Stochastic Movement Swarm Performing a Coverage Task with Physical Parameters

Krzysztof FALKOWSKI; Krzysztof FALKOWSKI

Pobierz

Ta publikacja jest chroniona prawem autorskim. Dostęp do jej cyfrowej wersji jest możliwy po zalogowaniu.

Tytuł: Stochastic Movement Swarm Performing a Coverage Task with Physical Parameters ; Stochastic Movement Swarm Performing a Coverage Task with Physical Parameters

Abstrakt:

This paper describes an attempt of implementing physical parameters into a virtual swarm algorithm solution. It defines which physical parameters of the single object need to be known to properly transfer a virtual algorithm into a physical system. Considerations have been based on a stochastic movement swarm performing a coverage task. Time to finish the task and energy consumptions were measured for different numbers of drones in a swarm allowing to designate an optimal size of the swarm. Additional tests for changing variables allowed us to determine their impact on the swarm performance. The presented algorithm is a discrete-time solution, and every test is divided into steps. Positions of the drones are calculated only in time corresponding to these steps. Their position is unknown between these steps and the algorithm does not check if the paths of two drones cross between subsequent positions. The lower the time interval, the more precise results, but simulating the test requires more computing power. Further work should consider the smallest possible time intervals or additional feature to check if the paths of the drones do not cross.
; This paper describes an attempt of implementing physical parameters into a virtual swarm algorithm solution. It defines which physical parameters of the single object need to be known to properly transfer a virtual algorithm into a physical system. Considerations have been based on a stochastic movement swarm performing a coverage task. Time to finish the task and energy consumptions were measured for different numbers of drones in a swarm allowing to designate an optimal size of the swarm. Additional tests for changing variables allowed us to determine their impact on the swarm performance. The presented algorithm is a discrete-time solution, and every test is divided into steps. Positions of the drones are calculated only in time corresponding to these steps. Their position is unknown between these steps and the algorithm does not check if the paths of two drones cross between subsequent positions. The lower the time interval, the more precise results, but simulating the test requires more computing power. Further work should consider the smallest possible time intervals or additional feature to check if the paths of the drones do not cross.