Abstract
This research aimed to develop a high-portability and multipurpose Microcontroller Trainer by employing Research and Development (R&D) with the ADDIE (Analysis, Des ign, Development, ...Implementation, Evaluation) model. It focuses on the development s tage (the design aspect) and the testing stage of the trainer’s performance. The results of this study indicate that the developed product of microcontroller trainer can be categorized as “very feasible” to use with a final score of 72. Meanwhile, for the aspect of trainer performance, it is also declared that the developed product is “very feasible” in terms of function and usefulness that can be used as a multipurpose media.
Depth estimation is crucial in several computer vision applications, and a recent trend in this field aims at inferring such a cue from a single camera. Unfortunately, despite the compelling results ...achieved, state-of-the-art monocular depth estimation methods are computationally demanding, thus precluding their practical deployment in several application contexts characterized by low-power constraints. Therefore, in this paper, we propose a lightweight Convolutional Neural Network based on a shallow pyramidal architecture, referred to as <inline-formula> <tex-math notation="LaTeX">\mu </tex-math></inline-formula>PyD-Net, enabling monocular depth estimation on microcontrollers. The network is trained in a peculiar self-supervised manner leveraging proxy labels obtained through a traditional stereo algorithm. Moreover, we propose optimization strategies aimed at performing computations with quantized 8-bit data and map the high-level description of the network to low-level layers optimized for the target microcontroller architecture. Exhaustive experimental results on standard datasets and an in-depth evaluation with a device belonging to the popular Arm Cortex-M family confirm that obtaining sufficiently accurate monocular depth estimation on microcontrollers is feasible. To the best of our knowledge, our proposal is the first one enabling such remarkable achievement, paving the way for the deployment of monocular depth cues onto the tiny end-nodes of distributed sensor networks.
One of the main challenges to design a proper protection scheme for a microgrid is the dynamic characteristics in the architecture for the microgrid. The relay settings need to change with every ...configuration in the system to properly isolate a fault. This article proposes a centralized control approach using a cosimulation platform to monitor operational modes and multiple microgrid topologies to find fault and clear them optimally. The simulation software, embedded microcontrollers, and a real communication architecture are collectively used to perform the protection scheme platform. The data distribution service middleware is used to link between the hardware and software environments. The proposed technique helps to identify the shortest path from the source to the fault location. The proposed main control center is responsible for adjusting the relay settings based on the active topology and the shortest path to the fault. This technique is investigated on IEEE 14 bus microgrid system for all the possible fault paths. The proposed scheme can clear the fault by isolating the minimum part of the system and improving the endurance of the power to supply the different loads in the system.
Proton beam testing was performed on several microcontrollers using proton energies between 2.4 MeV and 57.6 MeV. A remotely operated robotic testing platform was developed to improve testing ...throughput by minimizing the need for human intervention in the radiation vault. Sensitivity to SEU, SEL, and SEFI faults was determined. Using the measured SEU cross-section data along with NASA's radiation belt model, AP8, in-orbit SEU rates were determined for an ISS orbit. Testing was also done while varying the incidence angle of the proton beam which shows that the rate of SEL and SEFI faults increased significantly with shallower incidence angle.
Systems using microcontrollers make the interface of a transducer with the digital world much easier. They build smart, lucid, compact, cheap, and less power consuming electronic interfaces. Limited ...acquisition time is critical for many industrial applications. Single-element resistive sensors are used on a large scale. An attempt has been made to reduce the acquisition time of the measuring system compared to the existing work which converts the change in sensor resistance to the change in time with three charge-discharge cycles. Efforts have been made in this paper to reduce this to two charge-discharge cycles. The proposed work also compensates the errors due to lead wire resistance, the change in lead wire resistance, port pin resistances, and varying ambient temperature with reduced acquisition time. The mean of absolute errors, standard deviation, integral squared error (ISE), nonlinearity, and hysteresis of the proposed method have been computed. Simulation and experimentation study of the proposed system provided encouraging results. The mean of absolute errors and standard deviation of the proposed method in the simulation were 0.07 and 0.08 <inline-formula> <tex-math notation="LaTeX">\Omega </tex-math></inline-formula>, respectively. Furthermore, during experimentation, the values of the mean of absolute errors and standard deviation were 0.11 and 0.11 <inline-formula> <tex-math notation="LaTeX">\Omega </tex-math></inline-formula>, respectively. ISE was found to be 0.09 and 0.19 in simulation and experimentation, respectively. The proposed two-cycle method has an acquisition time of only 60% of the three-cycle method with no reduction in performance.
Outdoor acoustic event detection is an exciting research field but challenged by the need for complex algorithms and deep learning techniques, typically requiring many computational, memory, and ...energy resources. These challenges discourage IoT implementations, where an efficient use of resources is required. However, current embedded technologies and microcontrollers have increased their capabilities without penalizing energy efficiency. This paper addresses the application of sound event detection at the very edge, by optimizing deep learning techniques on resource-constrained embedded platforms for the IoT. The contribution is two-fold: firstly, a two-stage student-teacher approach is presented to make state-of-the-art neural networks for sound event detection fit on current microcontrollers; secondly, we test our approach on an ARM Cortex M4, particularly focusing on issues related to 8-bits quantization. Our embedded implementation can achieve 68% accuracy in recognition on Urbansound8k, not far from state-of-the-art performance, with an inference time of 125 ms for each second of the audio stream, and power consumption of 5.5 mW in just 34.3 kB of RAM.