Simulating Wireless Links Within a Dynamic EnvironmentJ. Alfredo Cruz-CarlónUniversidad Nacional Autónoma de MéxicoAv. Universidad 3000México D.F., Méxicojucruz@ciencias.unam.mxABSTRACTA couple of fundamental question for wireless networks si-mulators are: does the receiver hears a transmitter? andif so, how well? To answer them, the radio waves propa-gation characteristics of the environment must be consid-ered. This observation has motivated the creation of widelydif f erent models, they spread from only calculating the dis-tance between the nodes to fully describe a real deploymentarea with precise measurements of how a material interactswith a radio wave. As models become increasingly detailedin their description of the environment, the time requiredto describe and simulate it also increases. In this paper wepresent a model to simulate wireless links over environmentsthat may have many dif f erent propagation characteristics,each of them suited to be simulated by its own model, andpossibly changing over time. Our model is meant to be usedas a stepping stone between simulations using one propaga-tion model and simulations over a detailed description of thedeployment area.KeywordsPerformance evaluation and modeling, RF channel modelingand analysis,Tools, prototypes and testbeds,Wireless Com-munication and Mobile Networking1. INTRODUCTIONOne major challenge in wireless networks is to decidewhenever two given nodes can communicate with each other.On the theoretical side, the well-known unit-disk graph andquasi-unit disk graph models are among the most used ifthe position of the nodes is relevant. Otherwise, an edge-weighted directed graph or linear programming model is of-ten used.On the simulation side three options are mainly employed.The f i rst is to use a single propagation model, usually thelog-normal shadowing. The second is to describe the RFpropagations characteristics of the deployment area and useray tracing (or similar techniques). The third option is toPermission to make digital or hard copies of all or part of this work for personal orclassroom use is granted without fee provided that copies are not made or distributedfor prof i t or commercial advantage and that copies bear this notice and the full cita-tion on the f i rst page. Copyrights for components of this work owned by others thanACM must be honored. Abstracting with credit is permitted. To copy otherwise, or re-publish, to post on servers or to redistribute to lists, requires prior specif i c permissionand/or a fee. Request permissions from permissions@acm.org.DIVANet’16, November 13-17, 2016, Malta, Maltac ? 2016 ACM. ISBN 978-1-4503-4506-4/16/11...$15.00DOI: http://dx.doi.org/10.1145/2989275.2989291use an edge-weighted connection (directed) graph in whichusually if two nodes are adjacent, then the quality of thelink between them is expressed by the weight of their edge.Each approach has its advantages and disadvantages. Amongthe advantages of the log-normal shadowing model is thatthe environment can be described with two variables and todecide if two nodes can communicate with each other only asingle equation is used; its disadvantage is the accuracy of itspredictions. It seems it varies depending on environmentalfactors that can not be integrated in the model, like dif-ferent RF propagations characteristics. On the other hand,the second approach, a full description of RF propagationscharacteristics of the deployment area seem to give the bestpredictions but the description of the environment is a timeconsuming undertaking. Finally, the use of edge-weightedconnection graphs is not restricted to one model of the de-ployment area nor it requires a full description of it either,however how the adjacency function of the connection graphis created (and sometimes updated too) is not trivial.In this paper we present a simulation model for wirelesslinks that like the edge-weighted connection graph model, itis not restricted to a single propagation model and it doesnot require a full description of the propagation environ-ment, but unlike it the adjacency function of the connectiongraph is well def i ned as in log-normal shadowing and itsalso geographically consistent like when a full descriptionof the RF propagations characteristics is used. Moreoverout model is able to make the previously discussed modelsto work together in the same simulation. We believe thatin the simulations realm our model sits between simulationsperformed with one propagation model and simulations overa detailed description of the deployment area. In Section 2we brief l y expose some related work. In Section 3 we presentour model. In Section 4 we present a brief discussion of ourproposal. Finally, in Section 5 we present a few f i nal re-marks.2. RELATED WORKUnit Disk Graphs or UDGs and similar models were amongthe f i rst frameworks to be used to describe wireless con-nectivity. Their use included both, the theoretical and thevalidation/simulation side. It is widely believed that thesemodels are helpful when we are interested in the best casescenario, that is, the wireless signals are not disrupted by theenvironment or by each other and the antenna’s radiationpattern is a perfect disk; as a consequence, we can answerthe question whether a node can communicate with anotherby looking at the (euclidian) distance between them. In the115
