Batteryless Internet-of-Things (IoT) devices need to schedule tasks on very limited energy budgets from intermittent energy harvesting. Creating an energy-aware scheduler allows the device to schedule tasks in an efficient manner to avoid power loss during execution. To achieve this, we need insight in the Worst-Case Energy Consumption (WCEC) of each schedulable task on the device. Different methodologies exist to determine or approximate the energy consumption. However, these approaches are computationally expensive and infeasible to perform on all type of devices; or are not accurate enough to acquire safe upper bounds. We propose a hybrid methodology that combines machine learning-based prediction on small code sections, called hybrid blocks, with static analysis to combine the predictions to a final upper bound estimation for the WCEC. In this paper, we present our work on an automated testbench for the Code Behaviour Framework (COBRA) that measures and profiles the upper bound energy consumption on the target device. Next, we use the upper bound measurements of the testbench to train eight different regression models that need to predict these upper bounds. The results show promising estimates for three regression models that could potentially be used for the methodology with additional tuning and training.

Wildfires are becoming increasingly devastating, and detecting them early is essential to containing them. Deep learning-based wildfire detection systems have increased in complexity dramatically in recent years, and in order to manage this added complexity, techniques have been proposed to automate the design of neural network architectures. Such techniques are usually referred to as Neural Architecture Search (NAS). This paper showcases the use of a reinforcement learning-based NAS agent to design a small neural network to perform active fire detection on multispectral satellite imagery. Specifically, we aim to automatically design a neural network that can determine if a single multispectral pixel is a part of a fire, and do so within the constraints of a Low Earth Orbit (LEO) nanosatellite with a limited power budget, to facilitate on-board processing of sensor data. A regression model that predicts the F1 score obtained by a particular architecture following quantization is used as a reward function. This model is trained on the classification performance statistics of a sample of neural network architectures. Besides the F1 score, we also include the total number of parameters in our reward function to limit the size of the designed model. Finally, we deployed the best neural network to the Google Coral Micro Dev Board and evaluated its inference latency and power consumption. This neural network consists of 1,716 parameters, takes on average 984μs to inference, and consumes around 800mW to perform inference. These results show that our approach can be applied to new problems.

Mars has long captivated the human imagination as a potential destination for settlement and scientific exploration. After deploying individual rovers, the next step in our journey to Mars is the autonomous exploration of the Red Planet using a collaborative swarm of rovers, drones, and satellites. This concept paper envisions a sustainable Mars exploration scenario featuring energy-aware, collaborative, and autonomous vehicles, including rovers, drones, and satellites, operating around Mars. The proposed framework is designed to address key challenges in energy management, edge intelligence, communication, sensing, resource-aware task scheduling, and radiation hardening. This work not only identifies these critical areas of research but also proposes novel technological solutions drawn from terrestrial advancements to extend their application to extraterrestrial exploration.

Machine Learning (ML) has made its way into a wide variety of advanced applications, where high accuracies can be achieved when these ML models are evaluated in the same context as they were trained and validated on. However, when these high-accuracy models are exposed to out-of-distribution points such as noisy inputs, their performance could potentially degrade significantly. Recommending the most suitable ML model that retains a higher accuracy when exposed to these noisy inputs can overcome this performance degradation. For this, a mapping between the noise distribution at the input and the resulting accuracy needs to be obtained. Though, this relationship is costly to evaluate as this is a computationally intensive task. To minimize this computional cost, we employ metalearning to predict this mapping; that is, the performance of different ML models is predicted given the distribution parameters of the input noise. Although metalearning is an established research field, performance predictions based on noise distribution parameters have not been accomplished before. Hence, this research focuses on predicting the per-class classification performance based on the distribution parameters of the input noise. For this, our approach is twofold. First, in order to gain insights in this noise-to-performance relationship, we analyse the per-class performance of well-established convolutional neural networks through our multi-level Monte Carlo simulation. Second, we employ metalearning to learn this relationship between the input noise distribution and the resulting per-class performance in a sample-efficient way by incorporating Latin Hypercube Sampling. The noise performance analyses present novel insights about the per-class performance degradation when gradually increasing noise is augmented on the input. Additionally, we show that metalearning is capable of accurately predicting the per-class performance based on the noise distribution parameters. We also show the relationship between the number of metasamples and the metaprediction accuracy. Consequently, this research enables future work to make accurate classifier recommendations in noisy environments.

Hyperdimensional Computing (HDC) is an emerging AI algorithm, touted to be an efficient, neuro-inspired and reliable alternative to neural networks for Edge AI. HDC utilizes hypervectors with several thousand elements; the number of elements in these hypervectors denotes the HDC dimension. This dimension can be optimized for improving the efficiency and reliability of HDC inference against errors such as bit-flips, which can be caused by environmental radiation-induced soft errors. We hypothesize that, by reducing the runtime chip area and execution time utilized by HDC inference through lowering dimensionality, both efficiency and reliability against soft error-induced bit-flips can be simultaneously improved while trading off a negligible amount of accuracy and error threshold. We tested our hypothesis by executing an HDC inference algorithm with two different dimension values, 10000 (10k) and 1024, on a commercially available, low-power, bare-metal ARM platform with a Cortex-M4 processor. We conducted the efficiency analysis by measuring the CPU cycles and energy required for executing the algorithm, and the reliability analysis using real-world atmospheric-like neutron radiation from the ChipIr facility in Oxfordshire, UK. Analyses revealed that, by lowering the HDC dimension from 10k to 1024, the reliability of HDC inference against soft error-induced bit-flips was 3.5 times better and efficiency improved by more than 16 times. This innovative observation contrasts the prevailing understanding in the community that increasing the HDC dimension always improves robustness or reliability. To the best of our knowledge, our work is the first to study the reliability of HDC inference using real-world radiation.