NeTS: Small: Collaborative Research: Infrastructure Mobility

移动的计算传统上意味着连接到静态基础设施的移动的客户端。这个项目打破了这种观点，并设想了将移动性注入基础设施的可能性。PI设想了一个车轮上的WiFi接入点，它可以移动以优化所需的性能指标。移动不一定要在家里或办公室的地板上到处都是，也不一定要用电池操作，或者无线连接到互联网。在家里，他们可以保持拴在电源和以太网插座，而在小区域移动（也许在学习桌下）。在未来的办公室里，也许AP可以在安装在假天花板上的轨道上移动。云可能会根据用户在环境中的移动方式，或者网络中流量和干扰的变化来协调它们的移动性。该项目探讨了这一愿景的可行性，并将提供来自各种家庭/办公室环境的全面测量。在住宅和企业环境中移动WiFi AP的初步测量显示出了希望。分析发现，室内环境复杂的多径特性会导致链路质量的大幅波动，即使天线在一英尺范围内移动。移动的AP可以通过重新定位到对于其自身客户端较强而对于其服务器较弱的像素来利用这种空间变化。当多个AP由中央控制器协调时，需要联合规划AP的运动路径以优化全局网络参数。将开发启发式，首先假设AP位置的知识;基于此启发式的结果，假设可能会放宽。该项目还将考虑客户端的移动性以及AP如何适应它们，可能是基于当今WiFi卡中的信道状态信息（CSI）的信道变化质量。如果移动性被证明可以在当前技术进步的情况下提供额外的收益，那么机器人无线网络可能会成为未来一个重要而令人兴奋的方向。

Mobile computing has traditionally implied mobile clients connected to a static infrastructure. This project breaks away from this point of view and envisions the possibility of injecting mobility into infrastructure. The PIs envision a WiFi access point on wheels that moves to optimize desired performance metrics. Movements need not necessarily be all around the floor of a home or office, neither do they have to operate on batteries, or connect wirelessly to the Internet. At homes, they could remain tethered to power and Ethernet outlets while moving in small areas (perhaps under the study table). In offices of the future, perhaps APs could move on tracks installed on top of false ceilings. The cloud could perhaps coordinate their mobility based on how users move within the environment, or how traffic and interferences change in the network. This project explores the viability of this vision and will present thorough measurements from various home/office environments.Initial measurements with moving WiFi APs in residential and enterprise environments exhibit promise. Analysis finds that complex multipath characteristics of indoor environments cause large fluctuations in link quality even when the antenna moves in the scale of one foot. Mobile APs can leverage this spatial variation by relocating to a pixel that is strong for its own clients and yet weak from its interferers. When multiple APs are coordinated by a central controller, the motion paths of the APs need to be jointly planned to optimize global network parameters. Heuristics will be developed that first assume the knowledge of AP locations; based on the outcome of this heuristic, the assumption might be relaxed. The project will also consider client mobility and how APs adapt to them, perhaps based on the quality of channel changes available from channel state information (CSI) in today's WiFi cards. If mobility proves to offer additional gains despite the advances in current technologies, robotic wireless networking might become an important and exciting direction of the future.