Multilevel Adaptation in Teams of Unmanned Air and Ground Vehicles
Charles L. Ortiz, Régis Vincent, Pauline M. Berry, Andrew Agno
- Year
- 2002
- Citations
- 2
Abstract
We describe ongoing work to develop algorithms and software for the reprogrammable, coordinated command and control of teams of autonomous vehicles (AVs). This new software capability extends current work in distributed artificial intelligence and is intended to allow developers to build and test AV teams that collaboratively perform tasks - such as reconnaissance and surveillance - with minimal supervision in dynamic, unstructured environments. We describe a multilevel robot architecture which is adaptive along a number of dimensions, as well as several new technologies that have been developed to support behavior blending and inter-robot negotiation. 1. INTRODUCTION Over the last few years, and under ONR funding, we have been developing algorithms and software for the reprogrammable, coordinated command and control of teams of autonomous vehicles (AVs). This new software capability extends work in distributed artificial intelligence and is intended to allow developers to build and test AV teams that collaboratively perform tasks - such as reconnaissance and surveillance - with minimal supervision in dynamic, unstructured environments. In this paper, we describe our progress in realizing this vision. As motivation, Figure 1 illustrates a futuristic scenario involving the suppression of enemy air defenses (SEAD) using a team of UAVs. The icons representing the surveillance mission, helicopter mission, cover mission, and strike mission are all assumed to represent unmanned combat air vehicles (UCAVs). Control centers are shown on land, air, and sea. The order of events is as follows. The two surveillance UCAVs are launched from the ship to the surface-to-air missile (SAM) site to collect intelligence for prestrike planning. Using the information obtained, the operator then schedules the cover and strike missions. Two escorts responsible for the cover mission are subsequently launched and meet with a pair of UCAVs responsible for the air strike. During ingress, this SEAD package encounters an enemy mobile SAM site that attempts to break up the package. The escorts prcvide cover so that the strike mission can proceed. At the same time, a helicopter UCAV is launched to conduct mine countermeasure operations.
Keywords
Related papers
Statistical Learning Theory
Yuhai Wu, Vladimir Vapnik
1999
Artificial intelligence: a modern approach
1995
Fractional Differential Equations
Igor Podlubný
2025
Applied Nonlinear Control
Jean-Jacques Slotine, Weiping Li
1991