Algorithms for Location Estimation Based on （算法估计基于位置）.pdf

Algorithms for Location Estimation Based on RSSI Sampling Charalampos Papamanthou, Franco P. Preparata, and Roberto Tamassia Department of Computer Science and Center for Geometric Computing Brown University {cpap,franco,rt}@cs.brown.edu Abstract. In this paper, we re-examine the RSSI measurement model for location estimation and provide the first detailed formulation of the probability distribution of the position of a sensor node. We also show how to use this probabilistic model to efficiently compute a good esti- mation of the position of the sensor node by sampling multiple readings from the beacons (where we do not merely use the mean of the samples) and then minimizing a function with an acceptable computational effort. The results of the simulation of our method in TOSSIM indicate that the location of the sensor node can be computed in a small amount of time and that the quality of the solution is competitive with previous approaches. 1 Introduction Estimating the location of a roaming sensor is a fundamental task for most sensor networks applications. For example, if a sensor network has been deployed to provide protection against fire (in this case, sensor nodes report a sudden increase in temperature), we want to know the location of the sensor that triggers an alert so that action can be taken accordingly. Additionally, some routing protocols for sensor networks, such as geographical routing [15, 44], make routing decisions based on the knowledge of the locations of the sensor nodes. Common location estimation protocols that are widely adopted in practice assume that there are some fixed nodes (base station