We present MiDebug, a web-based Integrated Development Environment (IDE) for embedded system programming with in-browser debugging capabilities. This web application greatly reduces the time and effort required for rapid prototyping of microcontroller based devices.

We present DoubleDip, a low power monitoring system for enabling non-intrusive water flow detection. DoubleDip taps into minute thermal gradients in pipes for both replenishing energy reserves and performing low power wakeup. One of the remaining issues with wireless water monitoring in residences and offices is that current solutions require installing sensor nodes with access to electrical wiring or replacing batteries frequently. DoubleDip (DD) significantly extends the lifetime of vibration-based non-intrusive water flow sensors by harvesting thermal energy from hot pipes wherever accessible. DoubleDip requires less than an inch of exposed metal pipe to attach a coupler for gathering sufficient energy to power itself, in some cases, into perpetuity. We observe that water use in homes and offices is incredibly sporadic, making continuous monitoring both impractical and wasteful. Instead, DD puts a thermoelectric harvester into double duty. It uses thermal gradients not only for gathering energy but also for extremely low power (< 1μA) wakeup. In this paper, we describe the DoubleDip design and demonstrate that thermoelectric wakeup is essential for longevity and accuracy. Since DD wakes up from its low power state only when there is a water flow event, it replenishes the energy it uses in sensing and transmitting data by the energy it harvests from the corresponding heat gradient. While DD nodes installed on cold water pipes harvest far less than those installed on hot water pipes, our pilot deployment over four weeks and five locations suggests that thermoelectric wake up is only slightly worse in latency for cold water monitoring and there is sufficient energy harvested from the hot water that it can be shared to extend the lifetime of nearby cold water nodes too.

Wireless neural recording systems are subject to stringent power consumption constraints to support long-term recordings and to allow for implantation inside the brain. In this paper, we propose using a combination of on-chip detection of action potentials ("spikes") and compressive sensing (CS) techniques to reduce the power consumption of the neural recording system by reducing the power required for wireless transmission. We empirically verify that spikes are compressible in the wavelet domain and show that spikes from different neurons acquired from the same electrode have subtly different sparsity patterns or supports. We exploit the latter fact to further enhance the sparsity by incorporating a union of these supports learned over time into the spike recovery procedure. We show, using extra cellular recordings from human subjects, that this mechanism improves the SNDR of the recovered spikes over conventional basis pursuit recovery by up to 9.5 dB (6 dB mean) for the same number of CS measurements. Though the compression ratio in our system is contingent on the spike rate at the electrode, for the datasets considered here, the mean ratio achieved for 20-dB SNDR recovery is improved from 26:1 to 43:1 using the learned union of supports.

Data loss in wireless sensing applications is inevitable and while there have been many attempts at coping with this issue, recent developments in the area of Compressive Sensing (CS) provide a new and attractive perspective. Since many physical signals of interest are known to be sparse or compressible, employing CS, not only compresses the data and reduces effective transmission rate, but also improves the robustness of the system to channel erasures. This is possible because reconstruction algorithms for compressively sampled signals are not hampered by the stochastic nature of wireless link disturbances, which has traditionally plagued attempts at proactively handling the effects of these errors. In this paper, we propose that if CS is employed for source compression, then CS can further be exploited as an application layer erasure coding strategy for recovering missing data. We show that CS erasure encoding (CSEC) with random sampling is efficient for handling missing data in erasure channels, paralleling the performance of BCH codes, with the added benefit of graceful degradation of the reconstruction error even when the amount of missing data far exceeds the designed redundancy. Further, since CSEC is equivalent to nominal oversampling in the incoherent measurement basis, it is computationally cheaper than conventional erasure coding. We support our proposal through extensive performance studies.

To achieve more accurate global time synchronization, this paper argues for decoupling the clock distribution network from the routing tree in a multihop wireless network. We find that both flooding and routing-integrated time synchronization rapidly propagate node-level errors (typically due to temperature fluctuations) across the network. Therefore, we propose that a node chooses synchronization neighbors that offer the greatest frequency stability. We propose two methods to estimate a neighbor’s stability. The first approach selects the neighbor whose Frequency Error Variance, or simply FEV, is smallest with respect to the local clock. The second approach selects the neighbor that reports the lowest FEV relative to its synchronization parent. We also propose the node-level time-variance FEV as an additive metric for selecting more stable clock trees than either naïve flooding or routing-integrated time synchronization can provide. We incorporate these techniques into FTSP, a widelyused time synchronization protocol, and show that the mean error in global time significantly improved (by a factor of five) when some nodes are warmed and others are not.

The efficiency of the time synchronization service in wireless sensor networks is tightly connected to the design of the radio, the quality of the clocking hardware, and the synchronization algorithm employed. While improvements can be made on all levels of the system, over the last few years most work has focused on the algorithmic level to minimize message exchange and in radio architectures to provide accurate time-stamping mechanisms. Surprisingly, the influences of the underlying clock system and its impact on the overall synchronization accuracy has largely been unstudied. In this work, we investigate the impact of the clocking subsystem on the time synchronization service and address, in particular, the influence of changes in environmental temperature on clock drift in highly duty-cycled wireless sensor nodes. We also develop formulas that help the system architect choose the optimal resynchronization period to achieve a given synchronization accuracy. We find that the synchronization accuracy has a two region behavior. In the first region, the synchronization accuracy is limited by quantization error, while int he second region changes in environmental temperature impact the achievable accuracy. We verify our analytic results in simulation and real hardware experiments.

A key prerequisite for residential energy conservation is knowing when and where energy is being spent. Unfortunately, the current generation of energy reporting devices only provide partial and coarse grained information or require expensive professional installation. This limitation stems from the presumption that calculating per-appliance consumption requires per-appliance current measurements. However, since appliances typically emit measurable signals when they are consuming energy, we can estimate their consumption using indirect sensing. This paper presents ViridiScope, a fine-grained power monitoring system that furnishes users with an economical, self-calibrating tool that provides power consumption of virtually every appliance in the home. ViridiScope uses ambient signals from inexpensive sensors placed near appliances to estimate power consumption, thus no in-line sensor is necessary. We use a model-based machine learning algorithm that automates the sensor calibration process. Through experiments in a real house, we show that ViridiScope can estimate the end-point power consumption within 10% error.

Compressive Sensing (CS) is a recently developed mechanism that allows signal acquisition and compression to be performed in one inexpensive step so that the sampling process itself produces a compressed version of the signal. This significantly improves systemic energy efficiency because the average sampling rate can be considerably reduced and explicit compression eliminated.

Time keeping and synchronization are important services for networked and embedded systems. High quality timing information allows embedded network nodes to provide accurate time-stamping, fast localization, efficient duty cycling schedules, and other basic but essential functions - all of which are required for low power operation.

Water is nature's most precious resource and growing demand is pushing fresh water supplies to the brink of non-renewability. New technological and social initiatives that enhance conservation and reduce waste are needed. Providing consumers with fine-grained real-time information has yielded benefits in conservation of power and gasoline. Extending this philosophy to water conservation, we introduce a novel water monitoring system, NAWMS, that similarly empowers users.

Time synchronization is an essential service in distributed computing and control systems. It is used to enable tasks such as synchronized data sampling and accurate time-of-flight estimation, which can be used to locate nodes. The deviation in nodes' knowledge of time and inter-node resynchronization rate are affected by three sources of time stamping errors: network wireless communication delays, platform hardware and software delays, and environment-dependent frequency drift characteristics of the clock source. The focus of this work is on the last source of error, the clock source, which becomes a bottleneck when either required time accuracy or available energy budget and bandwidth (and thus feasible resynchronization rate) are too stringent. Traditionally, this has required the use of expensive clock sources (such as temperature compensation using precise sensors and calibration models) that are not cost-effective in low-end wireless sensor nodes. Since the frequency of a crystal is a product of manufacturing and environmental parameters, we describe an approach that exploits the subtle manufacturing variation between a pair of inexpensive oscillators placed in close proximity to algorithmically compensate for the drift produced by the environment. The algorithm effectively uses the oscillators themselves as a sensor that can detect changes in frequency caused by a variety of environmental factors. We analyze the performance of our approach using behavioral models of crystal oscillators in our algorithm simulation. Then we apply the algorithm to an actual temperature dataset collected at the James Wildlife Reserve in Riverside County, California, and test the algorithms on a waveform generator based testbed. The result of our experiments show that the technique can effectively improve the frequency stability of an inexpensive uncompensated crystal 5 times with the potential for even higher gains in future implementations.