This paper presents an Internet of Things (IoT) application that utilizes an AI classifier for fast-object detection using the frame difference method. This method, with its shorter duration, is the most efficient and suitable for fast-object detection in IoT systems, which require energy-efficient applications compared to end-to-end methods. We have implemented this technique on three edge devices: AMD AlveoTMU50, Jetson Orin Nano, and Hailo-8TMAI Accelerator, and four models with artificial neural networks and transformer models. We examined various classes, including birds, cars, trains, and airplanes. Using the frame difference method, the MobileNet model consistently has high accuracy, low latency, and is highly energy-efficient. YOLOX consistently shows the lowest accuracy, lowest latency, and lowest efficiency. The experimental results show that the proposed algorithm has improved the average accuracy gain by 28.314%, the average efficiency gain by 3.6 times, and the average latency reduction by 39.305% compared to the end-to-end method. Of all these classes, the faster objects are trains and airplanes. Experiments show that the accuracy percentage for trains and airplanes is lower than other categories. So, in tasks that require fast detection and accurate results, end-to-end methods can be a disaster because they cannot handle fast object detection. To improve computational efficiency, we designed our proposed method as a lightweight detection algorithm. It is well suited for applications in IoT systems, especially those that require fast-moving object detection and higher accuracy.

In the field of computer vision technology, deep learning of image processing has become an emerging research area. The semantic segmentation of an image is among the utmost essential and significant tasks in image-processing research, offering a wide range of application fields such as autonomous driving systems, medical diagnosis, surveillance security, etc. Thus far, many studies have suggested and developed neural network modules in deep learning. To the best of our knowledge, all existing neural networks for semantic segmentation have large parameter sizes and it is therefore unfeasible to implement those architectures in low-power and memory-limited embedded platforms such as FPGAs. Building an embedded platform with that architecture is possible after reducing the parameter size without affecting the module’s architecture. The quantization technique lowers the precision of the neural network parameters while mostly keeping the accuracy. In this paper, we propose a quantization algorithm for a semantic segmentation deep learning architecture, which reduces the parameter size by four to eight times with a negligible accuracy abatement. As long as the parameter size is reduced, the deep learning architecture is improved in terms of required storage, computational speed, and power efficiency.

In this article, a new Binary Fully Convolutional Neural Network (B-FCN) based on Taguchi method sub-optimization for the segmentation of robotic floor regions, which can precisely distinguish floor regions in complex indoor environments is proposed. This methodology is quite suitable for robot vision in an embedded platform and the segmentation accuracy is up to $84.80\%$ on average. A total of 6000 training datasets were used to improve the accuracy and reach convergence. On the other hand, to reach real-time computation, a PYNQ FPGA platform with heterogeneous computing acceleration was used to accelerate the proposed B-FCN architecture. Overall, robots would benefit from better navigation and route planning in our approach. The FPGA synthesis of our binarization method indicates an efficient reduction in the BRAM size to 0.5–1% and also GOPS/W is sufficiently high. Notably, the proposed faster architecture is ideal for low power embedded devices that need to solve the shortest path problem, path searching, and motion planning.

In recent years, plant electrophysiology has become a viable measurement tool for describing plants’ thinking properties. In this paper, we propose a new electrophysiological sensor as an autonomous device, which can transfer a plant’s potential electrical data into self-awareness information (a.k.a, phytosensing). Also, this device holds the potential to explore the electrophysiology of plants and its utilization in plant factories, along with its integration with intelligent monitoring and AI technology to evaluate a plant’s cognitive capabilities. Plant sensing can measure certain aspects of a plant’s natural internal state. The proposed sensor circuit design is based on two-stage adjustable processing layers in electrophysiology. The major contribution of this paper is to propose an electrophysiological sensor module that provides plant thinking data concerning environmental factors with dedicated surrounding conditions. Furthermore, the collected data were analyzed through a K-Nearest Neighbor algorithm, which provides a feasible and less complex way to implement the supervised machine learning algorithm. The experimental result shows that each dedicated sensing condition, such as light, heat, humidity, CO ₂ , etc., with the external stimuli through the proposed plant’s electrophysiology sensor module, can evaluate a plant’s thinking behavior. The proposed sensor module has a wide range of gains in circuits with less required complexity. Furthermore, it is suitable for larger-scale data measurement for multiple plant nodes up to 16-channels.

In this paper, a new algorithm that hardware acceleration has enabled to estimate floor regions in a single image for unmanned ground vehicle (UGV) robots is presented. In general, the cluttered indoor surroundings patterned floors, shadows, and reflections make it extremely difficult to identify floor regions. The proposed algorithm combines extracting surface texture characteristics with specific geometric areas to determine object boundaries and then uses SVM classification to distinguish between the floor and non‐floor regions. To achieve real‐time performance, the implementation of the proposed algorithm on an SoC FPGA embedded platform gives a heterogeneous hardware acceleration methodology. In experimental results, a public MIT scene dataset and indoor database were used to verify detection accuracy. Compared to other research, the proposed algorithm accuracy can reach up to 94.72% on average without the assistance of any other physical sensors, such as RGB‐D or laser ranger sensors.

One of the important tasks in the area of computer vision is semantic segmentation. The implementation of a semantic segmentation system in an embedded platform is a fruitful idea. But due to the limitations of embedded ability, it becomes a tough task. In this article, we proposed a novel and practical architecture i.e. quantized deep convolutional neural network for image segmentation (Q-SegNet). This architecture will be implemented on an FPGA device, which allows reducing the parameter size of the original architecture. Hence the required power also reduces. Thus, this paper proposed a high performance deep learning processor unit (DPU) based accelerator for Semantic segmentation neural network. This research is quite suitable for robot vision in an embedded platform and the segmentation accuracy is up to 89.60% on average. Notably, the proposed faster architecture is ideal for low power embedded devices that need to solve the shortest path problem, path searching, and motion planning, in the ADAS and Robot.