GPS World, September 2010

assuring time synchronization between sensors and nodes and limiting computational burden for real time applications Distributed GPS Apertures In the case of GPS signal degradation due to increased path loss and radio frequency interference RFI one can use an antenna array at the receiver site to increase the gain in the satellite signal direction as well as steer spatial nulls in the interfering signal directions For a network of GPS users one may be able to combine the signals received at various receivers nodes to achieve these goals beam pointing and null steering see FIGURE 2 However a network of GPS users represents a distributed antenna aperture with large hundreds of wavelengths interelement spacing This large thinned antenna aperture has some advantage and many drawbacks The main advantage is increased spatial resolution which allows one to discriminate between signals sources with small angular separations The main drawback is very high sidelobes in fact grating lobes which manifest as grating nulls sympathetic nulls in null steering The increased interelement spacing will also lead to the loss of correlation between the signals received at various nodes Thus space only processing will not be sufficient to increase SNR by combining the satellite signals received at various nodes One has to account for the large delay between the signals received at various nodes Similarly for adaptive null steering one has to use space time adaptive processing STAP for proper operation These research challenges must be solved for distributed GPS aperture to become a reality Investigate the increase in SNR that can be obtained by employing distributed GPS apertures accounting for inaccuracies in the inter nodal ranging measurements Investigate the improvement in the signal to interferenceplus noise ratio SINR that can be obtained over the upper hemisphere when a distributed GPS aperture is used for adaptive null steering to suppress RFI in GPS receivers Obtain an upper bound for inter node distances Based on the results of the above two investigations develop approaches for combined beam pointing and null steering using distributed GPS apertures Inter nodal Ranging Techniques In a wireless sensor network an RF signal can be used to measure ranges between the nodes in various modes For example WLAN observes the RF signal strength and UWB measures the time of arrival time difference of arrival or the angle of arrival There are known challenges for example signal fading interference or multipath to address for a RF based technique to reliably serve as internodal ranging method Ranging Based on Optical Sensing Inter nodal range measurements can be also acquired by active and passive imaging sensors such as laser and optical imaging sensors Laser range ÅFIGURE 3 Centralized extended Kalman filter finders that operate in the eye safe spectrum range can provide direct range measurements but the identification of the object is difficult Thus laser scanners are preferred delivering 3D data at the sensor level Using passive imagery such as digital cameras provides a 2D observation of the object space more information is needed to recover 3D information the most typical techniques is the use of stereo pairs or more generally multiple image coverage The laser has advantages over optical imagery as it preserves the 3D object shapes though laser data is more subject to artifacts due to non instantaneous image formation In general regardless whether 2D or 3D imagery is used the challenge is to recognize the landmark under various conditions such as occlusions and rotation of the objects when the appearance of the landmark alternates and the reference point on the landmark needs to be accurately identified to compute the range to the reference point with sufficient accuracy Network Configuration Nodes in the ad hoc network must be localized and ordered considering conditions such as type of sensors on the node grade of the IMU anti jamming capability positional accuracy accuracy of inter nodal ranging technique geometric confi guration computational cost requirements and so on There are two primary types of network confi gurations used in collaborative navigation centralized and distributed Centralized configuration is based on the concept of server master and client nodes Distributed configuration refers to the case where nodes in the network can be configured without a master node that is each node can be considered equal with respect to other nodes Sensor Integration The selection of data integration method is an important task it should focus on arriving at an optimal solution not only in terms of the accuracy but also taking the computational burden into account The two primary options are centralized and decentralized extended Kalman fi lter EKF GOVERNMENT Emergency Response GPS World September 2010 www gpsworld com 38