Localization In Noisy Environment Using Extended Kalman Filter

Abstract

Localization is an important aspect of Wireless Sensor Networks. Information regarding the position of the sensor nodes is not always known. Without the position information of the sensors, the data reported by the sensors is of little use. Various approaches have been used to perform localization using some information about the sensor node. Potential field approach for localization, using distance information has been successfully tested with satisfactory results.
\nHowever in case of noisy environment, the range measurements have greater inaccuracy. In such cases, localization using the above algorithm can provide some inaccuracy. To rectify such erroneous localization situations, Extended Kalman Filters are used to estimate the position. The Extended Kalman Filter has been used as the process for estimation of coordinates is a non-linear process. The EKF is a recursive filter which only needs the information from the previous state to predict the next state. My Contributions to the thesis :
\n1. Programmed the Cricket in TinyOS to store the distance measurement in
\narrays and then broadcast the same over radio to other crickets. The cricket
\nprogrammed as a base station only listens on the radio and then send the
\nreceived message to the PC over the serial UART.
\n2. Created a LabVIEW application which processes the message received from
\nthe cricket base station and deciphers the message to extract the distance
\nmeasurements. The node id is used to identify the transmitting node and the
\ndistances are stored in corresponding arrays. The ranging information in
\nthen written into a file.
\n3. Created a LabVIEW application to read the distances stored in the file and
\nthen arrange the readings depending on the number of nodes.
\n4. Implemented the Extended Kalman Filter Localization algorithm for relative
\nand absolute localization algorithm.
\n5. Implemented V-shaped swarm behavior demonstration using three Garcia as
\nthe robot and cricket as the motion guiding tool.