Abstract-Energy harvesters are being used to power autonomous systems, but their output power is variable and intermittent. To sustain computation, these systems integrate batteries or supercapacitors to smooth out rapid changes in harvester output. Energy storage devices require time for charging and increase the size, mass and cost of systems. The field of transient computing moves away from this approach, by powering the system directly from the harvester output. To prevent an application from having to restart computation after a power outage, approaches such as Hibernus allow these systems to hibernate when supply failure is imminent. When the supply reaches the operating threshold, the last saved state is restored and the operation is continued from the point it was interrupted. This work proposes Hibernus++ to intelligently adapt the hibernate and restore thresholds in response to source dynamics and system load properties. Specifically, capabilities are built into the system to autonomously characterize the hardware platform and its performance during hibernation in order to set the hibernation threshold at a point which minimizes wasted energy and maximizes computation time. Similarly, the system auto-calibrates the restore threshold depending on the balance of energy supply and consumption in order to maximize computation time. Hibernus++ is validated both theoretically and experimentally on microcontroller hardware using both synthesized and real energy harvesters. Results show that Hibernus++ provides an average 16% reduction in energy consumption and an improvement of 17% in application execution time over stateof-the-art approaches.

Energy harvesting sensor systems typically incorporate energy buffers (e.g., rechargeable batteries and supercapacitors) to accommodate fluctuations in supply. However, the presence of these elements limits the miniaturization of devices. In recent years, researchers have proposed a new paradigm, transient computing, where systems operate directly from the energy harvesting source and allow computation to span across power cycles, without adding energy buffers. Various transient computing approaches have addressed the challenge of power intermittency by retaining the processor’s state using non-volatile memory. However, no generic approach has yet been proposed to retain the state of peripherals external to the processing element. This paper proposes RESTOP, flexible middleware which retains the state of multiple external peripherals that are connected to a computing element (i.e., a microcontroller) through protocols such as SPI or I2C. RESTOP acts as an interface between the main application and the peripheral, which keeps a record, at run-time, of the transmitted data in order to restore peripheral configuration after a power interruption. RESTOP is practically implemented and validated using three digitally interfaced peripherals, successfully restoring their configuration after power interruptions, imposing a maximum time overhead of 15% when configuring a peripheral. However, this represents an overhead of only 0.82% during complete execution of our typical sensing application, which is substantially lower than existing approaches.

Energy harvesters are widely used to power wireless sensor systems, but the produced power is generally low, and can vary abruptly due to changes in the environment or the device's location. Energy buffers (batteries or supercapacitors) are normally incorporated into systems to smooth out these variations. However, they have a limited lifetime and increase system size and cost. Transient computing aims to address these issues by removing the energy buffer, and powering the system directly from the energy harvester. Approaches such as Hibernus++ deal with the resultant power intermittency by 'hibernating', i.e. saving a snapshot of the system state to non-volatile memory before a power failure, and restoring it after the power recovers. The overheads of this can be particularly costly with a low-current harvester, as the system may wake up and hibernate at a high frequency, doing little useful work in each power cycle. This paper proposes an enhancement to these approaches, providing an efficient method to avoid repeated hibernation. The introduction of a 'sleep' state, which is entered when the power supply is detected to be failing, allows the system's supply voltage to recover without taking a snapshot. Thus, the application can spend more time on useful work rather than checkpointing. If the supply voltage continues to decline, a snapshot will then be taken. The approach has been simulated and experimentally validated, with results demonstrating that the proposed scheme provides up to a 65% improvement in system active run-time with low-current harvesters vs. conventional Hibernus++.

Transient computing enables application execution to be performed despite power outages. Although it handles the non-deterministic nature of energy harvesting (EH), sensor systems envisioned by the IoT seek more cost-and volume-effective solutions, which are better tailored to application requirements. Additionally, a major drawback of transient computing, keeping track of time, hinders its widespread adoption in the IoT. To overcome these challenges, this paper proposes a control flow for sensor systems by combining two state-of-the-art transient computing schemes in an energy-aware manner, underpinned by a strategy for timekeeping. It enables application execution to be reliably performed even under the most severe EH conditions, with an improved cost and volume efficiency, i.e., smaller energy storage. Benefiting from the combination of the two schemes, dynamic adjustment of system performance is achieved, while the time is accurately tracked. To illustrate the applicability of this flow to actual sensor systems, two case studies: a bicycle trip computer and a step counter, are presented. Empirical results reveal that, even with a tiny amount of energy harvested (tens of µJ), our proposed approach can meet application requirements with smaller storage, i.e., 40% and 66% reduction in required capacitance for the presented case studies.

Systems operating from harvested sources typically integrate batteries or supercapacitors to smooth out rapid changes in harvester output. However, such energy storage devices require time for charging and increase the size, mass and cost of the system. A recent approach to address this is to power systems directly from the harvester output, termed transient computing. To solve the problem of having to restart computation from the start due to power-cycles, a number of techniques have been proposed to deal with transient power sources. In this paper, we quantitatively evaluate three state-of-the-art approaches on a Texas Instruments MSP430 microcontroller characterizing the application scenarios where each performs best. Finally, recommendations are provided to system designers for selecting the most suitable approach.