Maximum Likelihood Topology Maps for Wireless Sensor Networks Using an Automated Robot Ashanie Gunathillake School of Electrical Engineering and Telecommunication, University of New South Wales, Sydney, Australia NICTA, Eveleigh, NSW 2015, Australia ashanie@student.unsw.edu.au Andrey V. Savkin School of Electrical Engineering and Telecommunication, University of New South Wales, Sydney, Australia a.savkin@unsw.edu.au Anura P. Jayasumana Department of Electrical and Computer Engineering, Colorado State University, Fort Collins, CO 80523 USA anura.jayasumana@colostate.edu Abstract— Topology maps represent the layout arrangement of nodes while maintaining the connectivity. As it is extracted using connectivity information only, it does not accurately represent the physical layout such as physical voids, shape, and relative distances among physical positions of sensor nodes. A novel concept Maximum Likelihood-Topology Maps for Wireless Sensor Networks is presented. As it is based on a packet reception probability function, which is sensitive to the distance, it represents the physical layout more accurately. In this paper, we use a binary matrix recorded by a mobile robot representing the reception of packets from sensor nodes by the mobile robot at different loca- tions along the robots trajectory. Maximum likelihood topology coordinates are then extracted from the binary matrix by using a packet receiving probability function. Also, the robot trajectory is automated to avoid the obstacles and cover the entire network within least possible amount of time. The result shows that our algorithm generates topology maps for various network shapes under different environmental conditions accurately, and that it outperforms the existing algorithms by representing the physical layout of the network more accurately Keywords—Localization, Packet Receiving Probability, Signal Propagation, Topological Map, Wireless Sensor Network I. I NTRODUCTION Many Wireless Sensor Network (WSN) protocols require the location coordinates or map of the sensor nodes, as the data collected by the sensors are useful when considered in the context of the location from which it was collected. The location information can be obtained by the incorporation of hardware components such as Global Position Systems (GPS). However, cost prevents the use of expensive hardware devices in large- scale networks and also they do not work in all environments. Therefore, one of the major challenges in WSN is to de- termine the physical coordinates of the sensor nodes while minimizing the hardware cost. To address this issue, numerous localization algorithms have been proposed in the literature. These can be divided into two categories, namely range-based and range-free localization algorithms. Range-based algorithms use special hardware components to measure range-based pa- rameters such as received signal strength (RSS) [1], time of arrival (TOA) [2], angle of arrival (AOA) [3]. These algorithms are affected by noise, fading of the signals and interference [4] and as a result, their accuracy decreases in environments with obstacles. Range-free algorithms rely on the information about the connectivity of the sensors and known location of some of the senors (anchor nodes) instead of special hardware. However, the accuracy of these algorithms is highly depent on the number of anchor nodes and their distribution [5] [6]. A topology preserving map is an attractive alternative to the physical map of the network. Topology map is a representative of arrangement of node that preserves the connectivity. Topol- ogy map are obtained by Singular Value Decomposition (SVD) of Virtual Coordinate System (VCS) [7] , in which sensor node is identified by a vector that contains the distance in hops to a set of anchor nodes. Even though topology maps are based only on the connectivity, they have been demonstrated to preserve the general shapes of voids and boundaries. They are non-linearly distorted versions of physical maps. The objective of this research is to come up with a topology preserving map that is closer to the actual physical map, but without requiring the expense associated with localization based on actual physical distance measurements. n the process, we generalize the concept of topology map, by using measurements other than direct hop distances to obtain the map. To this end, we propose a Maximum Likelihood Topology Map (ML-TM) of a sensor network. We consider the problem of generating a map of the network by using a mobile robot. To avoid the disadvantages described above of geographic localization, It does not attempt to measure the actual physical distances. Unlike the case with topology preserving maps in [7], it does not rely solely on the connectivity. Here we use a binary matrix to calculate the maximum likelihood topology coordinates that reduce the dependency of the output on range-based parameters. The robot moves in the area where the network is deployed, and gathers a binary matrix based on the packets received from the nodes from different locations. Then, the topology coordinates are calculated by the binary matrix and a packet receiving probability function, which is sensitive to the distance. Received Signal Strength Indicator (RSSI) based algorithms extract the distances from received power, which encounters significant errors due to RF communication effects. The proposed scheme does not require such error prone and unreliable distance measurements. On the other hand, range-free algorithms use a 2016 IEEE 41st Conference on Local Computer Networks © 2016, Ashanie Gunathillake. Under license to IEEE. 339 2016 IEEE 41st Conference on Local Computer Networks © 2016, Ashanie Gunathillake. Under license to IEEE. DOI 10.1109/LCN.2016.62 339 2016 IEEE 41st Conference on Local Computer Networks © 2016, Ashanie Gunathillake. Under license to IEEE. DOI 10.1109/LCN.2016.62 339