Maximilian Gaul

Maximilian Gaul

Master's Thesis

Comparison of different radio modules for CSI sensing and positioning


Thomas Altstidl (M. Sc.), Maximilian Stahlke, Prof. Dr. Björn Eskofier


06 / 2023 – 12 / 2023


Passive indoor localization using commercial off-the-shelf WiFi hardware [5] gained considerable interest over the recent years due to its wide range of possible use-cases such as an integration into smart-home, office- or show-room applications. Radio systems in general collect spatial information about reflection and diffraction that occur within a given channel. These so called channel state information (CSI) [4] were leveraged for people counting [8], indoor localization [10], tracking [11] and human activity recognition [3]. Recent advances in WiFi technology, namely WiFi 6(E), come with an increased number of usable subcarriers as well as an increased bandwidth per channel compared to the older WiFi 5 standard [1]. A higher channel bandwidth could result in more accurate sensing because of the higher time resolution of the measured CSI [9]. Ultra-broadband (UWB) modules leverage that property to achieve even higher temporal resolution of the radio signals [7] but they require special hard- and software and are not ubiquitously installed. Existing literature mostly employs hardware with restricted bandwidth of below 40 MHz [12, 2, 6, 8] with only one receiver and transmitter combination, which limits the ability for sensing tasks. Therefore, we want to compare the performance of recent radio standards such as WiFi 6(E) and UWB and evaluate the impact of varying channel bandwidth on people detection and counting as well as positioning by exploiting CSI measurements. To achieve this, we create scenarios for each application in an indoor office space where we setup multiple radio transceivers and perform offline data collection and training. Eventually, we define key performance indicators and evaluate the performance of different machine learning based algorithms on defined sensing and localization tasks with different levels of complexity.


[1] Marco Cominelli, Francesco Gringoli, and Francesco Restuccia. Exposing the csi: A systematic investigation of csi-based wi-fi sensing capabilities and limitations, 2023.
[2] Shuya Ding, Zhe Chen, Tianyue Zheng, and Jun Luo. RF-net. In Proceedings of the 18th Conference on Embedded Networked Sensor Systems. ACM, nov 2020.
[3] Bing Li, Wei Cui, Wei Wang, Le Zhang, Zhenghua Chen, and Min Wu. Two-stream convolution augmented transformer for human activity recognition. Proceedings of the AAAI Conference on Artificial Intelligence, 35(1):286–293, May 2021.
[4] Yongsen Ma, Gang Zhou, and Shuangquan Wang. Wifi sensing with channel state information: A survey. ACM Comput. Surv., 52(3), jun 2019. [cs], Nov. 2022. [Online]. Available: 10307 (visited on 12/19/2022).
[5] Francesca Meneghello, Domenico Garlisi, Nicolo Dal Fabbro, Ilenia Tinnirello, and Michele Rossi. Sharp: Environment and person independent activity recognition with commodity ieee 802.11 access points. IEEE Transactions on Mobile Computing, pages 1–16, 2022.
[6] Muhammad Muaaz, Ali Chelli, Martin Wulf Gerdes, and Matthias Pätzold. Wi-sense: a passive human activity recognition system using wi-fi and convolutional neural network and its integration in health information systems. In Annals of Telecommunications, pages 163–175, 2022.
[7] Tommaso Polonelli, Simon Schläpfer, and Michele Magno. Performance comparison between decawave dw1000 and dw3000 in low-power double side ranging applications. In 2022 IEEE Sensors Applications Symposium (SAS), pages 1–6, 2022.
[8] Aryan Sharma, Wenqi Jiang, Deepak Mishra, Sanjay Jha, and Aruna Seneviratne. Optimised cnn for human counting using spectrograms of probabilistic wifi csi. In GLOBECOM 2022 – 2022 IEEE Global Communications Conference, pages 01–06, 2022.
[9] David Tse and Pramod Viswanath. Fundamentals of Wireless Communication. Cambridge University Press, Cambridge, 2005
[10] Xuyu Wang, Xiangyu Wang, and Shiwen Mao. Cifi: Deep convolutional neural networks for indoor localization with 5 ghz wi-fi. In 2017 IEEE International Conference on Communications (ICC), pages 1–6, 2017.
[11] Zhongqin Wang, Jian Andrew Zhang, Min Xu, and Jay Guo. Single-target real-time passive wifi tracking. IEEE Transactions on Mobile Computing, pages 1–1, 2022.
[12] Yi Zhang, Yue Zheng, Kun Qian, Guidong Zhang, Yunhao Liu, Chenshu Wu, and Zheng Yang. Widar3.0: Zero-effort cross-domain gesture recognition with wi-fi. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(11):8671–8688, 2022.