engleski

Resource Management in Cooperative Multi-Agent Networks Through Self-Triggering

In this study, the author’s interest is to reduce requirements posed on communication, sensing, processing and energy resources in multi-agent networks (MANs) without compromising MAN objectives. Consequently, the hardware expenses and energy consumption are driven down while MAN multitasking, such as inter-network collaboration, is facilitated. Scenarios, in which agents of one network need to achieve a common goal, call for the study of decentralised cooperative control schemes. To accomplish this goal in an uncertain and noisy setting and to detect changes in the communication topology, agents have to exchange information. Since each transmission and reception of information necessitates energy, communication should be induced only when the goal completion can no longer be guaranteed in order to prolong the MAN mission. Thus, the authors devise an information exchange mechanism which draws upon the ideas of self-triggered communication. The proposed mechanism is inspected both theoretically and experimentally (employing off-the-shelf wireless sensor platforms) for performance against lifetime trade-offs using a single-integrator consensus case study. Their mechanism is applicable to heterogeneous agents with exogenous disturbances, to directed switching communication topologies and to both initial-condition-dependent and initial-condition-independent long-term cooperative behaviours. The investigated stability notions include Lp- and input-to-state stability (with respect to a set).

Sampled Data Systems; Hybrid Systems; Robust Control; Linear Systems

nije evidentirano