Resource-Saving Multiobjective Task Distribution in the Fog- and Edge-Robotics

Abstract

In the current paper the question of the resource-saving tasks distribution in the robotic groups is under consideration. As a wide range of computational tasks in robotics are performed in a distributed manner, tasks can be assigned to the devices with a relatively low computational capacity. At the same time, data preprocessing, machine learning, SLAM problems are computationally complex, and so the participants of the computational process can be overloaded, while the latter causes the deterioration of average residual life of the computational nodes within the robots. In this paper the problem of resource-saving tasks distribution is formulated as structural-parametric multiobjective one, with paying attention to the workload of those robots in the group, which have to transmit sensor data. The general solution technique is proposed based on global problem decomposition, local time constraints estimations and simulated annealing technique. The a priory time estimations are used according to the tasks graph analysis, as well as time constraints are divided into shares considering the number of transit nodes. Also, some selected experimental results are presented, as well as comparison with the previously conducted results are made.