N02-097 TITLE: Cooperative Behavior and Control in Groups of Unmanned Air Vehicles (UAVs)

TECHNOLOGY AREAS: Air Platform, Information Systems

DOD ACQUISITION PROGRAM SUPPORTING THIS TOPIC: ACAT IV: PMR 51 Navy Low/Counterlow Observables Policy, Technology and Projects Office

OBJECTIVE: Develop control system approaches suitable for the intra-group command and control in expendable UAVs. Assess the extent of reliable communication required to ensure robust tactical performance. Identify limitations and advantages in particular control architectures.

DESCRIPTION: Small, low cost expendable UAVs are becoming more common. Their perceived value is not only their ability to accomplish the traditional functions of UAVs but also their ability to operate cooperatively in tactical situations where high attrition is expected. This cooperative behavior would allow groups of UAVs to identify their losses internally, reconfigure the force to accommodate these losses, select the most appropriate new "leader(s)", recombine with remnants of other groups when losses became too extensive, etc. Researchers have suggested several potential control architectures to implement these functions. These span the spectrum from very tightly controlled hierarchical designs through the almost flat designs suggested by the proponents of Artificial Intelligence (AI). The advantages and limitations of this spectrum of architectures needs to be examined for both their utility and their demand on communications bandwidth and reliability to accomplish their tactical goals.

This topic examines the computer simulation and real-world assessment of a variety of robust control structures for a group of expendable UAVs that would have a number of tactical functions in a battlespace. These functions would include (but not be limited to) reconnaissance, search, tracking, relay communications, target identification, and navigational guidance. The group would consist of varying amounts of redundant capability and subjected to a range of subsystem degradation (especially communications) in individual vehicles and total vehicle attrition rates. The simulations would track the average and peak communication rates (inter-vehicle), degradation in the tactical functionality of the total group in face of attrition, and the span, extent, and reliability of the information provided back to the command center. Hybrid or novel architectures are encouraged, particularly to overcome observed weaknesses and limitations identified in previous simulations.

A further aspect to this effort is the development of gradient control, i.e. from man-in-the-loop to supervisory to full autonomous control. This aspect is essential to allow a system to be tested and evaluated as it moves through the process of modeling into real vehicles.

PHASE I: Define potential architectures that can address the control of an expendable UAVs having a variety of capabilities. Perform a preliminary assessment of the architecture using a computer simulation of a UAV force of 10 vehicles selected from a group of vehicles possessing the functions of reconnaissance, search, tracking, relay communications, target identification, and navigational guidance. Determine the minimum level and frequency rate of communication necessary to maintain stable tactical group behavior. Document the results in a report identify portions of the architecture and algorithms exhibiting robust behavior and those portions requiring additional research.

PHASE II: Conduct a more detailed and realistic computer simulation than in Phase I, using a UAV force of 100 vehicles with the functions of reconnaissance, search, tracking, relay communications, target identification, and navigational guidance (as well as other functions to be provided). This simulation will include actual, observed vehicle characteristics measured in the field, and will use the results from sensor and communications system evaluations to enhance simulation accuracy. Determine the extent of redundancy required to accomplish a mission while the group is subjected to degradation of both its inter-vehicle and command communications and also experiences a variety of total vehicle attrition rates. Determine the minimum level and frequency rate of communication necessary to maintain stable tactical group behavior. Determine the capability to reconstitute tactical group behavior after the reestablishment of communications following communication pauses from seconds to hours. Document in detail the best (most robust) architecture in a final report.

PHASE III: Design, test, evaluate and integrate the most robust architecture into a hardware test control system suitable for use on a small, expendable UAV. Use a group of 30 actual UAVs for a battlefield simulation. Correlate battlefield and computer simulations and modify architectures and algorithms. Provide complete control system for field use.

COMMERCIAL POTENTIAL: Real time, robust control architectures are in demand for a variety of manufacturing and service operations. As these systems become more complicated, a more generic approach to them has become essential.

KEYWORDS: Control, Command, Navigation, Cooperative Behavior