Convolutional neural networks (CNNs) have become the dominant neural network architecture for solving many state-of-the-art (SOA) visual processing tasks. Even though graphical processing units are most often used in training and deploying CNNs, their power efficiency is less than 10 GOp/s/W for single-frame runtime inference. We propose a flexible and efficient CNN accelerator architecture called NullHop that implements SOA CNNs useful for low-power and low-latency application scenarios. NullHop exploits the sparsity of neuron activations in CNNs to accelerate the computation and reduce memory requirements. The flexible architecture allows high utilization of available computing resources across kernel sizes ranging from 1x1 to 7x7. NullHop can process up to 128 input and 128 output feature maps per layer in a single pass. We implemented the proposed architecture on a Xilinx Zynq field-programmable gate array (FPGA) platform and presented the results showing how our implementation reduces external memory transfers and compute time in five different CNNs ranging from small ones up to the widely known large VGG16 and VGG19 CNNs. Postsynthesis simulations using Mentor Modelsim in a 28-nm process with a clock frequency of 500 MHz show that the VGG19 network achieves over 450 GOp/s. By exploiting sparsity, NullHop achieves an efficiency of 368%, maintains over 98% utilization of the multiply-accumulate units, and achieves a power efficiency of over 3 TOp/s/W in a core area of 6.3 mm₂. As further proof of NullHop's usability, we interfaced its FPGA implementation with a neuromorphic event camera for real-time interactive demonstrations.

Low-latency, low-power portable recurrent neural network (RNN) accelerators offer powerful inference capabilities for real-time applications such as IoT, robotics, and human-machine interaction. We propose a lightweight Gated Recurrent Unit (GRU)-based RNN accelerator called EdgeDRNN that is optimized for low-latency edge RNN inference with batch size of 1. EdgeDRNN adopts the spiking neural network inspired delta network algorithm to exploit temporal sparsity in RNNs. Weights are stored in inexpensive DRAM which enables EdgeDRNN to compute large multi-layer RNNs on the most inexpensive FPGA. The sparse updates reduce DRAM weight memory access by a factor of up to 10x and the delta can be varied dynamically to trade-off between latency and accuracy. EdgeDRNN updates a 5 million parameter 2-layer GRU-RNN in about 0.5 ms. It achieves latency comparable with a 92 W Nvidia 1080 GPU. It outperforms NVIDIA Jetson Nano, Jetson TX2 and Intel Neural Compute Stick 2 in latency by 5X. For a batch size of 1, EdgeDRNN achieves a mean effective throughput of 20.2 GOp/s and a wall plug power efficiency that is over 4X higher than the commercial edge AI platforms.

The change from face-to-face work to teleworking caused by the pandemic has induced multiple workers to spend more time than usual in front of a computer; in addition, the sudden installation of workstations in homes means that not all of them meet the necessary characteristics for the worker to be able to position himself/herself comfortably with the correct posture in front of their computer. Furthermore, from the point of view of the medical personnel in charge of occupational risk prevention, an automated tool able to quantify the degree of incorrectness of a postural habit in a worker is needed. For this purpose, in this work, a system based on the postural detection of the worker is designed, implemented and tested, using a specialized hardware system that processes video in real time through convolutional neural networks. This system is capable of detecting the posture of the neck, shoulders and arms, providing recommendations to the worker in order to prevent possible health problems, due to poor posture. The results of the proposed system show that this video processing can be carried out in real time (up to 25 processed frames/sec) with a low power consumption (less than 10 watts) using specialized hardware, obtaining an accuracy of over 80% in terms of the pattern detected.

The study and monitoring of wildlife has always been a subject of great interest. Studying the behavior of wild animals is a difficult task due to the difficulties of tracking and classifying their actions. Nowadays, technology allows designing low-cost systems that make these tasks easier to carry out, and some of these systems produce good results; however, none of them obtains a high-accuracy classification because of the lack of information. Doñana National Park is a very rich environment with various endangered animal species. Thereby, this park requires a more accurate and efficient system of monitoring to act quickly against animal behaviors that may endanger certain species. In this letter, we propose a hierarchical, wireless sensor network installed in this park, to collect information about animals' behaviors using intelligent devices placed on them which contain a neural network implementation to classify their behavior based on sensory information. Once a behavior is detected, the network redirects this information to an external database for further treatment. This solution reduces power consumption and facilitates animals' behavior monitoring for biologists.

As localization represents the main core of various wireless sensor network applications, several localization algorithms have been suggested in wireless sensor network research. In this article, we put forward an iterative bounding box algorithm enhanced by a Kalman filter to refine the unknown node’s estimated position. In fact, several research efforts are currently in progress to extend the 2D positioning algorithm in WSNs to 3D that reflects reality and the most practical applications. Subsequently, we replace a large number of GPS-equipped anchors with a single mobile anchor. In our studies, we consider the type of range-free sensor network exploiting the wireless sensors connectivity. We assess the performance of our algorithm using exhaustive experiments on several isotropic and anisotropic topologies. Our proposed algorithm can fulfill the joint goals of algorithm transparency and accuracy for various scenarios by evaluating parameters such as localization accuracy whilst changing other simulation parameters such as the effect of communication range, mobile anchor node position and sensor node deployment topology. It has been proven by the results of the experiments that the proposed algorithm effectively reduces the location error without requiring more equipment or increasing the communication cost.