Wireless sensors form an integral part of the Internet of Things (IoT), standing at the edge between the cyber and physical domains. Ac-quiring and transmitting environmental data is an energy-intensive workload, especially when considering networks spanning large buildings or even cities. Many works aim to integrate energy har-vesting into wireless sensors, providing them with a greater level of energy autonomy. Initially deployed for energy-rich outdoor environments, recent advances have allowed wireless sensors to efficiently utilize the reduced energy harvested in indoor lighting conditions. This demo introduces a harvesting-based Dual Proces-sor Platform, the DPP3e, designed for energy harvesting in indoor environments. It features various sensors for advanced indoor en-vironmental sensing, e.g. air quality measurements, a low-power display for immediate visual feedback, and a powerful micro controller for energy-efficient inference of Tensorflow models. Fur-thermore, it has two separate RF interfaces: a 2.4 GHz Bluetooth Low Energy (BLE) radio for short-range communication, and a sub-GHz transceiver for long-range communication. Using configurable power domains and advanced power management, it can sustain sending BLE packets every 5 seconds while consuming only 37 µ W.

With the Internet of Things (IoT) large amounts of data can be gathered at the edge, centrally collected and subsequently utilized for various application domains. Efficient and reliable synchro-nous communication protocols are essential for automated data gathering, yet they typically require a stable energy supply. En-ergy harvesting enables long-term deployments, but it imposes widely varying energy budgets on each node in the network. To re-main efficient, synchronous protocols need to consider this energy variability. We propose a selective flooding protocol that employs low-energy communication rounds for all nodes and additional high-energy rounds only for nodes with high input power thus increasing their throughput. Low-energy rounds gather data with high reliability and maintain global network synchronization, while high-energy rounds use increased transmit power to overcome potentially-broken links that depend on low-power nodes. We eval-uate our proposed method on the FlockLab testbed and build a small network composed of standalone energy harvesting nodes. The average power consumption of almost 84 µW and 177 µW for low- and high-power nodes, respectively, are fully sustainable by indoor photovoltaic harvesting.

Low-power sensors are becoming ever more powerful, increasing both their energy efficiency as well as their processing capabilities. Much work in recent years has focused on optimizing machine learning models to low-power systems, typically to locally process sensor data. Significantly less attention has been paid to other artificial intelligence fields such as knowledge representation and automated reasoning, which may contribute to building autonomous devices. In this work, we present a low-power sensor node with an autonomous belief-desire-intention agent. This kind of agent simplifies the implementation of both proactive and reactive behaviors, promoting autonomy in our target applications. It does so by locally perceiving and reasoning, and then wirelessly broadcasting an intention, which can be forwarded to an actuator. The capabilities of the autonomous agent are demonstrated with a light-control application. Experiments demonstrate the feasibility of running intelligent agents in low-power platforms with little overhead.

Sensing systems powered by energy harvesting have traditionally been designed to tolerate long periods without energy. As the Internet of Things (IoT) evolves toward a more transient and opportunistic execution paradigm, reducing energy storage costs will be key for its economic and ecologic viability. However, decreasing energy storage in harvesting systems introduces reliability issues. Transducers only produce intermittent energy at low voltage and current levels, making guaranteed task completion a challenge. Existing ad hoc methods overcome this by buffering enough energy either for single tasks, incurring large data-retention overheads, or for one full application cycle, requiring a large energy buffer. We present Julienning : an automated method for optimizing the total energy cost of batteryless applications. Using a custom specification model, developers can describe transient applications as a set of atomically executed kernels with explicit data dependencies. Our optimization flow can partition data- and energy-intensive applications into multiple execution cycles with bounded energy consumption. By leveraging interkernel data dependencies, these energy-bounded execution cycles minimize the number of system activations and nonvolatile data transfers, and thus the total energy overhead. We validate our methodology with two batteryless cameras running energy-intensive machine learning applications. Using a solar testbed, we replay real-world illuminance traces to experimentally demonstrate optimized batteryless execution with a transducer-to-application energy efficiency of 74.5%. Partitioning results demonstrate that compared to ad hoc solutions, our method can reduce the required energy storage by over 94% while only incurring a 0.12% energy overhead.

Low-power operation of power-hungry MOX sensors is usually achieved by duty cycling them periodically. However, a difficulty arises if a sensor is operated at irregular time intervals due to intermittent energy availability in energy- neutral or batteryless applications. In this work, we propose a compensation method which re-maps on-demand measurements to virtually duty-cycled readings in the value domain by taking the duration of the last off-time into account. We evaluate our compensation algorithm based on Sensirion’s SGP30 sensor and achieve up to 79% accuracy improvement compared to uncompensated measurements.

Popular low-cost air quality sensors embedded into IoT and mobile devices are based on metal oxides (MOX) that change their electrical resistance in response to ambient pollutants emitted as gases. Operating MOX sensors continuously is expensive, since it requires to heat up and maintain a hotplate at several hundred degrees. To save energy, sensors are commonly duty cycled with short on-times and long off-times. However, doing so adversely affects the sensor's chemical reactions, which have slower transients as the off-time increases. As a result, sensor sensitivity to various gases deviates from a continuously powered sensor. In this paper, we show that it is possible to recover accurate continuous-sensor measurements from transient responses obtained from a duty cycled sensor and compensate for an altered multi-gas cross-sensitivity profile using machine learning methods. On a test set, we achieve a mean absolute error (MAE) of 24ppb between continuous ground-truth measurements and obtained model predictions of tVOC. This results in estimating 86.6% of Indoor Air Quality (IAQ) levels correctly compared to 68.1% if no correction is used. Our models are invariant to minor baseline shifts and work for both tVOC and CO2-eq signals provided by the sensor. Thanks to our models, 98.5% of the energy consumption can be reduced while maintaining high accuracy. This optimization enables energy-harvesting-based operation of IAQ sensors in indoor IoT scenarios.

Sensing systems for long-term monitoring constitute an important part of the emerging Internet of Things. In this domain, energy harvesting and infrastructure-less communication enable truly autonomous and maintenance-free operation of sensor nodes gathering long-term environmental data. Due to the infrastructure-less nature of the communication, receivers are not always available. The variable energy provided by the environment and the receiver’s mobility lead to non-deterministic node availability. In this work, we study infrastructure-less data transmission schemes to optimize communication when both senders and receivers exhibit intermittent behavior. We rely on the notion of data utility, describing the importance of sensed data to the receiver, to determine an optimal communication scheme. Deriving the communication policy that maximizes the utility of the received data is shown to be a convex optimization problem. The resulting scheme is implemented and validated on a batteryless Bluetooth Low Energy sensor node that communicates to commodity smartphones. Our evaluation demonstrates that the model accurately captures the application scenario with a maximum root-mean-square error of less than 0.016 in data reception probability. The communication scheme’s adaptiveness to variable harvesting conditions is experimentally demonstrated under varying harvesting conditions and is shown to significantly increase the data utility.

Detecting the amount of people occupying an environment is an important use case for surveillance in public spaces such as airports, stations and squares, but also for smaller environments such as classrooms (e.g. to track occupation of classrooms). Using visible imaging for this task is often suboptimal because 1) it potentially violates user privacy 2) to have a good final count, high resolution cameras are required. Long-wave infrared imaging is a viable solution to both these issues. In this paper, we developed a people counting algorithm on thermal images based on convolutional neural networks (CNNs) small enough that they can run on a limited-memory low-power platform. We created a dataset with 3k manually tagged thermal images and developed a fast and accurate CNN that is able to provide a completely error-free detection on 53.7% of the test images and an error bound within ±1 detection in 84.4% of the images, using only 308 kilobytes of system memory in a Cortex M4 platform.

While energy harvesting is generally seen to be the key to power cyber-physical systems in a low-cost, long-term, efficient manner, it has generally required large energy storage devices to mitigate the effects of the source’s variability. The emerging class of transiently powered systems embrace this variability by performing computation in proportion to the energy harvested, thereby minimizing the obtrusive and expensive storage element. By using an efficient Energy Management Unit (EMU), small bursts of energy can be buffered in an optimally sized capacitor and used to supply generic loads, even when the average harvested power is only a fraction of that required for sustained system operation. Dynamic Energy Burst Scaling (DEBS) can be used by the load to dynamically configure the EMU to supply small bursts of energy at its optimal power point, independent from the harvester’s operating point. Parameters like the maximum burst size, the solar panel’s area, as well as the use of energy-efficient Non-Volatile Memory Hierarchy (NVMH) can have a significant impact on the transient system’s characteristics such as the wake-up time and the amount of work that can be done per unit of energy. Experimental data from a solar-powered, long-term autonomous image acquisition application show that, regardless of its configuration, the EMU can supply energy bursts to a 43.4mW load with efficiencies of up to 79.7% and can work with input power levels as low as 140μW. When the EMU is configured to use DEBS and NVMH, the total energy cost of acquiring, processing and storing an image can be reduced by 77.8%, at the price of increasing the energy buffer size by 65%.

With the appearance of wearable devices and the IoT, energy harvesting nodes are becoming more and more important. The design and evaluation of these small standalone sensors and actuators, which harvest limited amounts of energy, requires novel tools and methods. Fast and accurate measurement systems are required to capture the rapidly changing harvesting scenarios and characterize leakage currents and energy efficiencies. The need for real-world experiments creates a demand for compact and portable equipment to perform in-situ power measurements and environmental logging. This work presents the RocketLogger, a hand-held measurement device that combines both properties: portability and accuracy. The custom analog front-end allows logging at sampling rates up to 64 kSPS. The fast range switching within 1.4 μ8 guarantees continuous power measurements starting from 4pW at 1 mV up to 2.75 W at 5.5 V. The software provides remote control and manages data acquisition of up to 13Mb/ sec in real-time. We extensively characterize the RocketLogger's performance, demonstrate the need for its properties in three use-cases at different stages of the system design flow, and show its advantages in measuring and validating new harvesting-driven devices for the IoT.

We demonstrate the RocketLogger, a mobile data logger designed for prototyping energy harvesting IoT devices. Novel IoT applications require new dataloggers with a highly increased dynamic range for current measurement to accommodate both ultra-low sleep currents of few nanoamperes as well as wireless communication currents in the range of hundreds of milliamperes. In parallel to ultra-low currents and high dynamic range measurements, novel applications require mobile measurements for easy in-situ characterization or wearable device testing. The RocketLogger is a solution that fulfills these requirements. While being fully mobile, it measures currents from 5 nA up to 500 mA with very fast and seamless range-switching. Using a sample energy harvesting application, we demonstrate its low-current measurement capabilities, fast, seamless auto-ranging and easy-to-use remote user interface.

Multicore systems are being increasingly used for embedded system deployments, even in safety-critical domains. Co-hosting applications of different criticality levels in the same platform requires sufficient isolation among them, which has given rise to the mixed-criticality scheduling problem and several recently proposed policies. Such policies typically employ runtime mechanisms to monitor task execution, detect exceptional events like task overruns, and react by switching scheduling mode. Implementing such mechanisms efficiently is crucial for any scheduler to detect runtime events and react in a timely manner, without compromising the system’s safety. This paper investigates implementation alternatives for these mechanisms and empirically evaluates the effect of their runtime overhead on the schedulability of mixed-criticality applications. Specifically, we implement in user-space two state-of-the-art scheduling policies: the flexible time-triggered FTTS [1] and the partitioned EDFVD [2], and measure their runtime overheads on a 60-core Intel R Xeon Phi and a 4-core Intel R Core i5 for the first time. Based on extensive executions of synthetic task sets and an industrial avionic application, we show that these overheads cannot be neglected, esp. on massively multicore architectures, where they can incur a schedulability loss up to 97%. Evaluating runtime mechanisms early in the design phase and integrating their overheads into schedulability analysis seem therefore inevitable steps in the design of mixed-criticality systems. The need for verifiably bounded overheads motivates the development of novel timing-predictable architectures and runtime environments specifically targeted for mixed-criticality applications.

The trend to integrate multiple functionalities on the same (off-the-shelf) hardware has made the selection of the right scheduling algorithm and configuration difficult. This selection requires the designer to validate any scheduling decision already during early design steps on the target architecture, e.g., by using a reconfigurable scheduling framework running in the user-space. In this paper, we first identify the requirements that such a scheduling framework must fulfill. Then, we propose SF3P: an open-source framework that meets these requirements. To this end, we define an interface common to all scheduling algorithms and separate the scheduling algorithm from its low-level implementation. With these features, SF3P can not only prototype a scheduler at high level of abstraction, but also execute the implemented task-set on specific hardware. Furthermore, SF3P can hierarchically compose scheduling algorithms, useful in the mixed criticality domain, and could also be used to explore different scheduling policies in the system optimization phase. We demonstrate these features by implementing SF3P on top of a POSIX-compliant operating system on two different platforms: Raspberry Pi and an Intel Core i7 desktop system.