TY - GEN
T1 - Foraging-Inspired Self-Organisation for Terrain Exploration with Failure-Prone Agents
AU - Rodriguez, Arles
AU - Gomez, Jonatan
AU - Diaconescu, Ada
N1 - Publisher Copyright:
© 2015 IEEE.
PY - 2015/10/23
Y1 - 2015/10/23
N2 - Mobile ad-hoc sensor systems are employed increasingly for distributed tasks in unreliable conditions, such as terrain exploration and measuring. Here, self-organising solutions can help ensure reliability, availability and scalability, while making use of unreliable components (or agents) with limited resources. These enable agents to act independently, and to exchange and combine their partial solutions into a (more) complete result, which can be transmitted to users before all agents fail. In previous work, we have studied how foraging-inspired self-organisation can help mobile agents achieve a collaborative task - terrain exploration and information gathering. Obtained results revealed two key aspects impacting success rates: the strategy for exploring as much uncharted terrain as fast as possible, and the strategy for encountering the highest number of agents that hold complementary information. In this paper, we explore further techniques for these aspects and introduce passive pheromone evaporation. Our results show that a hybrid approach improving exploration efficiency features higher success rates than basic stigmergy models, random exploration and Lévy walks.
AB - Mobile ad-hoc sensor systems are employed increasingly for distributed tasks in unreliable conditions, such as terrain exploration and measuring. Here, self-organising solutions can help ensure reliability, availability and scalability, while making use of unreliable components (or agents) with limited resources. These enable agents to act independently, and to exchange and combine their partial solutions into a (more) complete result, which can be transmitted to users before all agents fail. In previous work, we have studied how foraging-inspired self-organisation can help mobile agents achieve a collaborative task - terrain exploration and information gathering. Obtained results revealed two key aspects impacting success rates: the strategy for exploring as much uncharted terrain as fast as possible, and the strategy for encountering the highest number of agents that hold complementary information. In this paper, we explore further techniques for these aspects and introduce passive pheromone evaporation. Our results show that a hybrid approach improving exploration efficiency features higher success rates than basic stigmergy models, random exploration and Lévy walks.
KW - decentralized control
KW - failures
KW - foraging
KW - information gathering and sharing
KW - mobile sensors
KW - multi-agent systems
KW - reliability
KW - self-organising systems
KW - terrain exploration
U2 - 10.1109/SASO.2015.20
DO - 10.1109/SASO.2015.20
M3 - Conference contribution
AN - SCOPUS:84959261245
T3 - International Conference on Self-Adaptive and Self-Organizing Systems, SASO
SP - 121
EP - 130
BT - Proceedings - 2015 IEEE 9th International Conference on Self-Adaptive and Self-Organizing Systems, SASO 2015
PB - IEEE Computer Society
T2 - 9th IEEE International Conference on Self-Adaptive and Self-Organizing Systems, SASO 2015
Y2 - 21 September 2015 through 25 September 2015
ER -