Prediktion av radiokanaler för routing och trådlös reglering
Tidsperiod: 2014-01-01 till 2016-12-31
Projektledare: Anders Ahlén
Budget: 2 850 000 SEK
Sensor data for monitoring and control in the process industry are today communicated by means of cables. By instead using wireless sensors and actuators many advantages can be obtained, e.g., easy and fast deployment, low cost, and flexibility with respect to both localization of sensor nodes and reorganization of the processes.In the project we aim at developing efficient channel gain prediction algorithms for the purpose of improving routing of sensor data to the gateway. These algorithms will assist in the optimization of latency, energy expenditure and reliability, depending on what is of most interest. To accomplish this we will use a network state formulation representing strong reflecting or shadowing objects together with fading distributions representing all links between sensors and between sensors and gateway.Both off-line and on-line scenarios will be considered. The on-line scenario is particularly challenging as the available channel gain data among the sensors may be sparse. We will use the forward algorithm, expectation maximization, and maximum likelihood type methods in addition to dynamic programming to solve the optimization problems.The results of the project will contribute new knowledge about long term indoor radio channel behavior, how this behavior can be modeled, predicted, and exploited for efficient routing. It will also underpin the deploying process for wireless monitoring and control systems as well as development of new protocols.