of this paper will lead to more important techniques to have a reliable convoy of vehicles in a battlefield.

An omnidirectional mobile robot is able, kinematically, to move in any direction regardless of current pose. To date, nearly all designs and analyses of omnidirectional mobile robots have considered the case of motion on flat, smooth terrain. In this paper, an investigation of the design and analysis of an omnidirectional mobile robot for use in rough terrain is presented. Kinematic and geometric properties of the active split offset caster drive mechanism are investigated along with system and subsystem design guidelines. An optimization method is implemented to explore the design space. Use of this method results in a robot that has higher mobility than a robot designed using engineering judgment. A point design generated by the optimization method is shown for a meter-scale mobile robot.

The Black Knight is a 12-ton, C-130 deployable Unmanned Ground Combat Vehicle (UGCV). It was developed to demonstrate how unmanned vehicles can be integrated into a mechanized military force to increase combat capability while protecting Soldiers in a full spectrum of battlefield scenarios. The Black Knight is used in military operational tests that allow Soldiers to develop the necessary techniques, tactics, and procedures to operate a large unmanned vehicle within a mechanized military force. It can be safely controlled by Soldiers from inside a manned fighting vehicle, such as the Bradley Fighting Vehicle. Black Knight control modes include path tracking, guarded teleoperation, and fully autonomous movement. Its state-of-the-art Autonomous Navigation Module (ANM) includes terrain-mapping sensors for route planning, terrain classification, and obstacle avoidance. In guarded teleoperation mode, the ANM data, together with automotive dials and gages, are used to generate video overlays that assist the operator for both day and night driving performance. Remote operation of various sensors also allows Soldiers to perform effective target location and tracking. This document covers Black Knight's system architecture and includes implementation overviews of the various operation modes. We conclude with lessons learned and development goals for the Black Knight UGCV.

Simulation of robots in a virtual domain has multiple benefits. End users can use the simulation as a training tool to increase their skill with the vehicle without risking damage to the robot or surrounding environment. Simulation allows researchers and developers to benchmark robot performance in a range of scenarios without having the physical robot or environment present. The simulation can also help guide and generate new design concepts. USARSim (Unified System for Automation and Robot Simulation) is a tool that is being used to accomplish these goals, particularly within the realm of search and rescue. It is based on the Unreal Tournament 2004 gaming engine, which approximates the physics of how a robot interacts with its environment. A family of vehicles that can benefit from simulation in USARSim are Whegs^TM robots. Developed in the Biorobotics Laboratory at Case Western Reserve University, Whegs^TM robots are highly mobile ground vehicles that use abstracted biological principles to achieve a robust level of locomotion, including passive gait adaptation and enhanced climbing abilities. This paper describes a Whegs^TM robot model that was constructed in USARSim. The model was configured with the same kinds of behavioral characteristics found in real Whegs^TM vehicles. Once these traits were implemented, a validation study was performed using identical performance metrics measured on both the virtual and real vehicles to quantify vehicle performance and to ensure that the virtual robot's performance matched that of the real robot.

Generalized Field D* is a novel graph-based planning algorithm developed to improve autonomous navigation of robotic vehicles. The advantages of this algorithm include the possibility to reduce the search space by using multi-resolution graphs, faster planning, and smoother trajectories. The algorithm works with a wide range of simple graphs that have positive edge costs - there is no requirement for the graph to be a grid. Trajectories generated with generalized Field D* are not guaranteed to be optimal, but, in practice, they are often better than those obtained with conventional graph planners.

Both unmanned air vehicles (UAVs) and unmanned ground vehicles (UGVs) are being used in increasingly complex roles as this technology matures. There are many proposals for the joint use of UAVs and UGVs working together to achieve mission goals. Due to the spatial perception difficulties commonly reported in UGV operations, it has been postulated that the operation of UGVs could benefit from the use of live aerial views in scenarios where this is possible. In a series of experimental studies we examine the cognitive task of integrating air and ground views. Integrating these views is a key cognitive task component in UGV operations which usually require that maps for navigation be integrated with imagery from ground view cameras. Whether aerial views are used or not, integration with map views is crucial. In this paper we discuss a series of experimental studies relevant to the cognitive integration of air and ground views in UGV scenarios. The integration of map and ground camera views may be facilitated by the use of live aerial views. In urban search scenarios the addition of an aerial view to a UGV operator's task does not appear to have any measurable negative consequences and improves localization performance.

This paper documents the design and development of a low cost robot capable of autonomous navigation in unknown indoor environments. The proposed design uses only two complementary rotating sensors for navigation. The use of real time mapping allows for detection and avoidance of obstacles. The fusion of the sensors data helped improve accuracy of the online map of the robot environment. The robot builds an online map of its environment, and then automatically plans its navigation path. The feedback control keeps the robot moving along its planned path. The robot has been successfully tested in a cluttered environment in the Advanced Systems Lab. Preliminary tests carried out have shown the success of the robot in navigating autonomously.

We present the mechanical design of a bipedal walking robot named M2V2, as well as control strategies to be implemented for walking and balance recovery. M2V2 has 12 actuated degrees of freedom in the lower body: three at each hip, one at each knee, and two at each ankle. Each degree of freedom is powered by a force controllable Series Elastic Actuator. These actuators provide high force fidelity and low impedance, allowing for control techniques that exploit the natural dynamics of the robot. The walking and balance recovery controllers will use the concepts of Capture Points and the Capture Region in order to decide where to step. A Capture Point is a point on the ground in which a biped can step to in order to stop, and the Capture Region is the locus of such points.

The ability of autonomous unmanned ground vehicles (UGVs) to rapidly and effectively predict terrain negotiability is a critical requirement for their use on challenging terrain. Most methods for assessing traversability, however, assume precise knowledge of vehicle and terrain properties. In practical applications, uncertainties are associated with the estimation of the vehicle/terrain parameters, and these uncertainties must be considered while determining vehicular mobility. Here a computationally inexpensive method for efficient mobility prediction based on the stochastic response surface (SRSM) approach is presented that considers imprecise knowledge of terrain and vehicle parameters while analyzing various metrics associated with UGV mobility. A conventional Monte Carlo method and the proposed response surface methodology have been applied to two simulated cases of mobility analysis, and it has been shown that the SRSM method is an efficient tool as compared to conventional Monte Carlo methods for the analysis of vehicular mobility in uncertain environments.

This paper reports a methodology for inferring terrain parameters from estimated terrain forces in order to allow wheeled autonomous vehicles to assess mobility in real-time. Terrain force estimation can be used to infer the ability to accelerate, climb, or tow a load independent of the underlying terrain model. When a terrain model is available, physical soil properties and stress distribution parameters that relate to mobility are inferred from vehicle-terrain forces using multiple-model estimation. The approach uses Bayesian statistics to select the most likely terrain parameters from a set of hypotheses, given estimated terrain forces. The hypotheses are based on the extensive literature of soil properties for soils with cohesions from 1 - 70 kPa. Terrain parameter estimation is subject to mathematical uniqueness of the net forces resulting from vehicle-terrain interaction for a given set of terrain parameters; uniqueness properties are characterized in the paper motivating the approach. Terrain force and parameter estimation requires proprioceptive sensors - accelerometers, rate gyros, wheel speeds, motor currents, and ground speed. Simulation results demonstrate efficacy of the method on three terrains - low cohesion sand, sandy loam, and high cohesion clay, with parameter convergence times as low as .02 sec. The method exhibits an ability to interpolate between hypotheses when no single hypothesis adequately characterizes the terrain.

An ongoing effort within the US Naval EOD Technology Division (NAVEODTECHDIV) is exploring the integration of autonomous robotic technologies onto current and future Explosive Ordnance Disposal (EOD) robot platforms. The Cooperative Robotics program, though the support of the Joint Ground Robotics Enterprise (JGRE), has identified several autonomous robotic technologies useful to the EOD operator, and with the collaboration of academia and industry is in the process of bringing these technologies to EOD robot operators in the field. Initiated in January 2007, the Cooperative Robotics program includes the demonstration of various autonomous technologies to the EOD user community, and the optimization of these technologies for use on small EOD Unmanned Ground Vehicles (UGVs) in relevant environments. Through close interaction with actual EOD operators, these autonomous behaviors will be designed to work within the bounds of current EOD Tactics, Techniques, and Procedures (TTP). This paper will detail the ongoing and future efforts encompassing the Cooperative Robotics program including: technology demonstrations of autonomous robotic capabilities, development of autonomous capability requirements based on user focus groups, optimization of autonomous UGV behaviors to enable use in relevant environments based on current EOD TTP, and finally the transition of these technologies to current and future EOD robotic systems.

The Army Future Combat System (FCS) Operational Requirement Document has identified a number of advanced robot tactical behavior requirements to enable the Future Brigade Combat Team (FBCT). The FBCT advanced tactical behaviors include Sentinel Behavior, Obstacle Avoidance Behavior, and Scaled Levels of Human-Machine control Behavior. The U.S. Army Training and Doctrine Command, (TRADOC) Maneuver Support Center (MANSCEN) has also documented a number of robotic behavior requirements for the Army non FCS forces such as the Infantry Brigade Combat Team (IBCT), Stryker Brigade Combat Team (SBCT), and Heavy Brigade Combat Team (HBCT). The general categories of useful robot tactical behaviors include Ground/Air Mobility behaviors, Tactical Mission behaviors, Manned-Unmanned Teaming behaviors, and Soldier-Robot Interface behaviors. Many DoD research and development centers are achieving the necessary components necessary for artificial tactical behaviors for ground and air robots to include the Army Research Laboratory (ARL), U.S. Army Research, Development and Engineering Command (RDECOM), Space and Naval Warfare (SPAWAR) Systems Center, US Army Tank-Automotive Research, Development and Engineering Center (TARDEC) and non DoD labs such as Department of Energy (DOL). With the support of the Joint Ground Robotics Enterprise (JGRE) through DoD and non DoD labs the Army Maneuver Support Center has recently concluded successful field trails of ground and air robots with specialized tactical behaviors and sensors to enable semi autonomous detection, reporting, and marking of explosive hazards to include Improvised Explosive Devices (IED) and landmines. A specific goal of this effort was to assess how collaborative behaviors for multiple unmanned air and ground vehicles can reduce risks to Soldiers and increase efficiency for on and off route explosive hazard detection, reporting, and marking. This paper discusses experimental results achieved with a robotic countermine system that utilizes autonomous behaviors and a mixed-initiative control scheme to address the challenges of detecting and marking buried landmines. Emerging requirements for robotic countermine operations are outlined as are the technologies developed under this effort to address them. A first experiment shows that the resulting system was able to find and mark landmines with a very low level of human involvement. In addition, the data indicates that the robotic system is able to decrease the time to find mines and increase the detection accuracy and reliability. Finally, the paper presents current efforts to incorporate new countermine sensors and port the resulting behaviors to two fielded military systems for rigorous assessing.

The operational ability to project and sustain forces in distant, anti-access and area denial environments poses new challenges for combatant commanders. One of the new challenges is the ability to conduct sustainment operations at operationally feasible times and places on the battlefield. Combatant commanders require a sustainment system that is agile, versatile, and survivable throughout the range of military operations and across the spectrum of conflict. A key component of conducting responsive, operationally feasible sustainment operations is the ability to conduct sustainment convoys. Sustainment convoys are critical to providing combatant commanders the right support, at the right time and place, and in the right quantities, across the full range of military operations. The ability to conduct sustainment convoys in a variety of hostile environments require force protection measures that address the enemy threat and protect the Soldier. One cost effective, technically feasible method of increasing the force protection for sustainment convoys is the use of robotic follower technology and autonomous navigation. The Convoy Active Safety Technologies (CAST) system is a driver assist, convoy autopilot technology aimed to address these issues. Warfigher Experiment I, held at A.P. Hill, VA in the fall of 2007, tested the utility of this vehicle following technology not only in measures of system integrity and performance vs. manual driving, but also the physiological effects on the operators themselves. This paper will detail the Warfigher Experiment's methodology, analysis, results and conclusions.

This paper presents the motivation behind the new joint NIST/IEEE Virtual Manufacturing Automation Competition (VMAC). This competition strives to take the Automated Guided Vehicle (AGV) user community driven requirements and turn them into a low-entry-barrier competition. The objectives, scoring, performance metrics, and operation of the competition are explained. In addition, the entry-barrier lowering infrastructure that is provided to competitors is presented.

The Soldier Universal Robot Controller (SURC) is a modular OCU designed for simultaneous control of heterogeneous unmanned vehicles. It has a well defined, published API., defined using XML schemas, that allows other potential users of the system to develop their own modules for rapid integration with SURC. The SURC architecture is broken down into three layers: User Interface, Core Functions, and Transport. The User Interface layer is the front end module which provides the human computer interface for user control of robots. The Core layer is further divided into the following modules: Capabilities, Tactical, Mobility, and World Model. The Capabilities module keeps track of the known robots and provides a list of specifications and services. The Mobility module provides path planning via D*, while the Tactical module provides higher level mission planning (multi-agent/multi-mission) capabilities for collaborative operations. The World Model module is a relational database which stores world model objects. Finally, a Transport module provides translation from the SURC architecture to the robot specific messaging protocols (such as JAUS). This allows fast integration of new robot protocols into an existing SURC implementation to enable a new system to rapidly leverage existing SURC capabilities. The communication between different modules within the SURC architecture is done via XML. This gives developers and users the flexibility to extend existing messages without breaking backwards compatibility. The modularity of SURC offers users and developers alike the capability to create custom modules and plug them into place, as long as they follow the pre defined messaging API for that module.

PRIDE (PRediction In Dynamic Environments) is a hierarchical multi-resolutional framework for moving object prediction. PRIDE incorporates multiple prediction algorithms into a single, unifying framework. To date, we have applied this framework to predict the future location of autonomous vehicles during on-road driving. In this paper, we describe two different approaches to compute long-term predictions (on the order of seconds into the future) within PRIDE. The first is a cost-based approach that uses a discretized set of vehicle motions and costs associated with states and actions to compute probabilities of vehicle motion. The cost-based approach is the first prediction approach we have been using within PRIDE. The second is a fuzzy-logic-based approach that deals with the pervasive presence of uncertainty in the environment to negotiate complex traffic situations. Using the high-fidelity physics-based framework for the Unified System for Automation and Robot Simulation (USARSim), we will compare the performance of the two approaches in different driving situations at traffic intersections. Consequently, we will show how the two approaches complement each other and how their combination performs better than the cost-based approach only.

We address spatial ontologies for the areas of operations of tactical behaviors carried out by unmanned ground vehicles (UGVs). An ontology is a conceptualization of a domain and provides a common vocabulary for automated applications in the domain of interest. Ontological concepts are typically qualitative yet are rigorously defined. An ontology should provide abstract concepts that allow meaningful generalizations. The work reported here is the first known attempt to apply spatial ontologies to tactical behaviors. Some research on spatial ontologies is based on point set topology, although many find points to be unnatural primitives. Alternatives include relations defined in terms of the primitive binary relation "connected_to" on regions; the "part_of" relation is also important. This paper includes a focused survey driven by examples in which we evaluate the strengths and weaknesses of the different approaches for the domain in question. We also develop new concepts and techniques especially applicable to representing and reasoning about areas of operation in which UGVs perform missions.

We present an initial ontology for tactical behaviors conducted by unmanned ground vehicles (UGVs). We focus on activities, which are the denotations of verbs, notably 'move' but also 'look (for)' and several others. These take collective subjects, allowing activities to be attributed to units at various hierarchical levels. The semantics of verbs must consider the denotations of their grammatical complements; that is, we must consider entire verb frames. The thematic relations of the noun-phrase complements are critical, but prepositions also play an important role. FrameNet is an online lexical database of frames derived from text corpora. Our other major resource is Levin's classification of verbs according to how changes in their frames affect their meanings. Although natural languages have a large variety of words for aspects of tactical behaviors, there is motivation to get by with as few basic verbs as possible. A variety of meanings can often be associated with a verb by altering its frame, and we can impose co-reference constraints on combinations of frames to generate structures denoting more complex activities. A simple grammar is developed for the verbs of interest. Protege-Frames ontologies include classes that inherit from linguistically inspired classes but capture domain-specific notions.

This paper presents the libdrdc software standards library including internal nomenclature, definitions, units of measure, coordinate reference frames, and representations for use in autonomous systems research. This library is a configurable, portable C-function wrapped C++ / Object Oriented C library developed to be independent of software middleware, system architecture, processor, or operating system. It is designed to use the automatically-tuned linear algebra suite (ATLAS) and Basic Linear Algebra Suite (BLAS) and port to firmware and software. The library goal is to unify data collection and representation for various microcontrollers and Central Processing Unit (CPU) cores and to provide a common Application Binary Interface (ABI) for research projects at all scales. The library supports multi-platform development and currently works on Windows, Unix, GNU/Linux, and Real-Time Executive for Multiprocessor Systems (RTEMS). This library is made available under LGPL version 2.1 license.

There is a strong demand for efficient explosive detecting devices and deployment methods in the field. In this study we present a prototype mast that uses a telescoping pulley system for optimal performance on top of an unmanned ground vehicle to be able to be controlled wirelessly. The mast and payload reaches up eight feet from the platform with a gripper that can pick up objects. The current mobile platform operators using a remote-control devices to move the arm and the robot itself from a safe distance away. It is equipped with a pulley system that can also be used to extend a camera or explosive detection sensor under a vehicle. The mast is outfitted with sensors. The simple master-slave strategy will not be sufficient as the navigation and sensory inputs will become complex. In this paper we provide a tested software/hardware framework that allows a mobile platform and the expanded arm to offload operator tasks to autonomous behaviors while maintaining tele-operations. This will implement semi-autonomous behaviors. This architecture involves a server which communicates commands and receives sensor inputs via a wireless modem to the mobile platform. This server can take requests from multiple client processes which have prioritized access to on-board sensor readings and can command the steering. The clients would include the tele-operation soldier unit, and any number of other autonomous behaviors linked to particular sensor information or triggered by the operator. For instance, the behavior of certain tasks can be controlled by low-latency clients with sensory information to prevent collisions, place sensor pods precisely, return to preplanned positions, home the units location or even perform image enhancements or object recognition on streamed video.

A 9-DoF mobile robotic manipulator system consisting of a 7-DoF redundant manipulator and a differentially driven 2-DoF mobile non-holonomic platform was mathematically modeled to represent a general redundant mobile manipulator. The control of the 3-degree of redundancy system combines the mobility and manipulation, expands on the conventional control methods and introduces user-specified weights to the singularity-robust (S-R) inverse of the Jacobian. Criterion function weight was added to the weight matrix to optimize the control based on joint limit avoidance. A numerical example to apply and compare several control methods was presented. Singularity and joint limit avoidance along with user-defined motion preference were implemented in simulation. Possible applications in defense were explored.

In our earlier work, we proposed a synergistic design and control strategy to enable a holonomic mobile robot transport wheelchair bound residents in long-term-care facilities. Several simplifying assumptions were made and an adaptive control framework was proposed for wheelchair trajectory tracking. We remove some of the limiting assumptions by considering actuator saturation and actuator dynamics in this paper. We modify the adaptive controller and show that asymptotic trajectory tracking can be achieved provided that the reference trajectory does not result in complete loss of control authority. In the absence of control authority, we suspend trajectory tracking to maintain stability and revert to asymptotic trajectory tracking when control authority is regained. Although we focus on an application for long-term-care facilities, the technology being developed will find diverse applications such as autonomous transportation of hazardous materials.

In this study, behavior generation and self-learning paradigms are investigated for the real-time applications of multi-goal mobile robot tasks. The method is capable to generate new behaviors and it combines them in order to achieve multi goal tasks. The proposed method is composed from three layers: Behavior Generating Module, Coordination Level and Emotion -Motivation Level. Last two levels use Hidden Markov models to manage dynamical structure of behaviors. The kinematics and dynamic model of the mobile robot with non-holonomic constraints are considered in the behavior based control architecture. The proposed method is tested on a four-wheel driven and four-wheel steered mobile robot with constraints in simulation environment and results are obtained successfully.

This paper addresses a number of design issues that are associated with integrating lightweight and low-cost RF (Radio Frequency) sensors onto small UGV's (Unmanned Ground Vehicles) and UAV's (Unmanned Aerial Vehicles). Modular\integrated RF sub-systems functions that are discussed include lightweight software programmable radar (or software radar) using COTS software radio components and compact microstrip antenna design concepts for low-frequency surface penetration radars. A discussion on the potential for implementing lightweight multi-function RF systems as well as a discussion on novel futuristic concepts that explore the limits of sensor/platform integration is included.