Research On Channel Diversity Of Radio Tomographic Imaging In Multipath Environment

Radio Tomographic Imaging (RTI) is an emerging technology that uses thechange of received signal strength (RSS) caused by obstruction of the object to locatethe object within the area covered by wireless sensor network. Shadowing-based RTI(SRTI) divides the area into gird voxels uniformly, and solves the linear modelrelating the attenuation of RSS to voxels of attenuation field, which is established onweighting model between voxels and links, to reconstruct image and realizeslocalization and tracking of object.In an open environment in which the line-of-sight path is dominant, theperformance of localization is very good because the shadowing fading induced byobstruction of the object is obvious. However, in an indoor environment where themultipath is abundant, the change of RSS due to link obstructed by the objectbecomes unpredictable. The RSS can decrease, increase or even remain unchanged,which results in the degradation of localization performance of SRTI. In the past RTIis only to locate the object in a single channel, this thesis presents the use of channeldiversity that utilizes superposition characteristics of random phases of multipathcomponents in different channels, to reduce multipath interference, enhanceavailability of links and improve the localization performance of SRTI in multipathenvironments.When using all the measurement data of the whole channels, it is difficult toensure the real-time performance of imaging as the object moves rapidly. The qualityof signal will be affected when there exists other networks in the area, such as WLAN.The channel with the higher packet reception rate has the better signal quality.Changes of RSS under different channels about the same link are diverse. RSS of thelink which experiences the anti-fade usually decreases when the link is obstructed bythe object. Therefore, this thesis employs the two methods-packet reception rate andfade level, which are based on the statistics of links under different channels, to selectthe channel.In order to guarantee the real-time of data measurement, two kinds of datameasurement protocols are designed in this thesis, which are token passing protocol and TDMA protocol. The token passing protocol is suitable for the data measurementunder the single channel because of simple structure and easy to implement. Slotsynchronization and nodes’ autonomy enables TDMA protocol to settle thecommunication interrupt resulting from the failure of channel switching well. Theself-healing ability of TDMA protocol is very strong, which is applied to themulti-channel of data measurement.The result of experiment demonstrates that in multipath environment, the imagequality of SRTI using the multi-channel is much better than using the single channel.And using the attenuation statistics of RSS under the whole channels to image has thebest quality, which can improve the localization performance by70%compared withthe single channel.