Object detection (OD) has become vital for numerous computer vision applications, but deploying it on resource-constrained internet of things (IoT) devices presents a significant challenge. These devices, often powered by energy-efficient microcontrollers, struggle to handle the computational load of deep learning-based object detection (OD) models. This issue is compounded by the rapid proliferation of IoT devices, predicted to surpass 150 billion by 2030. TinyML offers a compelling solution by enabling OD on ultra-low-power devices, paving the way for efficient and real-time OD at the edge. Although numerous survey articles have been published on this topic, they often overlook the optimization challenges associated with deploying OD models in TinyML environments. To address this gap, this survey article provides a detailed analysis of key optimization techniques for deploying OD models on resource-constrained devices. These techniques include quantization, pruning, knowledge distillation, and neural architecture search. Furthermore, we explore both theoretical approaches and practical implementations, bridging the gap between academic research and real-world edge artificial intelligence (AI) deployment. Finally, we compare the key performance indicators (KPIs) of existing OD implementations on microcontroller devices, highlighting the achieved maturity level of these solutions in terms of both prediction accuracy and efficiency. We also provide a public repository to continually track developments in this fast-evolving field: Link.
This letter proposes the design and measurement of a periodic metasurface that achieves anomalous reflection with reduced RCS in a given parasitic direction. A previous study proposed a semianalytical model to predict the RCS behaviour of such a metasurface. However, this first study did not include any experimental exploration to verify the theoretical results. To complete this study, this work presents an experimental validation of the proposed design, with a focus on manufacturing and measurement issues. The synthesis, design specifications, fabrication method and experimental setup are presented and discussed. Measurement results are also examined in detail, highlighting some limitations in metasurface RCS measurements. The proposed metasurface effectively achieves the predicted RCS level reduction in the considered parasitic direction. The agreement between simulation and experimental results demonstrates the accuracy of the modelling and the efficiency of the optimisation procedure.
Current design codes, such as Eurocode 3 part 1-8 and AISC Design Guide 1, provide limited guidance on steel-to-concrete corner-bolted end-plate connections under uniaxial and biaxial bending moments. In previous studies on this bolt configuration, failures primarily occurred in the welds without a clear ductile failure mode. To address this gap, the authors conducted an experimental campaign involving 12 configurations designed to observe ductile failures caused by anchor bolt rupture with end-plate yielding. The study investigated three endplate thicknesses, each under four bending moment orientations: in-plane, out-of-plane, and biaxial bending moments with angle orientations of 30 degrees and 45 degrees. The connections were densely instrumented to assess the anchor bolt force evolutions and thus plastic redistribution between the different anchor bolts. Findings revealed that end-plate thickness substantially influences the global behaviour of the connection, including initial rotational stiffness and bending resistance. Additionally, the anchor bolt force distribution was strongly impacted by both the bending moment orientation and the end-plate thickness. These experimental observations were completed by results of a finite element model developed with ANSYS using solid/contact elements and considering an elasto-plastic behaviour. This model provided deeper insights into prying effects, evolution of the centre of compression, and contact pressure distributions in the tensile and compressive areas that could not be measured experimentally.
In this paper, we introduce a reliable positioning method for a smart wheelchair by using ultra-wideband (UWB) technology. This method provides confidence domains of the pose by assuming bounded measurement errors and proprioceptive information, without any assumptions of independence. Exploiting interval analysis and constraint propagation techniques, we characterize the sets of all feasible poses that are consistent with the measurements. Our method has been validated through experiments with a smart power wheelchair equipped with UWB sensors in realistic conditions. The results demonstrate that our approach consistently provides guaranteed uncertainty domains with 100% integrity across the tested dataset, even when faced with inconsistent measurements. In addition, we compare our interval-based method with M-estimator approaches, and show that while achieving slightly worse positioning accuracy, our method offers superior consistency.
