|
Test & Evaluation
OSD Sense-and-Avoid Flight Demonstration
David Gibbs, Ryan Schaefer, SRA International, Inc, Dyke Weatherington, OUSD/AT&L
The future of unmanned aviation rests, in a large part, to the ability of unmanned aircraft systems (UASs) to gain entrance to the National Airspace System (NAS). Current FAA restrictions prevent UASs from flying in the NAS due their inability to "see and avoid" other aircraft as required by CFR 91.113. Until UASs can demonstrate an "equivalent level of safety" to see and avoid, they will be restricted from the NAS. Taking the first step to solve this problem, SRA International, inc., in conjunction with Geneva Aerospace, Inc.; FlightStar Sport Planes, Inc.; the Tactical Analysis and Applications Center; and the USAF 46th Test Group; are working to develop a technology demonstration of an autonomous sense and avoid system.
This See & Avoid Flight Demonstration (SAFD) is to demonstrate an automated see & avoid (S&A) capability on an unmanned aircraft operating in unrestricted airspace. The purpose of the demonstration is to develop and refine detailed requirements for follow-on S&A systems for use in operational unmanned aircraft. The goal of the demonstration is to show current technology can provide a S&A system that meets the "see and avoid" and "equivalent level of safety" requirements of the FAA to allow unmanned aircraft safe, routine access to unrestricted airspace.
The SAFD will be conducted in a build-up approach in three flight phases, manned v. manned intruders testing, unmanned v. manned intruders testing, concluding with the demonstration for the FAA. Testing will be conducted against a variety of airborne objects ("intruders"), including tethered balloons and manned general aviation aircraft. Testing will progress from manned to unmanned flights and from stationary to maneuvering, then multiple, intruders.
Measuring Unmanned Vehicle System Performance: Challenges and Opportunities
Dr. Jerry Harbour, David Bruemmer, Douglas Few, Idaho National Laboratory
Technological advances are taking place within the unmanned vehicle system community at an ever accelerating pace. Yet despite such technical advances, there are few published studies that attempt to quantitatively assess the empirical relationship between innovations in unmanned vehicle systems and actual gains in performance. That is, how have (and have not) unmanned vehicle technologies fundamentally improved operational performance. To begin answering this important question, it is imperative that appropriate sets of operationally-defined metrics be developed that quantitatively compare and contrast both within unmanned vehicle system performance and between manned and unmanned system performance. Ongoing work at the Idaho National Laboratory involving multiple experiments in human-ground robot interaction coupled with measured field trials associated with the Autonomous (ground) Robotic Countermine System (ARCS) provide valuable insights into how unmanned system performance metrics may be formally developed and applied. It is suggested that the development, standardization, and acceptance of such quantitative measures will permit better and more informed decisions regarding unmanned vehicle system operational applicability, selection, and benefit, as well as trend and track ongoing improvement efforts. Additionally, quantitative metrics can assist government and private sector organizations alike in better assessing their Return on Innovation Investment (ROI2) in unmanned vehicle system technologies.
Use of Integrated Test to Reduce Cost and Schedule for UAS Acquisition
LCDR Jeff Laugle, US Navy COMOPTEVFOR
We all want "it" faster, better, and cheaper – whatever "it" is. Within today’s world of reduced budgets it is imperative that Operational Testers be involved as early as possible in the acquisition process. That can be accomplished through COMOPTEVFOR’s IT (Integrated Test) Planning Methodolgy.
Within DoD Acquisition, Joint and Service specific requirements become needed capabilities that have are expressed in a common language. A system will be designed to fill those capability gaps. That system will then be developed and tested by the contractors, development engineers, and potential end-users. Previously, each group conducted their own testing with OPEVAL being the “final exam” at the end. This led to inefficiencies that cost money and schedule when changes had to be applied to make a system effective and suitable for operational use. The IT Planning Methodology, adopted in 2005, allows for coordinated earlier involvement by the Testers so necessary changes may be made when a system is still in the engineering and development stage.
Through a program’s documentation and using the IT Planning Methodology, COMOPTEVFOR is able to give the US Navy’s Unmanned Aerial Systems the “open-book test study-guide” near the beginning of the system’s life-cycle. Here’s how it’s done.
Object Recognition and Reasoning for an Autonomous Maritime Navigation System
Les Elkins, Spatial Integrated Systems, Inc, Joe Fuller, Marshall University
Spatial Integrated Systems (SIS) of Rockville, Maryland, in collaboration with NSWC Carderock Division (NSWCCD), is applying digital 3D imaging technology, Artificial Intelligence (AI), sensor fusion, and system integration on a prototype nominal 40 foot, high performance Unmanned Surface Vehicle (USV) research and development vessel. The ongoing Autonomous Maritime Navigation (AMN) project has been discussed at AUVSI before. Following last year's paper discussing AMN and object recognition, this paper will discuss our continuing investigation into 3D object characterization and recognition for autonomous navigation. It will also discuss our efforts to improve the camera array data to generate better data for recognition. Our goal is to feed the AMN system's data into our C3 ( Composeable Command Center) system using its agent based framework to support data streams, applications, and displays for user awareness. The use of the C3's AI behavior tools to easily generate behavior scripts will also be discussed, as well as the use of intelligent agents to transport data from the vehicle to user displays as well as the boat's control system.
USDA Forest Service Small UAS Demonstration Series
Thomas Zajkowski, Everett Hinkley, USDA Forest Service (RSAC), Dr. Vince Ambrosia, Randell Berthold, NASA-Ames
The United States Forest Service in partnership with NASA-Ames research center has embarked upon a small UAS demonstration series to test the feasibility of using small unmanned aerial systems to support a variety wildland fire missions. The goal of this program is to develop UAS operational guidelines, safety standards, and contracting rules so that firefighters can include small UAS in their toolkit.
Currently wildland firefighters use a variety of manned fixed and rotary winged aircraft in a variety of mission including. The Forest Service feels that UAS can augment manned aircraft in missions that are underserved or where manned aircraft are not practical due to duration or safety issues.
The small UAS demonstrations series began in July 2005 at Moffitt Field, CA with three UAS flying in front of a group of wildland fire professions. A second demonstration incorporating their suggestions will be flown June 2006 utilizing 5 UAS and will explore how small unmanned vehicles can be best used by firefighter in operational conditions. Future demonstrations planned for fall 2006 and Spring/Summer 2007 will provide the Forest Service with the experience needed to contract and operate small UAS in regular operations over wildland fires throughout the United States.
Inverting the Ratio: One Operator to Many Vehicles
John Molberg, Terry Bandzul, CDL Systems
To make wider use of unmanned vehicle systems, it will be necessary to migrate from a one-vehicle, multiple-operator model to a one-operator, multiple-vehicle model. This transition has already taken place with target drones and boats. Meggitt Defence Systems Canada and EADS routinely operate two targets per operator, and with networked workstations have operated four jet-engined target drones or four unmanned sea surface vessels with just two operators.
This work is now being extended in a cooperative research program between Meggitt and CDL Systems to enable a team of four operators to control up to sixteen unmanned, high speed patrol boats. This is in support of work for the Canadian Navy to test defences against a coordinated attack of multiple Fast Incoming Attack Craft (FIAC).
Development is based on the CDL Systems' Vehicle Control System (VCS) software. Using a mission-centric, and generic GUI approach, operator workload is reduced and threat simulation capability is increased. Further, the data link interface conforms to NATO Standardization Agreement (STANAG) 4586, which facilitates interoperability. STANAG 4586 has been successfully used to operate a number of different UAVs - this will be one of the first times that it has been used to control USVs.
Queuing Theory Applied to UGV Operator Workload
Dr. Barry Bodt, U.S. Army Research Laboratory
We advance a queuing theory framework for exploring a key aspect of operator workload and span of control in the operation of unmanned ground vehicles (UGVs). Operator intervention is still sometimes necessary in experimentation with UGVs, even as autonomous mobility technology matures. The frequency and duration of intervention events places a stochastic limit on how much operator workload remains for mission tasks or, alternatively, how much operator workload remains for oversight of additional UGVs. In 2003, the U.S. Army Research Laboratory concluded an extensive experiment to assess the maturity of AM for UGVs. The UGVs in that study logged approximately 530 kilometers and 90 hours of autonomous operation over varying conditions, providing a large database of performance that included intervention events. By applying the queuing model to this baseline data we show the stochastic workload on operators that can be associated with AM technology at the time. We express that workload with varying levels of span of control involving multiple operators and multiple UGVs. If available, data from a planned 2006 experiment using current AM technology will be incorporated for workload comparison with the 2003 baseline.
Multi-Modal Command and Control in Embedded Unmanned System Training
Dr. Peter Drewes, CESI, Tim Roberts, Pat Garrity, RDECOM
RDECOM has been involved in research into tactical behaviors and multi-modal interfaces to control unmanned systems. This allows for higher level training and fielding capabilities. The unmanned systems of tomorrow will be integrated into individual and unit level operations and will require more resources than just tele-operation and remote piloting. This drives the need for advanced unmanned tactical behaviors and effective command and control methodologies. Since many of the unmanned systems explored are not yet fielded, an embedded training approach was utilized to test the capabilities, limitations and doctrinal implications of the research.
The goals of this RDECOM funded research included multi-unit tactical operations, scalability limitations between the air and ground unmanned system domains, training effectiveness, and overall simulation efficiency to supplement live exercises. This paper presents the tactics, techniques and procedures used, and their applicability to the multi-level training of soldiers operating with new air and ground unmanned systems, and expanding their effectiveness within their existing virtual and constructive environments.
This paper covers the following topic areas:
1) Embedded training methodologies
2) Approaches to unmanned tactical behavior testing
3) Architectural limitations and surrogate usage
4) Lessons learned
5) Applicability to the networked battlespace through the multi-level training
U.S. Army Family of UAS
COL Donald Hazelwood, US Army UAS Project Office
The Unmanned Aircraft Systems (UAS) Project Office provides a total Army perspective for the Life Cycle Management of the Army’s Unmanned Aerial Vehicle (UAV) Programs. PM UAS is responsible for management of the entire fleet of Army UAVs. From our hand launched systems to our high flying weaponized platforms, the warfighter is always central to the mission of PM UAS. Equipping every level of soldier, the UAS Project Office fortifies the Army with the technological superiority needed to provide a decisive advantage in today’s intelligence war. UAVs reduce the loss of life risk, while providing valuable information regarding enemy movements to tactical commanders. The UAV mission is growing rapidly to include weaponization, communications relay, specialty payload, and the linkage to manned aircraft. As the arsenal of unmanned vehicles expands, a One System approach ensures all aircraft and hardware work as a single unit, thus reducing costs, capitalizing on utility, and providing a decisive edge.
NUSE2 Perimeter Security Experiment
Kevin Hodges, US Air Force AFRL
The Air Force Research Laboratory, Robotics Research Group conducted an extended field experiment in May 2005 sponsored by the Joint Robotics Program Manager as a National Unmanned Systems Experimentation Environment (NUSE2) event. The experiment was conducted at F.E. Warren AFB, WY in order to place prototype unmanned systems into the hands of intended users in a relative environment.
The experiment was conducted to gather performance requirements for unmanned systems operating for extended periods performing persistent security operations. AFRL used the experiment to capture research requirement for advanced robotic behaviors and human/machine interface. Lessons learned include; 1) Unmanned systems must have a high degree of integration with existing security systems. 2) Unmanned systems must have a high degree of autonomy in the areas of navigation, incident response, stand-off alarm assessment, and intruder challenge. 3) Non-lethal and lethal weapons payloads must provide systems assisted target acquisition and tracking capability. The results of the experiment have been incorporated into the AFRL Robotics Research and Development roadmap for advanced behaviors and serves as the starting point for the next spiral development called Automated Perimeter Security. This presentation provides an overview of the NUSE2 Perimeter Security Experiment, system configuration, experimental results, and future plans.
| 
Events
