Using Radio Frequency (RF) signals as a way of sensing has seen great popularity and promising results in existing research. However, the technology is limited by the fact that an object has to be located between a transmitter and a receiver in order to be observable by such a system. While many publications have addressed this limitation by leveraging multipath propagation effects, these are often only found in closed indoor environments. In this paper, we introduce a novel method to utilize RF sensing from a distance that does not rely on multipath effects, allowing it to be used in scenarios where it would have previously been considered impractical such as outdoor applications. By receiving the ground reflection of a signal, we measure objects within its path as the signal propagates through them. This allows combining both antennas into a single compact unit that can be pointed at an object from a distance. We perform a series of experimental evaluations to gather preliminary data on the feasibility and performance of this approach. Using our setup, we are able to count water bottles positioned underneath the sensor with high accuracy based solely on the received signal strength of the reflected component.

Measuring soil moisture for agriculture is currently a hot research topic. Wireless sensor networks, that transmit readings from moisture sensors are a main contender for this task. However, sensors used in these networks often suffer from corrosion. Here, we propose a more scalable solution, using Commercial Off-The-Shelf (COTS) Bluetooth Low Energy (BLE) beacons with no additional sensor. Because of the variety of their application areas, they are available at smaller unit costs. In comparison to many moisture sensors, they can be fully encapsulated in a case for protection against corrosion. Similar solutions have been proposed using Received Signal Strength Indicator (RSSI) readings of the sensor nodes. We propose to instead use the Angle-of-Arrival (AoA), model the influence of different error sources, and calculate the expected measurement accuracy.

Indoor localization based on wireless communication networks has attracted a lot of attention in research. It can deliver better accuracy than a traditional Global Navigation Satellite System (GNSS) and operate in environments where satellite signals can not be received. Although more accurate methods have been published, fingerprinting based on Received Signal Strength (RSS) remains interesting. Such a system can be deployed on top of most existing networks without imposing additional requirements or restrictions to the communication. However, existing solutions usually require a larger number of base stations with known locations (anchors) within a mobile node’s reception range. This is required to prevent ambiguous fingerprints and deliver accurate results. However, it also increases the network’s energy consumption and operating cost. In this paper, we present a new approach that uses multi-frequency fingerprints to eliminate this requirement and operates with only a single anchor node while achieving accuracy that is comparable to existing solutions. In order to efficiently collect RSSI fingerprints across multiple frequencies, we introduce a measurement methodology using Software Defined Radio (SDR).

Wireless transmission of images in real time over large distances is a challenge for applications such as agricultural UAV-based remote sensing. The Wifibroadcast project uses standard WiFi hardware to provide a purely unidirectional transmission link for such situations. This paper examines the trade-off between image compression and FEC for image transmission via Wifibroadcast. The goal is to achieve highest robustness against transmission errors with regards to visual quality. For this, we conducted two experiments: First, we implemented a complete simulation of the transmission process to assess different degrees of packet and byte corruption. Afterward, we incorporated real network traces into the simulation in order to model realistic corruption patterns. The results show that for uniformly distributed byte corruption patterns, compression in combination with redundancy does not yield benefits to visual quality. However, if packet loss is the dominating corruption source, FEC is able to make image transmission more robust.

Some core advances in the early research on wireless networks have been made with simple analytical models for collisions. While complex simulations are nowadays often used to evaluate such networks, the more simple analytical models are still helpful for gaining fundamental understanding of effects. One challenge in using such models is finding appropriate parameters for them. To facilitate quick plausibility checks on these models, ready-to-use parameter sets for realistic example networks are desirable. In this paper, we provide such parameter sets for two networks, one in an office building and one in an industrial facility.

Using existing general purpose wireless technologies, it is difficult to transmit large amounts of data from Unmanned Aerial Vehicles (UAVs) to a ground station during flight. Thus, there is a demand for specialized solutions. Wifibroadcast proposes a solution for this use case, utilizing a unidirectional and connectionless design to transmit data using unmodified low-cost WiFi hardware. In this paper, we measure and evaluate the performance of Wifibroadcast for UAV to ground transmissions over large distances and provide a direct comparison against standard IEEE 802.11. We demonstrate that Wifibroadcast is able to outperform WiFi in all tested scenarios while identifying specific characteristics, limitations, and requirements of a long-distance transmission on both the 5 GHz and 2.4 GHz frequency bands.

Due to the COVID-19 pandemic, smartphone-based proximity tracing systems became of utmost interest. Many of these systems use Bluetooth Low Energy (BLE) signal strength data to estimate the distance between two persons. The quality of this method depends on many factors and, therefore, does hardly deliver accurate results. We present a multi-channel approach to improve proximity classification, and a novel, publicly available data set that contains matched IEEE 802.11 (2.4 & 5 GHz) and BLE signal strength data, measured in four different environments. We utilize these data to train machine learning models. The evaluation showed significant improvements in the distance classification and consequently also the contact tracing accuracy. However, we also encountered privacy problems and limitations due to the consistency and interval at which such probes are sent. We discuss these limitations and sketch how our approach could be improved to make it suitable for real-world deployment.

This work in progress paper presents an automated approach for network coverage prediction in real-world environments by combining mobile mapping, 3D mesh generation, and a ray launching based network simulator. We identify the challenges and demonstrate the functionality of such a pipeline. We preview an empirical evaluation in a realistic real-world environment.

The quality of Bluetooth Low Energy (BLE) connections can be improved by deactivating wireless channels when external interference is detected. The BLE standard contains basic requirements to implement this behavior. However, it does not specify how channels should be classified and disabled, leaving the details up to the implementations. In this paper, an implementation is presented and evaluated, based partially on existing work. A new algorithm for re-activating channels that were previously disabled in order to allow for operation in dynamic environments with changing background interference is proposed and demonstrated to deliver improvements over existing work, resulting in better connection stability. Finally, an evaluation under real-world conditions confirms the improvements to BLE connections that were previously only shown in controlled testing environments.

Wireless networks are indispensable for current and future elements of industry 4.0 and the digital transformation of companies. Industrial environments with arrays of machines mostly made of metal differ significantly from home and office environments, with consequential effects on wireless communication. We deployed wireless sensor networks built with commercial-off-the-shelf hardware in industrial and office scenarios and evaluated various aspects of wireless communications with different technologies and parameters. In this paper, we present our measurement architecture and the core findings. The results show highly spatial- and time-dependent differences between the two scenarios and also provide a comparative evaluation of the applied technologies. In addition, we also share parameters for established simulation models making future simulative evaluations particularly easy.

Increasing the efficiency of farming has been one of, if not the, most important factor in human development since the neolithic age. It laid the path for mankind's accomplishments in the present age. Even today, new improvements are required. One of the most promising improvements is precision farming which demands a spatial distribution of crop parameters in order to cultivate plants in every location in precisely the correct way, e.g. by supplying the perfect amount of fertilizer. One convenient way to gain such a distribution is measuring data on a combine harvester while driving across the field. When data is simply collected on board and transferred and evaluated later, huge amounts of data may be collected. However, being able to estimate the distribution while being on the field has some promising advantages ranging from getting improved status information or satisfying the curiosity of farmers to being able to predict when the harvester needs to empty its tank. The major challenge in transmitting the data is having only access to low bandwidth links due to underdeveloped public land mobile network infrastructures in rural areas. We propose Compressive Field Estimate (CFE), an approach for sending compressed data on the go in order to gain an early estimate of the whole spatial scalar field of plant and ground parameters. Lastly, we demonstrate CFE's performance in a simulative evaluation based on a trace recorded by a harvester on a spacious field.

On trains, predetermined maintenance of wheels requires significant amounts of resources. Therefore, it is desirable to replace the scheduled maintenance by a condition-based maintenance strategy, where manual maintenance only takes place when it is actually necessary. For a minimally invasive retrofitting of wagons, a system consisting of a Wireless Sensor Network (WSN) powered by energy-harvesting offers the most convenient and least costly solution. However, tight time synchronization and high energy efficiency is essential for a successful realization. Established time synchronization techniques for WSNs do not meet the special requirements for this scenario. Looking for an alternative, we noticed that one of the main components of WSNs, the sensor, has been barely exploited for this problem. We will demonstrate the feasibility of using sensor-data-based time synchronization instead of classical message-based synchronization in wireless networks for the example scenario of on-board wheel diameter measurements on trains. In contrast to popular current time synchronization protocols for WSNs, our approach is compatible with every wireless hardware. We perform a simulative evaluation of our approach on various wireless protocols based on message delay distributions measured on Raspberry Pi.

Although a realistic simulation of wireless networks mainly depends on a proper model of radio wave propagation, it is common to simply assume a fixed, circular range of communication. It has been shown that mobility models affect the results of the performance evaluation of wireless networks, e.g., contact and inter-contact times for opportunistic networks. However, even when using more realistic mobility models that consider street maps, the calculation of contact metrics widely relies on simplified propagation models. More realistic models such as ray launchers require 3-dimensional maps of the environment. We assume that the low availability of such maps is a major reason for falling back to more simplified models. We show that the level of detail in widely available map data from the OpenStreetMap-project is sufficient in certain scenarios such as the calculation of contact metrics. Our results open up a path to more realistic simulative performance evaluations using arbitrary cities.