Showing 1–2 of 2 results for author: Savkin, A V

Optimized Deployment of Autonomous Drones to Improve User Experience in Cellular Networks

Authors: Hailong Huang, Andrey V. Savkin, Ming Ding, Mohamed Ali Kaafar

Abstract: Modern wireless traffic demand pushes Internet Service Providers to develop effective strategies to improve user experience. Since deploying dense Base Stations (BSs) is not cost efficient, an alternative is to deploy autonomous drones to supplement existing BSs. A street graph is adopted to represent the area of interest. The outdoor User Equipments (UEs) to be served locate near streets and the… ▽ More Modern wireless traffic demand pushes Internet Service Providers to develop effective strategies to improve user experience. Since deploying dense Base Stations (BSs) is not cost efficient, an alternative is to deploy autonomous drones to supplement existing BSs. A street graph is adopted to represent the area of interest. The outdoor User Equipments (UEs) to be served locate near streets and the 2D projections of drones are restricted to streets to avoid collision with buildings. We build up a UE density function based on a real dataset, reflecting the traffic in the area. We study four problems: where to deploy single drone to cover maximum UEs, where to deploy $k$ drones cover maximum UEs subject to an inner drone distance constraint, where to deploy $k$ drones cover maximum UEs subject to inner drone distance constraint and drones' battery constraints, and the minimum drones to cover a given percentage of UEs subject to inner drone distance constraint. We prove that the latter three problems are NP-hard and propose greedy algorithms with theoretical analysis. To our best knowledge, this is the first paper to consider the battery constraints for drone deployments. Extensive simulations have been conducted to verify the effectiveness of our approaches. △ Less

Submitted 6 December, 2017; originally announced December 2017.

Proofs of the Technical Results Justifying an Algorithm of Reactive 3D Navigation for a Surface Scan by a Nonholonomic Mobile Robot

Authors: Alexey S. Matveev, Kirill S. Ovchinnikov, Andrey V. Savkin

Abstract: A single nonholonomic under-actuated mobile robot with a bounded control range travels in a 3D workspace. This workspace also hosts a compact unknown domain with a smooth boundary. The robot has access to the horizontal distance to this domain and to a certain direction (typically, vertical) in the space, along with its coordinate (typically, altitude) in this direction. We present a new navigatio… ▽ More A single nonholonomic under-actuated mobile robot with a bounded control range travels in a 3D workspace. This workspace also hosts a compact unknown domain with a smooth boundary. The robot has access to the horizontal distance to this domain and to a certain direction (typically, vertical) in the space, along with its coordinate (typically, altitude) in this direction. We present a new navigation law that drives the robot to the desired distance to the domain and ensures subsequent full scan of its surface within a desired range of "altitudes". The proposed strategy is computationally inexpensive and achieves full scan at the lowest control level via generating the current control as a simple, reflex-like reaction to the current observation. The paper presents and proves key technical facts underlying mathematically rigorous analysis and justification of the proposed guidance approach. △ Less

Submitted 9 May, 2016; originally announced May 2016.