Mid Sweden University

Autonomous condition monitoring is essential for advancing intelligent systems in both industrial and domestic Internet of Things (IoT) applications. However, continuous long-term condition monitoring is challenged by the limited energy availability for wireless sensor nodes (WSNs). Therefore, energy harvesting offers a promising approach by converting ambient or host energy into electrical power to sustain WSN operation. To bridge the gap between energy harvesting and condition monitoring, this review provides an overview and synthesis of recent advances in energy harvesting technologies tailored for condition monitoring applications. State-of-the-art developments in energy harvesting are categorized into six domains: healthcare, ocean, machinery, grid, railway, and infrastructure. The characteristics of these energy sources and their domain-specific monitoring requirements are analyzed. Furthermore, this review examines harvesting transducers, structural designs, and optimization methods employed in energy harvesters. Finally, the review discusses current challenges and future prospects for energy-autonomous condition monitoring systems, aiming to support the deployment of sustainable IoT sensing solutions. 

Accurate and robust recognition of daylily flower buds at the pre-bloom stage is essential for timely harvesting and quality preservation, yet remains highly challenging under natural field conditions due to the buds' slender morphology, diverse orientations, dense distribution, and frequent occlusion by foliage. Existing horizontal-box object detectors struggle to capture the orientation and geometric structure of daylily buds, leading to inaccurate localization and unreliable guidance for automated harvesting. To address these challenges, we propose DDM-YOLO, a lightweight orientation-aware detection model tailored for daylily production environments. The model integrates three key components: (i) a Multi-scale Adaptive Feature Pyramid Network (MAFPN) that enhances the extraction and fusion of multi-dimensional features for densely distributed and occluded slender buds; (ii) a Lightweight Adaptive Direction-aware Head (LADH) that dynamically optimizes angle regression for rotated bounding boxes, improving orientation stability and reducing localization bias; and (iii) an Adaptive Down-sampling module (Adown) that preserves structurally critical spatial cues while reducing model complexity. Experiments conducted on a custom daylily field dataset demonstrate that DDM-YOLO achieves 96.8% precision and 98.1% mAP50, outperforming the baseline YOLOv11n-OBB by 1.3 percentage points in mAP while reducing model parameters by 17.0% to 2.2M. Deployment verification using a PySide5-based visualization prototype demonstrated a total system-level latency of less than 0.2 s, a duration encompassing the cumulative overhead of image input and output, pre-processing, post-processing including non-maximum suppression, and interface rendering. Furthermore, physical deployment on an NVIDIA Jetson AGX Orin embedded platform utilizing TensorRT optimization achieved an impressive inference speed of 114.5 FPS, corresponding to approximately 8.7 ms per frame. This performance confirms that the model meets the stringent real-time requirements for edge computing in mobile agricultural robotics. The model efficiently and accurately performs oriented detection and harvesting pose estimation for daylily buds, providing critical technical support for the visual perception system of harvesting robots.

Achieving reliable leak detection remains a core challenge in pipeline structural health monitoring (SHM), where diagnostic errors can lead to catastrophic failures or costly false alarms that compromise system safety. Multi-sensor SHM combined with Graph Neural Networks (GNNs) offers a promising paradigm for capturing non-Euclidean sensor topologies, yet its reliance on statistical correlations undermines reliability in industrial environments. Confounding factors such as environmental vibrations and fluid turbulence induce spurious correlations among physically disconnected sensors, degrading diagnostic trustworthiness. To enhance detection reliability, this study proposed the Dynamic Causal Graph Network (DCGN), which leveraged causal inference to disentangle spurious connections and recover genuine physical propagation paths. DCGN employs a Causal Node Encoder (CNE) that extracts state embeddings from sensor signals, while the Causal Skeleton Learner (CSL) and Dynamic Causal Mechanism (DCM) collaboratively reconstruct causal connectivity and model time-varying effects using Partial Correlation Momentary Independence (PCMI) priors. A Transformer-based Causal Inference and Reasoning (CIR) module aggregates long-range causal evidence for robust state inference. Experiments on branched and looped pipeline topologies demonstrate that DCGN achieves 96.32% and 96.60% accuracy respectively, significantly outperforming baselines with statistical significance validated by ANOVA. Under extreme conditions including 5-dB signal-to-noise ratio (SNR) and strong impulse noise, DCGN maintains over 86% accuracy, and achieves above 81% accuracy with only 10% training samples, demonstrating strong robustness for reliable operation. Ablation studies confirm the critical contribution of causal modules to performance gains. Causal graph structure validation confirms that the learned skeleton aligns with the physical pipeline topology, and cross-regime transfer experiments demonstrate accuracy degradation within 2.9% across both topologies, further evidencing the distributional invariance of causal representations. DCGN establishes an intelligent diagnostic framework centered on replacing statistical correlation graphs with causal graphs and substituting spurious connections with genuine physical propagation paths, offering a new paradigm for reliable pipeline safety assurance in complex industrial environments. © 2026 Elsevier Ltd.

Accurate and timely weed identification is fundamental to sustainable crop management, particularly for autonomous agricultural systems operating under strict energy and hardware constraints. While deep learning has significantly advanced image-based weed recognition, most existing models rely on GPU-based inference and therefore cannot be deployed directly in low-power field devices. In this study, we propose a hardware-efficient lightweight convolutional neural network (CNN), named TinyWeedNet, designed specifically for real-time on-device weed identification in precision agriculture. The model integrates multi-scale feature extraction, depthwise separable inverted residual blocks, and compact channel attention to enhance discriminative ability while maintaining a minimal computational footprint. To evaluate its suitability for field deployment, TinyWeedNet was trained and tested on the public DeepWeeds dataset and implemented on an STM32H7 microcontroller via the TinyML workflow. Experimental results demonstrate that the model achieves 97.26% classification accuracy with only 0.48 M parameters, supporting sub-90 ms inference and low energy consumption during fully embedded execution. A comprehensive analysis, including benchmark comparisons, hyperparameter sensitivity tests, and ablation studies, demonstrates that TinyWeedNet provides a good balance of accuracy, speed, and energy efficiency for resource-constrained agricultural platforms. Overall, this work demonstrates a practical pathway for integrating real-time, low-power weed identification into field robots, UAVs, and distributed sensing nodes, contributing to more autonomous and energy-aware weed management strategies in precision agriculture.

Understanding subsurface processes is a central task in geophysical sciences, with direct implications for energy resource development, seismic hazard assessment, and environmental monitoring. Among the available techniques, microseismic monitoring provides critical information on fracture propagation and reservoir stimulation in unconventional oil and gas operations. However, surface-based measurements are often severely degraded by environmental and anthropogenic noise, which compromises the accuracy of event detection, localization, and subsequent interpretation. Conventional denoising methods based on sparse transforms or time–frequency analysis rely on rigid assumptions, while existing U-Net based approaches tend to introduce boundary artifacts and loss of fine signal details. This study proposes MFU-Net, a hybrid denoising framework that integrates multivariate singular spectrum analysis (MSSA) with a frequency-domain-aware U-Net. MSSA reconstructs dominant signal components to suppress random noise and provides physics-informed priors as two-channel inputs. The network further incorporates multi-scale temporal convolution (MSTC) and frequency-domain processing (FDP) modules, enabling joint time–frequency feature learning for robust suppression of low-frequency coupled noise. Experiments on synthetic datasets show that MFU-Net consistently outperforms MSSA and conventional U-Net across a wide input signal-to-noise ratio(SNR) range (–20 dB to 20 dB). At –20 dB, MFU-Net achieves an output SNR improvement to 10.271 dB, while preserving waveform fidelity. Field validation on hydraulic fracturing data from the Sichuan Basin confirms the robust performance of MFU-Net in orangesurface-based microseismic denoising under low-SNR conditions. These results indicate that MFU-Net provides a reliable solution for the studied monitoring scenario, effectively improving data quality and supporting subsequent microseismic analysis and interpretation.

Structural health monitoring of offshore oil and gas pipelines is critical for energy security and environmental protection. Acoustic emission technology has been widely adopted as a non-destructive approach for pipeline valve leakage detection. However, it faces severe challenges in real marine environments. Offshore platform pipelines exhibit strong background noise interference that significantly undermines leakage signal identifiability. This requires distributed sensor deployment to expand monitoring coverage. But installation constraints cause spatially uneven distributions that limit information propagation and create monitoring blind spots. Consequently, collaborative response patterns among multiple sensors are difficult to extract effectively. Traditional fusion methods fail to exploit spatial dependencies between sensors. To address these challenges, this paper proposes a novel graph learning-based end-to-end intelligent monitoring method. The method employs a time-frequency domain graph to suppress noise interference and encode spatial relationships between sensors. Building upon this, a spatial-topological adaptive graph neural network (STAG) captures global collaborative patterns and balances information propagation in non-uniform networks. On datasets simulating real offshore platform pipeline leakage, the proposed method achieved 94.64%-97.74% accuracy and maintained 91.56% under -15 dB noise. Generalization and superiority were validated on public benchmark datasets. Statistical tests confirmed the method significantly outperformed existing approaches. With only 30% training data, accuracy exceeded 88%. Ablation studies validated component effectiveness. STAG required only three layers for high-precision detection and localization. This research provides an effective solution for intelligent offshore pipeline monitoring with significant engineering value.

Intelligent acoustic emission-based pipeline leak detection technology plays a critical role in Internet of Things structural health monitoring for offshore platforms. However, traditional deep networks possess large parameter counts and high computational complexity, making them infeasible for deployment on resource-constrained edge nodes, while lightweight methods universally adopt single-scale feature extraction and cannot simultaneously capture short-duration burst and long-range attenuation characteristics of acoustic emission signals, resulting in insufficient discriminative capability for adjacent valves. To address this, this paper proposes Tiny Leak Sense Net (TinyLSN), a novel lightweight leak localization framework specifically designed for Internet of Things edge nodes. TinyLSN achieves optimal balance between computational efficiency and detection performance through three innovative components we designed including the Inverted Residual Block (IRB), Multi-Scale Dilated Perception Module (MSDPM), and Large Kernel Feed-Forward Network (LK-FFN), which respectively enhance cross-channel interactions, capture multi-scale temporal features, and extract global attenuation patterns. On our self-constructed experimental dataset simulating real offshore platform operational pipeline leakage, TinyLSN achieved detection accuracy of 97.11% to 97.45% and an extremely low false positive rate of 0.27% to 0.32%, significantly outperforming lightweight baseline methods. Validation on publicly available benchmark datasets further confirmed its generalization capability. When deployed on the STM32H7B3I-DK microcontroller, TinyLSN requires only 267.16 KiB Flash memory and achieves 3.417 ms inference latency. Furthermore, TinyLSN maintains over 90% accuracy under strong noise and achieves 94.83% accuracy with only 10% training samples, fully validating its reliability in harsh industrial environments and providing an efficient and feasible solution for offshore platform Internet of Things edge intelligence.

Structural health monitoring (SHM) of aging civil, aerospace, and energy infrastructure increasingly relies on unmanned aerial vehicles (UAVs) equipped with vision sensors for efficient and large-scale inspections. Among these applications, automated crack classification using deep learning models has emerged as a key use case. However, cloud-based inference for such tasks imposes bandwidth, power, connectivity, and privacy costs that are unacceptable for safety-critical assets. To address these limitations, this study presents a fully self-contained Tiny Machine Learning (TinyML) solution that performs onboard crack classification on a milliwatt-level STM32H7 microcontroller (MCU). Using MobileNetV1x0.25 as a baseline, we systematically evaluate an end-to-end measurement processing pipeline, including image capturing, image preprocessing, and model inference on a low-power embedded system. To identify the optimal pipeline configuration, we compare two image preprocessing strategies consisting of a handcrafted grayscale–contrast–denoise–median–binarization method and a greedy algorithm–based composite approach. We further assess four model compression techniques, including 8-bit post-training quantization (PTQ), quantization-aware training (QAT), pruning, and weight clustering, both individually and in combination. The proposed pipeling achieves an F1-score of 0.938, which outperforms the state-of-the-art by 11.4\%. At the same time, it only requires 2.9 MB of RAM and 309 KB of flash memory. The deployed solution has an end-to-end latency of 461.6 ms and an energy cost of 623.16 mJ per inference. For a DJI Mini 4 Pro UAV, continuous operation is estimated to shorten the flight time by merely 1.31 minutes (i.e., 4\%). In contrast, previously reported deployments based on NVIDIA Jetson NX implementations reduce flight time by 8 minutes (i.e., 24\%). This work thus provides a reproducible benchmark and a practical trade-off of accuracy, resource usage, and energy consumption for on-device crack classification in highly resource-constrained, UAV-based SHM scenarios.

Weed classification is a fundamental perception task for agricultural robots and an essential enabler of precision and sustainable farming. Existing solutions often rely on high-power edge computing platforms, which limit long-term autonomous operation in Internet of Things (IoT) environments. Meanwhile, the computational complexity of high-accuracy deep learning models hinders their deployment on resource-constrained micro-controllers (MCUs), a critical component of IoT edge nodes. To address these challenges, this paper proposes a TinyML-enabled energy-efficient IoT framework for on-device weed classification, integrating a novel Three-Dimensional Alignment Knowledge Distillation (TDA-KD) strategy with a lightweight multi-layer dilated-convolution student network. The framework enhances knowledge transfer by jointly aligning (i) individual predictions, (ii) inter-sample correlations, and (iii) class semantics, further strengthened through a multi-temperature calibration mechanism. Experimental results on the DeepWeeds and 4Weeds datasets demonstrate that the proposed student model achieves over 95% classification accuracy with only 240K parameters and 87.51 MFLOPs. The model is successfully deployed on an OpenMV H7 Plus board with an STM32H7 MCU, requiring just 105.68 KB Flash memory and achieving an inference time of 378.3 ms with 510.7 mJ energy consumption per sample. A runtime analysis on the Vitirover horticultural robot shows that, compared with a Jetson Nano-based implementation, the proposed IoT pipeline extends operational time by approximately 30.5%. These results highlight the feasibility of deploying high-accuracy weed classification directly on ultra-low-power IoT devices, thereby significantly enhancing the autonomy, energy efficiency, and scalability of agricultural robots.

Enabling broad use of electromagnetic energy harvesting in wireless sensor screws requires robust systems that work with weak vibrations and varying frequency profiles. This contribution presents an energy management circuit incorporating two cooperating DC-DC converters controlled by self-powered MOSFET switches and a passive voltage multiplier enabling low-voltage start-up. The circuit operates effectively over an acceleration range of 0.07-0.21 g. It consistently harvests energy across a wider frequency range than energy management circuits based on single DC-DC converters. For example, at an acceleration of 0.21 g, the frequency range is 20–30 Hz. Thereby it realizes, e.g. at 25 Hz, an efficiency of 72%. The proposed circuit enables robust energy harvesting in a wide frequency range, supporting wireless sensor operation even under low-vibration conditions typical of industrial predictive maintenance.