Articles | Volume 17
19 Sep 2019
19 Sep 2019
A machine learning joint lidar and radar classification system in urban automotive scenarios
Rodrigo Pérez et al.
No articles found.
Miroslav Lach, Christian Looschen, and Erwin Biebl
Adv. Radio Sci., 19, 127–137,Short summary
A typical logistical RFID use-case is presented. The signal coverage is computed using accurate simulation techniques. A subsequent evaluation of the simulation results is performed. Emphasis is put on the derivation of expressive quantities to facilitate the evaluation and interpretation of the simulation results. Accurate simulations of the planned environment are essential to predict the real behavior and uncover flaws in the planned system, to avoid costly a-posteriori modifications.
Vera Kurz, Hannes Stuelzebach, Florian Pfeiffer, Carlo van Driesten, and Erwin Biebl
Adv. Radio Sci., 19, 165–172,Short summary
On the way to autonomous driving a safety proof is inevitable. Here simulators can help, but their accuracy is not good enough, yet. In order to improve such simulations, we analyzed the behaviour of road surfaces on radar sensors. Therefore, laboratory measurements and open space measurements were performed. Differences depending on material and roughness could be proved. Furthermore, material parameters were calculated, which can be mapped to roads in existing simulation setups.
Philipp Eschlwech and Erwin Biebl
Adv. Radio Sci., 17, 109–118,Short summary
A target simulator for RFID direction of arrival estimation systems is proposed. This simulator can be used in the evaluation of such systems and offers a method for fast, economical and, above all, reproducible and transferable analysis of their performance. Signal models and descriptions of two typical detrimental propagation effects are derived. The developed system structure and hardware modules are presented and exemplary evaluation results are given, showing the application of this method.
Bruna Cruz, Andreas Albrecht, Philipp Eschlwech, and Erwin Biebl
Adv. Radio Sci., 17, 119–127,Short summary
In this work, silver and gold nanoparticle inks were printed by a consumer inkjet printer on eco-friendly substrates like paper and PET in order to make the RFID tag manufacturing process less harmful to the ecosystem. A dipole antenna was designed and based on this design, simulations of the greener tags are presented and compared as a proof of concept for the different materials. First measurements are conducted and simulations with the optimized antenna designs are shown.
Christian Buchberger, Florian Pfeiffer, and Erwin Biebl
Adv. Radio Sci., 17, 197–203,Short summary
This article investigates the properties of dielectric corner reflectors for use in a number of millimeter wave applications, such as road safety for autonomous driving. Material characterizations of different typical plastics using transmission measurements are presented, as well as an analysis of their respective radar cross section (RCS) when used as corner reflectors. They exhibit similar behavior as conventional metallic reflectors, while intrinsic dielectric losses reduce the overall RCS.
Nils Hirsenkorn, Timo Hanke, Andreas Rauch, Bernhard Dehlink, Ralph Rasshofer, and Erwin Biebl
Adv. Radio Sci., 14, 31–37,Short summary
Increased complexity and severity of future driver assistance systems demand extensive testing and validation. For driver assistance functions the perception of the sensors is crucial. Therefore, sensors also have to be modeled. In this contribution, a statistical data-driven sensor-model, is described. The method is widely applicable and able to adapt to complex behavior. As exemplary implementation, a model of an automotive radar system, using a high precision measurement system, is presented.
Timo Patelczyk, Andreas Löffler, and Erwin Biebl
Adv. Radio Sci., 14, 63–69,
A. Ascher, M. Eberhardt, M. Lehner, and E. Biebl
Adv. Radio Sci., 14, 71–76,Short summary
In the present publication a GPS- based localization system for saving fawns during pasture mowing was introduced. Fawns were first found by a UAV before mowing began. They were then tagged with active RFID transponders, and an appropriate reader was installed on a mowing machine. Conventional direction-of-arrival approaches require a large antenna array, while our transponders were equipped with a small GPS module, allowing a transponder to determine its own position on request from the reader.
Michael Eberhardt, Philipp Eschlwech, and Erwin Biebl
Adv. Radio Sci., 14, 181–190,Short summary
This paper investigates the practicability of some well known antenna array calibration algorithms and gives a deeper insight into existing obstacles. Analysis on the validity of the common used calibration model is presented. A new effect in modeling errors is revealed and simulation results substantiate this theory. The research goal was to simplify calibration procedures for antenna arrays which are used for direction-of-arrival estimation.
A. Ascher, M. Lehner, M. Eberhardt, and E. Biebl
Adv. Radio Sci., 13, 81–86,
M. Eberhardt, M. Lehner, A. Ascher, M. Allwang, and E. M. Biebl
Adv. Radio Sci., 13, 87–94,Short summary
This paper presents a localization concept for fawn saving. Fawns can be saved during pasture mowing by marking them with an active UHF RFID transponder. The system consists of a mowing-machine mounted direction-of-arrival (DOA) system and a handheld DOA device. Only by estimating the DOA one can reliable localize the marked fawns. The whole RFID and DOA system is presented and communication concept is shown.
F. Pfeiffer, M. Rashwan, E. Biebl, and B. Napholz
Adv. Radio Sci., 13, 181–188,Short summary
In this paper, the coexistence of KLEER (a proprietary wireless standard in the 2.4GHz-ISM-band for high quality audio transmission) in presence of Bluetooth and WLAN IEEE 802.11b/g is analyzed. The study shows that Bluetooth and WLAN can be a very serious interferer, especially when no Bluetooth device is connected and Bluetooth will periodically actively search for devices (Bluetooth paging).
A. Koenig, T. Rehg, and R. Rasshofer
Adv. Radio Sci., 13, 197–202,
Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., Corrado, G. S., Davis, A., Dean, J., Devin, M., Ghemawat, S., Goodfellow, I., Harp, A., Irving, G., Isard, M., Jia, Y., Jozefowicz, R., Kaiser, L., Kudlur, M., Levenberg, J., Mané, D., Monga, R., Moore, S., Murray, D., Olah, C., Schuster, M., Shlens, J., Steiner, B., Sutskever, I., Talwar, K., Tucker, P., Vanhoucke, V., Vasudevan, V., Viégas, F., Vinyals, O., Warden, P., Wattenberg, M., Wicke, M., Yu, Y., and Zheng, X.: TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems, available at: https://www.tensorflow.org/ (last access: 13 August 2019), 2015. a
European Commission: Traffic Safety Basic Facts on Cyclists, available at: https://ec.europa.eu/transport/road_safety/sites/roadsafety/files/pdf/statistics/dacota/bfs2017_cyclists.pdf (last access: 23 November 2018), 2017a. a
European Commission: Traffic Safety Basic Facts on Pedestrians, available at: https://ec.europa.eu/transport/road_safety/sites/roadsafety/files/pdf/statistics/dacota/bfs2017_pedestrians.pdf (last access: 23 November 2018), 2017b. a
Pérez, R., Schubert, F., Rasshofer, R., and Biebl, E.: Single-Frame Vulnerable Road Users Classification with a 77 GHz FMCW Radar Sensor and a Convolutional Neural Network, 19th International Radar Symposium (IRS), 1–10, IEEE, Piscataway, New Jersey, USA, https://doi.org/10.23919/IRS.2018.8448126, 2018. a, b, c
Qi, C. R., Yi, L., Su, H., and Guibas, L. J.: PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space, CoRR, abs/1706.02413, http://arxiv.org/abs/1706.02413, 2017. a
This work presents a deep learning approach to classify road users as pedestrians, cyclists or cars using a lidar sensor for detection and a radar sensor for classification. A dataset was gathered on urban roads to train and test the deep learning algorithms. The results show that the system reliably classifies cars, but has trouble with pedestrians and cyclists. The results are improved after aggregating decisions with a Bayes filter. Overlapping targets remain a challenge for the system.
This work presents a deep learning approach to classify road users as pedestrians, cyclists or...