Adaptive Remote-Sensing Techniques Implementing Swarms of Mobile Agents
This paper focuses on our recent work at Sandia National Laboratories toward engineering a physics-based swarm of mobile vehicles for distributed sensing applications. Our goal is to coordinate a sensor array that optimizes sensor coverage and multivariate signal analysis by implementing artificial intelligence and evolutionary computational techniques. These intelligent control systems integrate both globally operating decision-making systems and locally cooperative information-sharing modes using genetically-trained neural networks. Once trained, neural networks have the ability to enhance real-time operational responses to dynamical environments, such as obstacle avoidance, responding to prevailing wind patterns, and overcoming other natural obscurants or interferences (jammers). The swarm realizes a collective set of sensor neurons with simple properties incorporating interactions based on basic community rules (potential fields) and complex interconnecting functions based on various neural network architectures, Therefore, the swarm is capable of redundant heterogeneous measurements which furnishes an additional degree of robustness and fault tolerance not afforded by conventional systems, while accomplishing such cognitive tasks as generalization, error correction, pattern recognition, and sensor fission. The robotic platforms could be equipped with specialized sensor devices including transmit/receive dipole antennas, chemical or biological sniffers in combination with recognition analysis tools, communication modulators, and laser diodes. Our group has been studying the collective behavior of an autonomous, multi-agent system applied to emerging threat applications. To accomplish such tasks, research in the fields of robotics, sensor technology, and swarms are being conducted within an integrated program. Mission scenarios under consideration include ground penetrating impulse radar (GPR) for detection of under-ground structures, airborne systems, and plume detection/remediation. We will describe our research in these areas and give a status report on our progress, including simulations and laboratory-based sensor experiments.
- Research Organization:
- Sandia National Lab. (SNL-NM), Albuquerque, NM (United States); Sandia National Lab. (SNL-CA), Livermore, CA (United States)
- Sponsoring Organization:
- US Department of Energy (US)
- DOE Contract Number:
- AC04-94AL85000
- OSTI ID:
- 5682
- Report Number(s):
- SAND99-0770C; TRN: AH200115%%89
- Resource Relation:
- Conference: SPIE 13th Annual International Symposium on Aerospace/Defense Sensing, Simulation, and Controls, Orlando, FL (US), 04/05/1999--04/09/1999; Other Information: PBD: 1 Apr 1999
- Country of Publication:
- United States
- Language:
- English
Similar Records
Distributed Sensing and Cooperating Control for Swarms of Robotic vehicles
Integrate Light-Weight Deep Learning Tools with Internet of Things