dc.contributor.authorSubramanian Rathi Push
dc.description.abstractIn our day to day life, the need for indoor positioning systems has increased drastically. The reason behind is its wide range of applications starting from industries to healthcare. The Global Positioning Systems (GPS) is suitable only for outdoor space because of its significant power loss within the indoor environment, thereby affecting the area of coverage for indoor space. Recent research is being carried out for the integration of GPS with indoor positioning systems to increase accuracy and robustness. This dissertation focuses on SIMTech developed indoor positioning system namely the Real Time Location Systems (RTLS). The work aims at a deep study of the localization algorithm implemented and to enhance the overall performance, i.e., accuracy of the system. This dissertation approaches the localization algorithm based on RSSI Fingerprinting technique. The other concepts involved are K -Nearest Neighbour and Clustering filter. Each of them has its own contribution towards reliability of the system. After a study about the various ways to improve the accuracy, the implementation and analysis of Kalman filter were experimented. Here it produces more appropriate RSSI value after refinement. Other alternatives include the use of neural networks, strong KNN function, LQI parameter for estimation of position and subclustering. Importance was given to the calibration of database as it plays a vital role in the stability ofthe system. The already installed experimental setup at Infinitus lab is used for testing purposes. The calibration was done for both stable and unstable environment. In the former, the orientation, direction and height of asset tag from ground level was kept constant whereas in the latter it was not constant. Three different test cases were performed to study the behaviour of the system under different environment. The results of the test for real time walking condition in a dynamic environment proved better than the current localization algorithm used in RTLS. The positioning error for few random points in the Infinitus lab is about 0 - 2.81m and the average positioning error is about 1.67m. This implicitly describes the reliability of the proposed algorithm.en_US
dc.format.extent68 p.en_US
dc.subjectDRNTU::Engineering::Electrical and electronic engineeringen_US
dc.titlePerformance enhancement of indoor real time location systemsen_US
dc.contributor.supervisorSoong Boon Heeen_US
dc.contributor.schoolSchool of Electrical and Electronic Engineeringen_US

Files in this item


This item appears in the following Collection(s)

Show simple item record