by Alexis Duque, R&D & Security Leader
From November 25th to 29th, I attended the ACM MSWiM conference, that took place in Miami Beach, Florida, USA, to present our full paper on “Performance Evaluation of LED-to-Camera Communication”, co-authored with Razvan Stanica, Hervé Rivano, and Adrien Desportes.
During the conference, I introduced our work on the modeling and simulation of LED-to-Camera communications by presenting CamComSim, the first simulator in this area. This work has been included in the conference proceedings and the presentation slides are available on my speakerdeck.
This also allowed me to attend interesting paper presentations, during the main conference, as well as the collocated workshops, PE-WASUN & DIVANET, and symposia: Q2SWinet & MobiWac.
Among them, the 6 following presentations particularly interest me. Note that this selection is unfair and highly influenced by my fields of interest
HiPR: High-Precision UWB Ranging for Sensor Networks
Daniel Neuhold (Alpen Adria Universität, Austria); Christian Bettstetter (University of Klagenfurt, Austria); Andreas Molisch (University of Southern California, USA)
The author presented his work on improving the state of the art precision of UWB localization using the Decawave module. His approach leverage time-of-arrival measurements with the double-sided two-way ranging (DS-TWR). His technique is about 30 times faster than Decawave's out-of-the-box solution and improves the precision by one order of magnitude.
ns-3 meets OpenAI Gym: The Playground for Machine Learning in Networking Research
Piotr Gawłowicz and Anatolij Zubow (Technische Universität Berlin, Germany)
The speaker introduced ns3-gym: an ns-3 module that relies on OpenAI Gym to add Reinforcement Learning capabilities to the widely used network simulator ns-3.
Fingerprinting using Fine Timing Measurement
Nicolas Montavont (Institut Mines Telecom / IMT Atlantique, France); Jerome Henry (Cisco, USA)
J.Henry started its presentation by introducing the time-of-flight based ranging mechanism defined in 802.11-2016, called FTM. He then showed that simple FTM frame observations by a passive sniffer can easily allow for individual chipset identification. He leverages machine learning with LSTM to recognize individual machines performing FTM exchanges, even when these machines implement the same chipset or the same hardware platform. He concluded by proposing several ways to mitigate individual device patterning based on FTM exchange observation.
Simulation and Performance Evaluation of the Intel Rate Adaption Algorithm
Rémy Grünblatt (Université de Lyon, France); Isabelle Guérin Lassous (Université Claude Bernard Lyon 1 - LIP, France); Olivier Simonin (INSA Lyon, France)
Remy described the Wifi rate adaptation algorithm implemented in the Intel IwlWifi Linux Driver, called Iwl-Mvm-Rs. He has reverse-engineered it relying on the Iwl-Mvm-Rs open-source implementation. He then summed-up the different mechanisms used to obtain the best throughput according to the network conditions and integrated the Iwl-Mvm-Rs algorithm in the ns-3 simulator. Thanks to this implementation, he presented the performance of Iwl-Mvm-Rs and compare them with the ones of Minstrel-HT and IdealWifi, also implemented in the ns-3 simulator.
Evaluation of LoRaWAN Transmission Range for Wireless Sensor Networks in Riparian Forests
Fabian Astudillo-Salinas, Andres Vazquez-Rodas, Pablo E Avila and Alcides Araujo (University of Cuenca, Ecuador)
The author proposed an empirical study of LoRa with LoRaWAN transmission range in riparian forests, based on path-loss modeling, using both received signal strength indicator (RSSI) and signal-to-noise-ratio (SNR). The measurement conducted in the riparian forest of three local rivers at urban, semi-urban, and rural environments gives the following insight: there is a significant distribution difference among measurement places, a high correlation between two banks of the same river, a higher standard deviation in urban measurements and a larger coverage in rural areas.
Benchmarking the Physical Layer of Wireless Cards using Software-Defined Radios
Liangxiao Xin, Johannes K Becker, Stefan Gvozdenovic and David Starobinski (Boston University, USA)
Johannes introduced a new testbed architecture for software-defined radio (SDR) wireless device performance benchmarking. The testbed is capable of accessing and measuring physical layer protocol features of real wireless and IoT devices. The testbed allows tight control of timing events, at a microsecond time granularity. Using the testbed, he showed marked differences in the performance of Wifi devices, including a variation of the receiver sensitivity.
You want to go deeper with your IoT project? Discover how to build a successful connected solution with our free guide.