Unmanned Aerial Vehicles (UAVs) have demonstrated tremendous capability in recent military operations. Recently swarm technology has been suggested as a possible solution to automatically control and coordinate multiple UAVs. The idea behind a swarm is that simple local rules that govern the behavior of individual entities can lead to complex emergent behavior of the system as a whole. Although such systems have achieved limited success in simulated applications, finding good rules can be difficult for humans. Moreover, such rules can result in odd behavior or unnecessarily long missions. This paper describes a swarm-based multi-UAV system, called SmartSwarms, using a radically different approach: instead of operating with human-defined rules, each individual reasons using Simulated LookAhead (SLA), thus incorporating a model of its world and nearby entities in decision-making. Our results show that this approach can improve swarm behavior in UAVs. SLA is affordable, scalable to a large number of UAVs, deconflicts in real-time, learns over time, is interoperable, reusable, fault tolerant, and error tolerant, and can handle uncertainty.
SmartSwarms: Distributed UAVs that Think
2004
23 pages
Report
Keine Angabe
Englisch
Aircraft , Computer Software , Military Operations, Strategy, & Tactics , Military operations , Decision making , Artificial intelligence , Drones , Interoperability , Reusable equipment , Simulation , Scaling factor , Models , Fault tolerant computing , Real time , Swarm , Sla(Simulated lookahead) , Agent-based combat modeling , Swarms
Distributed Finite-Time FTCC of Multi-UAVs with Multiple Leader UAVs
Springer Verlag | 2023
|What Do I Think About Machines that Think?
NTRS | 2015
|A Distributed Avionics Package for Small UAVs
AIAA | 2005
|Engineering Index Backfile | 1953
|Distributed Multi-UAVs Coupled Tasks Consensus Assignment
Springer Verlag | 2023
|