The growing demand for food in terms of quality and quantity has increased the need for industrialization and intensification in the agriculture field. Internet of Things (IoT) is a highly promising technology that is offering many innovative solutions to modernize the agriculture sector. Research institutions and scientific groups are continuously working to deliver solutions and products using IoT to address different domains of agriculture. This paper presents a systematic literature review (SLR) by conducting a survey of IoT technologies and their current utilization in different application domains of the agriculture sector. The underlying SLR has been compiled by reviewing research articles published in well-reputed venues between 2006 and 2019. A total of 67 papers were carefully selected through a systematic process and classified accordingly. The primary objective of this systematic study is the collection of all relevant research on IoT agricultural applications, sensors/devices, communication protocols, and network types. Furthermore, it also discusses the main issues and challenges that are being investigated in the field of agriculture. Moreover, an IoT agriculture framework has been presented that contextualizes the representation of a wide range of current solutions in the field of agriculture. Similarly, country policies for IoT-based agriculture have also been presented. Lastly, open issues and challenges have been presented to provide the researchers promising future directions in the domain of IoT agriculture.

Role of IoT Technology in Agriculture: A Systematic Literature Review

Deep learning requires a large amount of data to perform well. However, the field of medical image analysis suffers from a lack of sufficient data for training deep learning models. Moreover, medical images require manual labeling, usually provided by human annotators coming from various backgrounds. More importantly, the annotation process is time-consuming, expensive, and prone to errors. Transfer learning was introduced to reduce the need for the annotation process by transferring the deep learning models with knowledge from a previous task and then by fine-tuning them on a relatively small dataset of the current task. Most of the methods of medical image classification employ transfer learning from pretrained models, e.g., ImageNet, which has been proven to be ineffective. This is due to the mismatch in learned features between the natural image, e.g., ImageNet, and medical images. Additionally, it results in the utilization of deeply elaborated models. In this paper, we propose a novel transfer learning approach to overcome the previous drawbacks by means of training the deep learning model on large unlabeled medical image datasets and by next transferring the knowledge to train the deep learning model on the small amount of labeled medical images. Additionally, we propose a new deep convolutional neural network (DCNN) model that combines recent advancements in the field. We conducted several experiments on two challenging medical imaging scenarios dealing with skin and breast cancer classification tasks. According to the reported results, it has been empirically proven that the proposed approach can significantly improve the performance of both classification scenarios. In terms of skin cancer, the proposed model achieved an F1-score value of 89.09% when trained from scratch and 98.53% with the proposed approach. Secondly, it achieved an accuracy value of 85.29% and 97.51%, respectively, when trained from scratch and using the proposed approach in the case of the breast cancer scenario. Finally, we concluded that our method can possibly be applied to many medical imaging problems in which a substantial amount of unlabeled image data is available and the labeled image data is limited. Moreover, it can be utilized to improve the performance of medical imaging tasks in the same domain. To do so, we used the pretrained skin cancer model to train on feet skin to classify them into two classes-either normal or abnormal (diabetic foot ulcer (DFU)). It achieved an F1-score value of 86.0% when trained from scratch, 96.25% using transfer learning, and 99.25% using double-transfer learning.

https://www.mdpi.com/2072-6694/13/7/1590/pdf

Novel Transfer Learning Approach for Medical Imaging with Limited Labeled Data.

The Atmosphere-Space Interactions Monitor (ASIM) is an instrument suite on the International Space Station (ISS) for measurements of lightning, Transient Luminous Events (TLEs) and Terrestrial Gamma-ray Flashes (TGFs). Developed in the framework of the European Space Agency (ESA), it was launched April 2, 2018 on the SpaceX CRS-14 flight to the ISS. ASIM was mounted on an external platform of ESA’s Columbus module eleven days later and is planned to take measurements during minimum 3 years. The instruments are an x- and gamma-ray monitor measuring photons from 15 keV to 20 MeV, and an array of three photometers and two cameras measuring in bands at: 180–250 nm, 337 nm and 777.4 nm. Additional objectives that can be addressed with the instruments relate to space physics like aurorae and meteors, and to Earth observation such as dust- and aerosol effects on cloud electrification. The paper describes the scientific objectives of the ASIM mission, the instruments, the mission architecture and the international collaboration supported by the ASIM Science Data Centre. ASIM is the first space mission with a comprehensive suite of instruments designed to measure TLEs and TGFs. Two companion papers describe the instruments in more detail (Ostgaard et al. in Space Sci. Rev., 2019; Chanrion et al. in Space Sci. Rev., 2019).

/pdf/the-asim-mission-on-the-international-space-station-30a43our20.pdf

The ASIM Mission on the International Space Station

The Kingdom of Saudi Arabia is known for its extreme climate where temperatures can exceed 50 °C, especially in summer. Improving agricultural production can only be achieved using innovative environmentally suitable solutions and modern agricultural technologies. Using Internet of Things (IoT) technologies in greenhouse farming allows reduction of the immediate impact of external climatic conditions. In this paper, a highly scalable intelligent system controlling, and monitoring greenhouse temperature using IoT technologies is introduced. The first objective of this system is to monitor the greenhouse environment and control the internal temperature to reduce consumed energy while maintaining good conditions that improve productivity. A Petri Nets (PN) model is used to achieve both monitoring of the greenhouse environment and generating the suitable reference temperature which is sent later to a temperature regulation block. The second objective is to provide an Energy-Efficient (EE) scalable system design that handles massive amounts of IoT big data captured from sensors using a dynamic graph data model to be used for future analysis and prediction of production, crop growth rate, energy consumption and other related issues. The design tries to organize various possible unstructured formats of raw data, collected from different kinds of IoT devices, unified and technology-independent fashion using the benefit of model transformations and model-driven architecture to transform data in structured form.

/pdf/an-intelligent-iot-based-system-design-for-controlling-and-2cen08gdiw.pdf

An Intelligent IoT-Based System Design for Controlling and Monitoring Greenhouse Temperature

The rapid growth of the Internet of Things (IoT) has accelerated strong interests in the development of low-power wireless sensors. Today, wireless sensors are integrated within IoT systems to gather information in a reliable and practical manner to monitor processes and control activities in areas such as transportation, energy, civil infrastructure, smart buildings, environment monitoring, healthcare, defense, manufacturing, and production. The long-term and self-sustainable operation of these IoT devices must be considered early on when they are designed and implemented. Traditionally, wireless sensors have often been powered by batteries, which, despite allowing low overall system costs, can negatively impact the lifespan and the performance of the entire network they are used in. Energy Harvesting (EH) technology is a promising environment-friendly solution that extends the lifetime of these sensors, and, in some cases completely replaces the use of battery power. In addition, energy harvesting offers economic and practical advantages through the optimal use of energy, and the provisioning of lower network maintenance costs. We review recent advances in energy harvesting techniques for IoT. We demonstrate two energy harvesting techniques using case studies. Finally, we discuss some future research challenges that must be addressed to enable the large-scale deployment of energy harvesting solutions for IoT environments.

/pdf/energy-harvesting-techniques-for-internet-of-things-iot-3gwdaz1e0d.pdf

Energy Harvesting Techniques for Internet of Things (IoT)

Reward criteria impact on the performance of reinforcement learning agent for autonomous navigation

Bollard is a vital component of mooring system. It is the anchor point for mooring ropes to be fixed in order to secure the vessel or ship. An algorithm that translates the segmented mask of bollard output from masked R-CNN along with bounding box and associated class probability to its corresponding edge coordinate and finally to the single reference point for efficient detection and classification of bollard towards autonomous mooring is presented. At first stage, Mask R-CNN framework is trained with custom built bollard. The model obtained from the training is inferred with real data resulting in instance segment of bollard. The segmented mask obtained contains relatively large amount of the data points representing the whole area of bollard, which typically is not desirable. In order to precisely localize the bollard with one reference co-ordinate, the proposed algorithm is applied to segmented mask. Firstly, it translates the segmented mask to only four co-ordinate points, where each point correspond to the edge of bollard. Further, from the edges, the reference point is estimated. This causes significant reduction in point of interest (POI) and has potential to reduce the error encountered during pose estimation of the bollard in 3D thus making the autonomous mooring more precise and accurate.

Autonomous Mooring towards Autonomous Maritime Navigation and Offshore Operations

Capacitive micromachined ultrasonic transducers (cMUT) technology is a new trend for intravascular ultrasound (IVUS) imaging. Large bandwidth, high sensitivity and compatibility to CMOS processes makes the cMUT a better choice compared to the conventional piezoelectric transducer. To exploit the merits of cMUT technology, an accurately designed front end circuit is required. The circuit functions as an output pulse driver for the generation of the acoustic signal and buffers the return echo. For an accurate evaluation before tape-out, the circuit has to be simulated using the post-layout extracted netlist of the IC with the electrical equivalent circuit that models the transducer pulse-echo behavior. In this paper, we present two different designs of front-end IC for 2D cMUT arrays that can be used for intravascular ultrasound imaging system. To simulate the response of the front-end circuit, we first developed a pulse-echo model for an array element using Mason Equivalent Circuit. The model is then combined with the front-end circuit using Cadence Spectre. The simulation results are verified by comparing them to experimental data obtained from the manufactured front-end IC. The results show that the front-end circuit tested with the equivalent circuit model of the cMUT elements is promising for the optimization of the overall system performance before manufacturing.

Front-end IC design for intravascular ultrasound imaging

Abstract Speech enables easy human-to-human communication as well as human-to-machine interaction. However, the quality of speech degrades due to background noise in the environment, such as drone noise embedded in speech during search and rescue operations. Similarly, helicopter noise, airplane noise, and station noise reduce the quality of speech. Speech enhancement algorithms reduce background noise, resulting in a crystal clear and noise-free conversation. For many applications, it is also necessary to process these noisy speech signals at the edge node level. Thus, we propose implicit Wiener filter-based algorithm for speech enhancement using edge computing system. In the proposed algorithm, a first order recursive equation is used to estimate the noise. The performance of the proposed algorithm is evaluated for two speech utterances, one uttered by a male speaker and the other by a female speaker. Both utterances are degraded by different types of non-stationary noises such as exhibition, station, drone, helicopter, airplane, and white Gaussian stationary noise with different signal-to-noise ratios. Further, we compare the performance of the proposed speech enhancement algorithm with the conventional spectral subtraction algorithm. Performance evaluations using objective speech quality measures demonstrate that the proposed speech enhancement algorithm outperforms the spectral subtraction algorithm in estimating the clean speech from the noisy speech. Finally, we implement the proposed speech enhancement algorithm, in addition to the spectral subtraction algorithm, on the Raspberry Pi 4 Model B, which is a low power edge computing device. 

/pdf/single-channel-speech-enhancement-using-implicit-wiener-1xd03pmd.pdf

Single-channel speech enhancement using implicit Wiener filter for high-quality speech communication

This paper proposes Particle Swarm Optimization (PSO) based fault-tolerant application mapping on to ZMesh topology based Network-on-Chip (NoC) design. Permanent faults in application cores has been considered and performed application mapping using PSO. The major contribution of this paper is to find out the best position for the spare core to be placed in the network using PSO. Experimentations have been carried out by scaling the ZMesh network size and percentage of network faults. The results show that the proposed approach leads to minimum overhead in communication cost over fault-free result.

Linga Reddy Cenkeramaddi

Papers

Reward criteria impact on the performance of reinforcement learning agent for autonomous navigation

Autonomous Mooring towards Autonomous Maritime Navigation and Offshore Operations

Front-end IC design for intravascular ultrasound imaging

Single-channel speech enhancement using implicit Wiener filter for high-quality speech communication

Fault-Tolerant Application Mapping on to ZMesh topology based Network-on-Chip Design