Philipp M. Glatz

A measurement platform for energy harvesting and software characterization in WSNs

Wireless Sensor Network (WSN) nodes are resource- constrained computing devices. Adaptive behavior of autonomously working WSNs tries to maximize the cost efficiency of deployments. This includes maximizing the lifetime through power consumption optimization and recharging energy reservoirs with the use of energy harvesting. The adaptive behavior that leads to efficient resource usage needs information about the WSNs energy balance for decision making. We present a novel platform to measure the harvested, stored and dissipated energy. For being applicable to different environments it allows to attach different energy harvesting devices (EHDs). EHDs do not provide power continuously. Power availability patterns are used to determine how these sources can be used efficiently. Models from harvesting theory try to adapt to it. We implement a model that targets energy neutrality on our platform. It is used to evaluate the model and improve it. Our novel platform can be used to evaluate theories that model different sources. It can utilize and characterize thermoelectric, piezoelectric and magnetic induction generators and solar cells. The measurement platform tracks energy dissipation too. Mote software is implemented to establish communication to the platform. A sample application on top of it shows that the system can be used for software characterization. This paper contributes a novel modular and low-power design for measurement platforms for WSNs. It shows utilization of different energy sources and the ability to supply different mote types. Our work shows how theories for energy harvesting can be evaluated and improved. Our work also contributes to the field of simulation and emulation through online software characterization. The approach improves in accuracy and completeness over the capabilities of offline simulation.

Evaluation of component-aware dynamic voltage scaling for mobile devices and wireless sensor networks

Energy efficiency is very important for mobile devices and wireless sensor networks (WSNs), because the consumable energy is limited. Therefore, the operating time of such devices depends mainly on the capacity of the energy storage component and on the average power consumption of the device. The power consumption depends on the supply voltage and on the activated components of the hardware. This work presents the evaluation of component-aware dynamic voltage scaling (CADVS). This low power technique combines the power-down of unused components and the minimization of the supply voltage. Typically, each component of the hardware (microcontroller, transceiver, sensors) has its own supply voltage range. Therefore, the minimum allowed supply voltage depends on the activated components. However, the activated components and consequently the minimum allowed supply voltage varies over time. CADVS uses voltage converter to adjusts the supply voltage of the hardware to save as much energy as possible. This work presents the evaluation of six different voltage converters. It has been shown that CADVS can be used to save up to 38.7% of the energy compared to a constant voltage supply using the introduced scenario while achieving the same end-user performance.

A system for accurate characterization of wireless sensor networks with power states and energy harvesting system efficiency

Wireless sensor network (WSN) nodes have to cope with severe power supply constraints. Energy harvesting system (EHS) technology is used for prolonging network lifetime. Robust operation of such systems heavily relies on accurate models of EHS efficiency and node power dissipation for calculating sustainable operation modes. A nodes energy balance can be described with a power state model (PSM). While for battery operated WSNs PSM measurement errors and battery effects have to be considered, this paper widens the point of view to EHS properties. We analyze varying PSMs at runtime, EHS efficiency and measurement errors impacts on duty cycled WSNs. We show a measurement system setup and results from profiling a PSM for Mica2 nodes with an EHS. We explain important considerations for such systems with showing the variance of a nodes energy consumption depending on its supply and we profile EHS efficiency for deducing an energy storage level operating point. The results contribute to problems in the field of low power WSNs and especially EHS supported ones. Robustness of duty cycled WSNs heavily depends on measurement error bounds and is impacted by EHS efficiency. The reader is provided with a methodology for measuring and modeling robust WSNs.

On distributed computation of information potentials

A common task of mobile wireless ad-hoc networks is to distributedly extract information from a monitored process. We define process information as a measure that is sensed and computed by each mobile node in a network. For complex tasks, such as searching in a network and coordination of robotic swarms, we are typically interested in the spatial distribution of the process information. Spatial distributions can be thought of as information potentials that recursively consider the richness of information around each node. This paper describes a localized mechanism for determining the information potential on each node based on local process information and the potential of neighboring nodes. The mechanism allows us to distributedly generate a spectrum of possible information potentials between the extreme points of a local view and distributed averaging. In this work, we describe the mechanism, prove its exponential convergence, and characterize the spectrum of information potentials. Moreover, we use the mechanism to generate information potentials that are unimodal, i.e., feature a single extremum. Unimodality is a very valuable property for chemotactic search, which can be used in diverse application tasks such as directed search of information and rendezvous of mobile agents.

TOSPIE2: tiny operating system plug-in for energy estimation

TinyOS2 (TOS2) is the state of the art operating system for wireless sensor network (WSN) programming. There are a number of testbeds and simulation tools for checking functional correctness and a few integrated development environments (IDEs) that support graphical user interface (GUI) and power profiling of WSN simulation as well. All these systems and environments are tailored towards design optimization of WSNs subject to functional correctness and energy conservation. Unfortunately, all these approaches lack self-contained support for energy harvesting systems (EHSs) which is the state of the art technique for tackling the energy conservation problem. Here, we present the design and implementation of a self-contained X-in-the-loop approach. TOSPIE2 is an Eclipse plug-in that integrates TOS2 installation and compilation support, power profiles and EHS efficiency in simulation and measurement.

A first step towards energy management for network coding in wireless sensor networks

Network coding is a suitable mean to come up against the effects of crossing information flows in wireless sensor networks (WSNs) where these areas are heavily impacted in terms of channel bandwidth, message delay and energy balance.

Opportunistic Network Coding for Energy Conservation in Wireless Sensor Networks

Wireless sensor networks (WSN) continuously enhance processing capabilities and miniaturization. However, there exists a design gap to energy and bandwidth availability. Especially battery technology cannot keep pace with demands of novel versatile services. A common approach for conserving channel capacity and energy is optimizing power-aware routing and different kind of duty cycling (DC) and harvesting technology. While these optimizations are usually dealt with separately, we provide a novel framework with integrating and analyzing these different aspects in a practical TinyOS implementation at the same time. We implement the essential combination of energy harvesting aware routing (EHAR) together with radio and application DC and we add the novel approach of opportunistic network coding (ONC) for WSNs. We give detailed analysis of the applicability of application-level DC compared to low-power MAC and power save modes for state-of-the-art WSN and harvesting system hardware. We elaborate static and dynamic aspects of EHAR, scalable network coding (SNC) and ONC. Combining analytical models, comprehensive simulation and detailed highly accurate hardware power profiling measurement results, we demonstrate energy conservation from 13% to 50% when applying ONC and SNC.

Energy Conservation with Network Coding for Wireless Sensor Networks with Multiple Crossed Information Flows

The theory of network coding is hardly ever used and cannot be mapped to general wireless sensor network (WSN) topologies without careful consideration of technology constraints. Severe energy constraints and low bandwidth are faced by platforms of low computational power. We show how network coding methods can be implemented with low computational power. We discuss extensive experimentation in simulation of scalability for arbitrarily large networks. It is configured according to the results from mote hardware measurements. The testbed implementation and the mote software implementation is kept completely generic. All phases, including the initialization, are implemented and work for mesh networks of any size without modification. Our work explains important considerations for applying network coding to WSNs. The gain from applying the method over ad-hoc on-demand (AODV) like routing closely approaches the optimum value of 2. Thereby WSN resource constraints are relaxed and network lifetime is prolonged.

Designing of efficient energy harvesting systems for autonomous WSNs using a tier model

Energy harvesting systems (EHSs) are the key to perpetual operation of electronic devices in application areas with bad infrastructure or mobility. Wireless sensor networks (WSNs) are often used in such areas. Normal WSN nodes are powered by batteries. Therefore, the lifetime is limited and the batteries have to be replaced manually after a certain period of time. This problem can be solved by EHSs. They exploit energy sources of the environment and store the harvested energy in energy buffers. The EHS supplies the electronic device and ensures a continuous operation. WSNs can benefit from these developments, because the lifetime can be enhanced dramatically. However, the EHS have to be adapted to the requirements of the application area and of the supplied device. This enhances the overall efficiency of EHS. To be able to do that, the fundamental mode of operation of an EHS has to be well-understood. We introduce a novel tier model for EHSs. It structures the EHS into tiers with special functions. This enhances the design process of an EHS, because tiers can be adapted to each other and the overall efficiency of can be increased. The tier model is applied to RiverMote, a WSN node for in-river water level monitoring. Each node is supplied by solar cells and the energy is stored in double layer capacitors (DLCs). The hardware of the EHS of RiverMote is divided into the tiers of the model. These tiers are adapted to each other carefully. Although no maximum power point tracker has been implemented, it has been shown that the available power of the solar cell is greater than 80 % of the maximum power point if the energy level of the DLCs is between 42 % and 100 %. This result was only possible by a careful design and an adaption of the tiers.

Towards an on-site characterization of energy harvesting devices for wireless sensor networks

Wireless sensor networks (WSNs) suffer from the lack of wired infrastructure. Each node needs its own power supply, e.g. batteries or energy harvesting systems (EHSs). Typically, EHSs can extend the lifetime of a sensor node or even enable perpetual operation. Due to the high variation of harvestable energy of the environment, the design of the EHSs has to be done very carefully. The design process can be enhanced by using simulation of the WSN including energy harvesting. However, a realistic simulation needs accurate data of the harvestable energy of the environment. This paper presents the concept of an on-site characterization instrument for different types of energy harvesting devices. These instruments can be connected like a WSN.