Autonomous Learning of Target Decision Strategies without Communications for Continuous Coordinated Cleaning Tasks

Yoneda, Keisuke; Kato, C.; Sugawara, Toshiharu

doi:10.1109/wi-iat.2013.112

Cited by 8 publications

(10 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Note that agents with the AMTDS are assumed to know the dirt accumulation probabilities [11], so P i v = Pv for ∀v ∈ V and ∀i ∈ A and do not learn P i v . The proposed method in this paper is called AMTDS with learning of dirt accumulation probabilities (AMTDS/LD), after this.…”

Section: Proposed Methodsmentioning

confidence: 99%

“…Before describing the model, we will explain three assumptions that were also assumed in [11]. First, agents, which are control programs on portable cleaner robots capable of autonomously deciding their actions, know their own and others' locations.…”

Section: Assumptionsmentioning

confidence: 99%

“…This section briefly describes a number of path planning strategies used in our experiments below, which correspond to the elements of S. The strategies used here are identical to those used in [11], since we would like to compare the difference of performance their method and ours.…”

Section: Path Planning Strategiesmentioning

confidence: 99%

“…Cheng and Dasguputa [3] studied a decentralized management system where the communication range and memory were limited and assumed noise was included in communication. Yoneda et al [11] proposed a method, called the adaptive meta-target decision strategy (AMTDS), in which agents autonomously learned strategies to determine targets to move to and clean next according to knowledge on where dirt easily accumulates but did not perform complex reasoning for coordination/cooperation (so lightweight cooperation), such as recognition of plans in other agents and negotiation among agents. However, they assumed that this knowledge was provided to agents in advance, although such environmental information is often unknown in real situations.…”

Section: Introductionmentioning

confidence: 99%

“…Therefore, we have extended Yoneda et al's method [11] by incorporating it with learning to identify the locations that are easy to become dirty, in order to make it applicable to situations where the environmental information is unknown. In the beginning, agents assume that the deployed area in the environment is uniformly made dirty.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Meta-strategy for cooperative tasks with learning of environments in multi-agent continuous tasks

Sugiyama

Sugawara

2015

Proceedings of the 30th Annual ACM Symposium on Applied Computing

Self Cite

View full text Add to dashboard Cite

With the development of robot technology, we can expect selfpropelled robots working in large areas where cooperative and coordinated behaviors by multiple (hardware and software) robots are necessary. However, it is not trivial for agents, which are control programs running on robots, to determine the actions for their cooperative behaviors, because such strategies depend on the characteristics of the environment and the capabilities of individual agents. Therefore, using the example of continuous cleaning tasks by multiple agents, we propose a method of meta-strategy that decide the appropriate planning strategies for cooperation and coordination through with the learning of the performance of individual strategies and the environmental data in a multi-agent systems context, but without complex reasoning for deep coordination due to the limited CPU capability and battery capacity. We experimentally evaluated our method by comparing it with a conventional method that assumes that agents have knowledge on where agents visit frequently (since they are easy to become dirty). We found that agents with the proposed method could operate as effectively as and, in complex areas, outperformed those with the conventional method. Finally, we describe that the reasons for such a counterintuitive phenomenon is induced from splitting up in working by autonomous agents based on the local observations. We also discuss the limitation of the current method.

show abstract

Section: Proposed Methodsmentioning

confidence: 99%

Section: Assumptionsmentioning

confidence: 99%

Section: Path Planning Strategiesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations