David Hasenfratz

Sensing the air we breathe: the opensense Zurich dataset

Monitoring and managing urban air pollution is a significant challenge for the sustainability of our environment. We quickly survey the air pollution modeling problem, introduce a new dataset of mobile air quality measurements in Zurich, and discuss the challenges of making sense of these data.Molecular oxygen (O 2 )is a basic requirement for cellular growth and viability and many aspects of anatomy and physiology are dedicated to achieving reliable distribution. Recent work has identified a specific sensing and response system, centred around a transcription complex called Hypoxia-inducible Factor 1 (HIF-1), which forms the focus of this review. The HIF-system operates in all cell types and modulates a very broad range of cellular pathways, consistent with the broad importance of oxygen. It is implicated in a rapidly expanding range of developmental, physiological and pathological settings, and is potentially relevant to almost all areas of clinical medicine. Excitingly, the pathway can be activated with low molecular weight compounds which should offer therapeutic benefit, especially in diseases where oxygen supply is compromised.

On-the-Fly calibration of low-cost gas sensors

Air quality monitoring is extremely important as air pollution has a direct impact on human health. Low-cost gas sensors are used to effectively perceive the environment by mounting them on top of mobile vehicles, for example, using a public transport network. Thus, these sensors are part of a mobile network and perform from time to time measurements in each others vicinity. In this paper, we study three calibration algorithms that exploit co-located sensor measurements to enhance sensor calibration and consequently the quality of the pollution measurements on-the-fly. Forward calibration, based on a traditional approach widely used in the literature, is used as performance benchmark for two novel algorithms: backward and instant calibration. We validate all three algorithms with real ozone pollution measurements carried out in an urban setting by comparing gas sensor output to high-quality measurements from analytical instruments. We find that both backward and instant calibration reduce the average measurement error by a factor of two compared to forward calibration. Furthermore, we unveil the arising difficulties if sensor calibration is not based on reliable reference measurements but on sensor readings of low-cost gas sensors which is inevitable in a mobile scenario with only a few reliable sensors. We propose a solution and evaluate its effect on the measurement accuracy in experiments and simulation.

Analysis, Comparison, and Optimization of Routing Protocols for Energy Harvesting Wireless Sensor Networks

Energy harvesting has been steadily gaining interest in the wireless sensor network community. Instead of minimizing the energy consumption and maximizing a network’s operational time, the main challenge in energy harvesting sensor networks is to maximize the utility of the application subject to the harvested energy. One major challenge is to maximize the data delivery rates by exploiting the spatial variations of environmental energy. While there exists a multiplicity of energy-aware routing protocols for sensor networks without energy harvesting capabilities, only a small number of routing protocols have been published which explicitly account for energy harvesting. In this paper, we analyze and compare three state-of-the-art routing algorithms. While the original algorithms assume an idealized medium access control (MAC), a lossless wireless channel and global knowledge, we show that these assumptions lead to delusive results. We detail these findings by showing the influence of a low-power MAC protocol, a realistic wireless channel and the protocol overhead. Moreover, we show how to optimize the parameters of the MAC protocol for a given network configuration. By conducting various evaluations, we identify that our modified version of the R-MPRT algorithm outperforms the evaluated algorithms in scenarios where little energy is harvested from the environment.

Pushing the spatio-temporal resolution limit of urban air pollution maps

Up-to-date information on urban air pollution is of great importance for health protection agencies to assess air quality and provide advice to the general public in a timely manner. In particular, ultrafine particles (UFPs) are widely spread in urban environments and may have a severe impact on human health. However, the lack of knowledge about the spatio-temporal distribution of UFPs hampers profound evaluation of these effects. In this paper, we analyze one of the largest spatially resolved UFP data set publicly available today containing over 25 million measurements. We collected the measurements throughout more than a year using mobile sensor nodes installed on top of public transport vehicles in the city of Zurich, Switzerland. Based on these data, we develop land-use regression models to create pollution maps with a high spatial resolution of 100m × 100 m. We compare the accuracy of the derived models across various time scales and observe a rapid drop in accuracy for maps with subweekly temporal resolution. To address this problem, we propose a novel modeling approach that incorporates past measurements annotated with metadata into the modeling process. In this way, we achieve a 26% reduction in the root-mean-square error-a standard metric to evaluate the accuracy of air quality models-of pollution maps with semi-daily temporal resolution. We believe that our findings can help epidemiologists to better understand the adverse health effects related to UFPs and serve as a stepping stone towards detailed real-time pollution assessment.

Reducing multi-hop calibration errors in large-scale mobile sensor networks

Frequent sensor calibration is essential in sensor networks with low-cost sensors. We exploit the fact that temporally and spatially close measurements of different sensors measuring the same phenomenon are similar. Hence, when calibrating a sensor, we adjust its calibration parameters to minimize the differences between co-located measurements of previously calibrated sensors. In turn, freshly calibrated sensors can now be used to calibrate other sensors in the network, referred to as multi-hop calibration. We are the first to study multi-hop calibration with respect to a reference signal (micro-calibration) in detail. We show that ordinary least squares regression---commonly used to calibrate noisy sensors---suffers from significant error accumulation over multiple hops. In this paper, we propose a novel multi-hop calibration algorithm using geometric mean regression, which (i) highly reduces error propagation in the network, (ii) distinctly outperforms ordinary least squares in the multi-hop scenario, and (iii) requires considerably fewer ground truth measurements compared to existing network calibration algorithms. The proposed algorithm is especially valuable when calibrating large networks of heterogeneous sensors with different noise characteristics. We provide theoretical justifications for our claims. Then, we conduct a detailed analysis with artificial data to study calibration accuracy under various settings and to identify different error sources. Finally, we use our algorithm to accurately calibrate 13 million temperature, ground ozone (O3), and carbon monoxide (CO) measurements gathered by our mobile air pollution monitoring network.

Model-Driven Accuracy Bounds for Noisy Sensor Readings

Wireless sensor networks are increasingly used in application scenarios where a high data quality is inevitable, e.g., the control of industrial production areas. Nevertheless, many deployments must live with strict constraints regarding the sensing hardware and may not employ newest sensing technologies, e.g., due to limited energy budget, size, and bandwidth. Additionally, many applications would benefit from not only gathering absolute sensor readings but also knowing the quality of their low-cost sensor measurements. In this paper, we introduce a model-driven approach that (i) provides reliable accuracy bounds for individual noisy sensor readings and (ii) detects systematic and transient sensor errors. We apply our method to static and mobile real-world deployments of noisy and unstable low-cost sensors by analyzing large sets of urban temperature and ozone measurements. We find that the proposed algorithm successfully calculates precise accuracy bounds. We compare them to measurements of high-quality instruments and show that up to 96 % of the reference measurements are inside the computed accuracy bounds in the static scenario and up to 94 % in the mobile scenario. This is surprisingly high for the used low-cost sensors. By analyzing data from our static longterm deployment, we reveal that the ozone sensors reliability is dependent on seasonal weather conditions.

On Rendezvous in Mobile Sensing Networks

A rendezvous is a temporal and spatial vicinity of two sensors. In this chapter, we investigate rendezvous in the context of mobile sensing systems. We use an air quality dataset obtained with the OpenSense monitoring network to explore rendezvous properties for carbon monoxide, ozone, temperature, and humidity processes. Temporal and spatial locality of a physical process impacts the number of rendezvous between sensors, their duration, and their frequency. We introduce a rendezvous connection graph and explore the trade-off between locality of a process and the amount of time needed for the graph to be connected. Rendezvous graph connectivity has many potential use cases, such as sensor fault detection. We successfully apply the proposed concepts to track down faulty sensors and to improve sensor calibration in our deployment.

Spatially Resolved Monitoring of Radio-Frequency Electromagnetic Fields

Radio-frequency electromagnetic fields are emitted by many applications, such as radio broadcasting and mobile communication. A part of the general public is increasingly concerned about the long-term effects of electromagnetic radiation on human health. However, the accurate exposure assessment in peoples everyday life remains a formidable challenge. State-of-the-art personal exposure meters are expensive and tedious to use. Epidemiological large-scale studies are rare and governmental compliance measurements can only cover a small number of locations of high interest (e.g., schools). In this paper, we demonstrate that accurate, spatially resolved electromagnetic field measurements are feasible with commodity sensor nodes. We show the design, implementation, and evaluation on mobile air quality sensor nodes, which traverse a large urban area on top of public transport vehicles in Zurich, Switzerland. We collect a data set with over 4 million measurements and use it to develop the first exposure map of Zurich with a spatial resolution of 100 m. Further, we compare the found exposure levels to measurements from different urban cities across Europe.

Transactional Memory: How to perform load adaption in a simple and distributed manner

We analyze and present different strategies to adapt the load in transactional memory systems based on contention. Our experimental results show a substantial overall improvement for our best performing strategies QuickAdapter and AbortBackoff on the throughput when compared to the best existing contention management policies (without load adaption). Opposed to prior work our load adapting schemes are simple and fully distributed, while maintaining the same throughput rate. Our theoretical analysis gives insights into the usefulness of load adaption schemes. We show a constant expected speed-up compared to systems without load adaption in several important scenarios, but also illustrate that the worst-case behavior can result in an exponential increase in the running time.

Route selection for mobile sensor nodes on public transport networks

The sensing range of a sensor is spatially limited. Thus, achieving a good coverage of a large area of interest requires installation of a huge number of sensors which is cost and labor intensive. For example, monitoring air pollution in a city needs a high density of measurement stations installed throughout the area of interest. As alternative, we install a smaller number of mobile sensing nodes on top of public transport vehicles that regularly traverse the city. In this paper, we consider the problem of selecting a subnetwork of a city’s public transport network to achieve a good coverage of the area of interest. In general case, public transport vehicles are not assigned to fix lines but rather to depots where they are parked overnight. We introduce an algorithm that selects the installation locations, i.e., number of vehicles within each host depot, such that sensing coverage is maximized. Since we are working with low-cost sensors, which exhibit failures and drift over time, vehicles selected for sensor installation have to be in each other’s vicinity from time to time to allow comparing sensor readings. We refer to such meeting points as checkpoints. Our algorithm optimizes sensing coverage while providing a sufficient number of checkpoint locations. We evaluate our algorithm based on the tram network of Zurich and show how an accurate selection of vehicles for installing measurement stations affects the overall system quality. We show that our algorithm outperforms random search, simulated annealing, and the greedy approach.