NnD: Shallow Neural Network Based Collision Decoding in IoT Communications
Internet of Things (IoT) has been widely used in intelligent warehouse, environment monitoring and smart buildings. In these application scenarios, concurrent transmissions frequently occur in which multiple transmitters send packets to one receiver simultaneously, causing severe collisions and low throughput. The state-of-the-art methods are able to decompose collided packets from different transmitters. However, they rely heavily on random time offsets and has poor performance under inferior channel conditions. In this paper, we present a new physical layer mechanism Open image in new window (nnD) to resolve multi-packet collisions. We first collect collision-free symbols or history single packets as the training set. In order to improve the decoding accuracy, we model the mapping relationship between overlapped symbols and their symbol values by neural networks. Since overlapping combinations of symbols are limited which are decided by corresponding chips’ value, we can predict values of unknown symbols by classifying different kinds of overlapping combinations. By introducing neural networks, nnD can not only achieve a high decoding precision but also can dynamically choose neural network architecture to adapt to different collision scenarios. To evaluate the performance of nnD, extensive trace-driven simulations are conducted. The results demonstrate that nnD outperforms existing methods in terms of bit error rate and the number of concurrent transmissions.
This work was supported in part by the National Key R&D Program of China 2018YFB1004703, NSFC grant 61972253, 61672349, 61672353.
- 1.Alliance, Z.: Introduction to Zigbee (2018). http://www.zigbee.org
- 2.Gollakota, S., Katabi, D.: Zigzag decoding: combating hidden terminals in wireless networks. In: ACM SIGCOMM (2008)Google Scholar
- 3.Halperin, D., Hu, W., Sheth, A., Wetherall, D.: Predictable 802.11 packet delivery from wireless channel measurements. In: ACM SIGCOMM (2010)Google Scholar
- 8.Kong, L., Liu, X.: mZig: enabling multi-packet reception in ZigBee. In: ACM MOBICOM (2015)Google Scholar
- 11.Ronen, E., Shamir, A., Weingarten, A.O., Flynn, C.O.: IoT goes nuclear: creating a ZigBee chain reaction. In: IEEE S&P (2017)Google Scholar
- 12.Sobrinho, J.L., de Haan, R., Brazio, J.M.: Why RTS-CTS is not your ideal wireless LAN multiple access protocol. In: IEEE WCNC (2005)Google Scholar
- 14.Ziouva, E., Antonakopoulos, T.: CSMA/CA performance under high traffic conditions: throughput and delay analysis. Computer Communications (2002)Google Scholar