Embedded Development Platform Research Articles

TinyML Model for Classifying Hazardous Volatile Organic Compounds Using Low-Power Embedded Edge Sensors: Perfecting Factory 5.0 Using Edge AI

In this letter, we present an intelligent chemical sniffer, capable of detecting hazardous volatile organic compounds (VOCs) using artificial intelligence deployed at the edge. The proposed solution utilizes a multichannel metal–oxide–semiconductor gas sensor for generating unique signature responses from the organic compounds. These signature responses are then used for training a quantized neural network model in the cloud and then deployed onto an embedded development platform for edge inference. In this approach, we utilized TensorFlow Lite for microcontrollers, an edge computing framework (developed by Google) to quantize a 32-bit floating point precision neural network model into an 8-bit integer precision model for deployment on low-power and low-memory footprint embedded edge devices. Our quantized model was able to successfully classify all the six VOCs (i.e., Xylene, Hexane, Acetone, Toluene, Methanol, and Butanol) with accuracies of 99.8 and 100% on the validation and test datasets, respectively. More than 50% reduction in on-device RAM and Flash memory usage were measured for the 8-bit quantized model when compared with the equivalently performing 32-bit floating point model for relatively same inference speed on the edge sensing device. The proposed solution paves the way for the mass-scale deployment of such intelligent chemical sensing solutions in the next-generation automated smart factories.

A novel method based on neural architecture search for Diptera insect classification on embedded devices

Diptera insects have the characteristics of spreading diseases and destroying forests. There are similarities among different species, which makes it difficult to identify a Diptera insect. Most traditional convolutional neural networks have large parameters and high recognition latency. Therefore, they are not suitable for deploying models on embedded devices for classification and recognition. This paper proposes an improved neural architecture based on differentiable search method. First, we designed a network search cell by adding the feature output of the previous layer to each search cell. Second, we added the attention module to the search space to expand the searchable range. At the same time, we used methods such as model quantization and limiting the ReLU function to the ReLU6 function to reduce computer resource consumption. Finally, the network model was transplanted to the NVIDIA Jetson Xavier NX embedded development platform to verify the network performance so that the neural architecture search could be organically combined with the embedded development platform. The experimental results show that the designed neural architecture achieves 98.9% accuracy on the Diptera insect dataset with a latency of 8.4 ms. It has important practical significance for the recognition of Diptera insects in embedded devices.

Development of Multi-Function Vehicle Intelligent System Based on ARM

At present, with the progress of science and technology in the world and the improvement of people’s living standards, the number of cars is increasing rapidly. People’s demand for tourism has become a new bright spot, and the demand for transportation system services is getting higher and higher. Taxi has become an indispensable means of travel for people. Based on the original functions of taxi management system, this paper presents a multi-functional taxi intelligent system, which integrates security alarm, traffic record, wireless communication and remote monitoring, based on Cortex-A8 embedded development platform and Linux kernel programming. With the advantage of embedded system, the algorithm is optimized by using the idea of communication between threads. The communication between modules is realized, and the reliability and stability of the system are effectively improved. For remote video surveillance system, through the CGI process writing and the design of human-computer interaction interface, the system realizes the web surveillance of alarm settings, driving records and real-time monitoring of vehicle data, which further enhances the functions of the system. The test results show that the system has good traffic safety functions, and the data exchange between the web monitoring system and the on-board system is smooth, which meets the expected design requirements.

Real-Time Embedded Implementation of Improved Object Detector for Resource-Constrained Devices

Artificial intelligence (A.I.) has revolutionised a wide range of human activities, including the accelerated development of autonomous vehicles. Self-navigating delivery robots are recent trends in A.I. applications such as multitarget object detection, image classification, and segmentation to tackle sociotechnical challenges, including the development of autonomous driving vehicles, surveillance systems, intelligent transportation, and smart traffic monitoring systems. In recent years, object detection and its deployment on embedded edge devices have seen a rise in interest compared to other perception tasks. Embedded edge devices have limited computing power, which impedes the deployment of efficient detection algorithms in resource-constrained environments. To improve on-board computational latency, edge devices often sacrifice performance, creating the need for highly efficient A.I. models. This research examines existing loss metrics and their weaknesses, and proposes an improved loss metric that can address the bounding box regression problem. Enhanced metrics were implemented in an ultraefficient YOLOv5 network and tested on the targeted datasets. The latest version of the PyTorch framework was incorporated in model development. The model was further deployed using the ROS 2 framework running on NVIDIA Jetson Xavier NX, an embedded development platform, to conduct the experiment in real time.

DeepFake Video Detection through Facial Sparse Optical Flow based Light CNN

DeepFake detection has become an attractive research topic with tremendous growth of interests recently. However, existing DeepFake detection studies spare no effort to improve accuracy or Area Under Curve metric, regardless of computing costs. In this work, the tradeoff between result accuracy and computing resources is taken into consideration. A facial sparse optical flow method is proposed to extract spatio-temporal features representing facial expression incoherence, which helps to distinguish fake videos and real videos. The features fed into a light CNN model are highly compact and low-dimensional. The proposed method has an amazing small amount of parameters with high training speed and low usage of GPU memory. The low resource requirement makes it possible to port to embedded development platform.

Performance Evaluation of Symmetric Cryptographic Algorithms in resource constrained hardware for Wireless Sensor Networks

Wireless sensor networks (WSN) allow to exchange information and to take immediate and remote actions in natural, industrial, military or domestic environment systems. These networks are vulnerable to cyber-attacks, hence, they are vulnerable to being intercepted, interrupted or modified. However, for the last twenty years, the amount of information exchanged through communication networks around the world has considerably increased and thus its vulnerability. As a result, encryption algorithms are fundamental to protect information. In this context, security, performance and energy consumption become a paramount design factor for an engineer when designing and implementing WSNs. It is not an easy task to reach an optimal balance between these factors. In this paper, we evaluate three well-known symmetric encryption algorithms in an embedded development platform: the ARDUINO MEGA 2560. Our experiment measures the algorithms encryption average execution time and energy consumption. Additionally, we measure the impact of the channels of a small wireless sensor network with two nodes. The evaluation demonstrates the feasibility of implementing cryptographic algorithms in devices with limited resources such as memory, computing power and life span (energy consumption). Furthermore, the low impact on the network channels when comparing unencrypted communication vs encrypted communication evidence the potential of symmetric encryption algorithms.

Multicore development environment for embedded processor in arduino IDE

Internet of things (IoT) technology has found more applications that require complex computation while still preserving power. Embedded processors as the core of the IoT system approaches the need for computation by employing a parallel processor system, namely MPSoC. While various MPSoCs hardware is widely available, there is limited software support form of user-friendly libraries and development platform. There is a need for such a platform to facilitate both the study and development of parallel embedded software. arduino as the widely used embedded development platform is yet to officially support multicore programming. This work proposes an arduino-based development environment that supports multicore programming while maintaining arduino’s simple program structure, targeted at specific low-power MPSoC, the RUMPS401. The environment is fully functional, and while it targets only specific MPSoC, the proposed environment can easily be adopted to other MPSoCs with similar structures with minimal modification.

Secure IoT solution for wearable health care applications, case study Electric Imp development platform

SummaryInternet expansion affects currently diverse entities on the physical world and daily life. The democratization of the technology called Internet of Things (IoT) brings back many problems associated with the security of the transmitted data, autonomy, and ease of use, with the central constraint of gathering momentum. The use of an embedded development platform for IoT operating through Wi‐Fi and cloud technology is in our case the solution that fits the internet model objects for wearable health care application. This work involves the study and implementation of wearable IoT secure solution dedicated to health care devoted to old and dependent people. The proposed hardware, powered by solar cells, operate through smart sleep and wake up events. The relevant part of the developed IoT operating software exploits the multiagent behavior. Case study, based on the exploitation of the Electric Imp platform, is proposed.

Virtual duplication and mapping prefetching for emerging storage primitives in NAND flash memory storage systems

NAND flash memory has become the mainstream storage medium for both enterprise high performance computers and embedded systems. However, over the past several decades, the storage primitives that access secondary storage have remained unchanged, forcing NAND flash memory to serve merely as a block device like hard disk drive. Recently, several emerging storage primitives have been presented to explore the potential value of non-volatile memory devices. Although these primitives can significantly boost the access performance by providing virtual to logical address mappings, they still suffer from large RAM footprint to maintain the address mapping table and require further support for update operations.This paper presents ESP to optimize E merging S torage P rimitives with virtualization for flash memory storage systems. We propose two optimization strategies, virtual duplication and mapping prefetching to solve the critical issues in existing emerging storage primitives. The objective is to reduce unnecessary flash memory accesses and keep RAM footprint of address mapping table well under control. We have evaluated ESP on an embedded development platform. Experimental results show that ESP can significantly improve the write/read performance and reduce over 30% of garbage collection operations.

Energy Conservation in a Smart Home via an Embedded Platform

Home automation is a popular field of application which combines the concepts of embedded systems and pervasive computing to control a living space for modern times. A smart home system aims to free up the user to carry out his or her daily tasks without worrying on the trivial aspects of personal home management, while at the same time providing the user with more control of the home than ever before. The implementation comprises of a large number of sensors which can be used to control or monitor objects distributed in three-dimensional space around the house. The sensors can be specialized in measuring temperature, humidity, pressure, light, noise, dust air, and upon intelligently computed triggers, compel the system to perform a specific task, or a set of tasks. In this project, a solution to transform a normal house into a smart, autonomous house in order to reduce the energy consumption of the household is proposed. This can be realized with the help of wired sensor networks where we control and interface with electronic appliances via a Raspberry Pi, an embedded development platform. The embedded platform can be controlled remotely to turn switches on and off, and eventually aims to become as autonomous as possible via the help of a Bayesian Network. The base workings of the project are aimed at being as scalable as possible, to be able to potentially fit into industrial as well as office environments in the future.

Design of Remote Monitoring System Based on GoAhead Web Server

This paper covers the current situation of remote monitoring system, combined with the basic features and running environment of GoAhead web server, Put forward the solution to integrate GoAhead web server, Internet, embedded development platform and Linux operation system, design a remote monitoring system. By which the user can monitor voltage and control the stepper motor’s running in real time. The application results show that GoAhead web server reacts rapidly manipulates data quickly. It’s feasible to put into place and to achieve satisfied effect for the management and control of the embedded product.

Embedded PWM DC Motor Control Cystem Based on ARM

This paper introduces an ARM-based embedded PWM DC motor control system. The system uses a variable voltage control; the entire system uses the speed and current closed-loop control structure. The main circuit using the structure of high-power system GTR switching devices and the bipolar H-bridge circuit for the power amplifier circuit, to complete irreversible speed regulation also finished reversible speed regulation. PWM modulation part of the ARM embedded development platform, using the C language programming to control the timer to generate two complementary width adjustable rectangular waves. GTR H circuit is controlled by adjusting the width of the waveform from two ways off time in order to achieve the purpose of speed regulation. It increased the flexibility and accuracy of the system while simplifying the process of generation of the PWM pulse.

Image Capture and Display Based on Embedded Linux

Abstract For the requirement of building a highly reliable communication system, SpaceWire was selected in the integrated electronic system. There was a need to test the performance of SpaceWire. As part of the testing work, the goal of this paper is to transmit image data from CMOS camera through SpaceWire and display real-time images on the graphical user interface with Qt in the embedded development platform of Linux & ARM. A point-to-point mode of transmission was chosen; the running result showed the two communication ends basically reach a consensus picture in succession. It suggests that the SpaceWire can transmit the data reliably.