Search

Recently, agriculture has gained much attention regarding automation by artificial intelligence techniques and robotic systems. Particularly, with the advancements in machine learning (ML) concepts, significant improvements have been observed in agricultural tasks. The ability of automatic feature extraction creates an adaptive nature in deep learning (DL), specifically convolutional neural networks to achieve human-level accuracy in various agricultural applications, prominent among which are plant disease detection and classification, weed/crop discrimination, fruit counting, land cover classification, and crop/plant recognition. This review presents the performance of recent uses in agricultural robots by the implementation of ML and DL algorithms/architectures during the last decade. Performance plots are drawn to study the effectiveness of deep learning over traditional machine learning models for certain agricultural operations. The analysis of prominent studies highlighted that the DL-based models, like RCNN (Region-based Convolutional Neural Network), achieve a higher plant disease/pest detection rate (82.51%) than the well-known ML algorithms, including Multi-Layer Perceptron (64.9%) and K-nearest Neighbour (63.76%). The famous DL architecture named ResNet-18 attained more accurate Area Under the Curve (94.84%), and outperformed ML-based techniques, including Random Forest (RF) (70.16%) and Support Vector Machine (SVM) (60.6%), for crop/weed discrimination. Another DL model called FCN (Fully Convolutional Networks) recorded higher accuracy (83.9%) than SVM (67.6%) and RF (65.6%) algorithms for the classification of agricultural land covers. Finally, some important research gaps from the previous studies and innovative future directions are also noted to help propel automation in agriculture up to the next level.

Air pollution poses significant risk to environment and health. Air quality monitoring stations are often confined to a small number of locations due to the high cost of the monitoring equipment. They provide a low fidelity picture of the air quality in the city; local variations are overlooked. However, recent developments in low-cost sensor technology and wireless communication systems like Internet of Things (IoT) provide an opportunity to use arrayed sensor networks to measure air pollution, in real time, at a large number of locations. This article reports the development of a novel low-cost sensor node that utilizes cost-effective electrochemical sensors to measure carbon monoxide (CO) and nitrogen dioxide (NO2) concentrations and an infrared sensor to measure particulate matter (PM) levels. The node can be powered by either solar-recharged battery or mains supply. It is capable of long-range, low power communication over public or private long-range wide area network (LoRaWAN) IoT network and short-range high data rate communication over Wi-Fi. The developed sensor nodes were co-located with an accurate reference CO sensor for field calibration. The low-cost sensors’ data, with offset and gain calibration, show good correlation with the data collected from the reference sensor. Multiple linear regression (MLR)-based temperature and humidity correction results in mean absolute percentage error (MAPE) of 48.71% and $R^{2}$ of 0.607 relative to the reference sensor’s data. Artificial neural network (ANN)-based calibration shows the potential for significant further improvement with MAPE of 38.89% and $R^{2}$ of 0.78 for leave-one-out cross-validation.

The collagen hydrolysate, a proteinic biopeptide, is used for various key functionalities in humans and animals. Numerous reviews explained either individually or a few of following aspects: types, processes, properties, and applications. In the recent developments, various biological, biochemical, and biomedical functionalities are achieved in five aspects: process, type, species, disease, receptors. The receptors are rarely addressed in the past which are an essential stimulus to activate various biomedical and biological activities in the metabolic system of humans and animals. Furthermore, a systematic segregation of the recent developments regarding the five main aspects is not yet reported. This review presents various biological, biochemical, and biomedical functionalities achieved for each of the beforementioned five aspects using a systematic approach. The review proposes a novel three-level hierarchy that aims to associate a specific functionality to a particular aspect and its subcategory. The hierarchy also highlights various key research novelties in a categorical manner that will contribute to future research.

The identification of plant disease is an imperative part of crop monitoring systems. Computer vision and deep learning (DL) techniques have been proven to be state-of-the-art to address various agricultural problems. This research performed the complex tasks of localization and classification of the disease in plant leaves. In this regard, three DL meta-architectures including the Single Shot MultiBox Detector (SSD), Faster Region-based Convolutional Neural Network (RCNN), and Region-based Fully Convolutional Networks (RFCN) were applied by using the TensorFlow object detection framework. All the DL models were trained/tested on a controlled environment dataset to recognize the disease in plant species. Moreover, an improvement in the mean average precision of the best-obtained deep learning architecture was attempted through different state-of-the-art deep learning optimizers. The SSD model trained with an Adam optimizer exhibited the highest mean average precision (mAP) of 73.07%. The successful identification of 26 different types of defected and 12 types of healthy leaves in a single framework proved the novelty of the work. In the future, the proposed detection methodology can also be adopted for other agricultural applications. Moreover, the generated weights can be reused for future real-time detection of plant disease in a controlled/uncontrolled environment.

Recently, plant disease classification has been done by various state-of-the-art deep learning (DL) architectures on the publicly available/author generated datasets. This research proposed the deep learning-based comparative evaluation for the classification of plant disease in two steps. Firstly, the best convolutional neural network (CNN) was obtained by conducting a comparative analysis among well-known CNN architectures along with modified and cascaded/hybrid versions of some of the DL models proposed in the recent researches. Secondly, the performance of the best-obtained model was attempted to improve by training through various deep learning optimizers. The comparison between various CNNs was based on performance metrics such as validation accuracy/loss, F1-score, and the required number of epochs. All the selected DL architectures were trained in the PlantVillage dataset which contains 26 different diseases belonging to 14 respective plant species. Keras with TensorFlow backend was used to train deep learning architectures. It is concluded that the Xception architecture trained with the Adam optimizer attained the highest validation accuracy and F1-score of 99.81% and 0.9978 respectively which is comparatively better than the previous approaches and it proves the novelty of the work. Therefore, the method proposed in this research can be applied to other agricultural applications for transparent detection and classification purposes.

Air pollution poses significant risks to environment and health. Air quality monitoring stations are often confined to a small number of locations due to the high cost of the monitoring equipment. They provide a low fidelity picture of the air quality in the city; local variations are overlooked. However, recent developments in low cost sensor technology and wireless communication systems like Internet of Things (IoT) provide an opportunity to use arrayed sensor networks to measure air quality, in real time, at a large number of locations. This paper reports the development of a novel low cost sensor node that utilizes cost-effective electrochemical sensors to measure Carbon Monoxide (CO) and Nitrogen Dioxide (NO2) concentrations and an infrared sensor to measure Particulate Matter (PM) levels. The node can be powered by either solar-recharged battery or mains supply. It is capable of long-range, low power communication over public or private LoRaWAN IoT network and short-range high data rate communication over Wi-Fi. The developed sensor nodes were co-located with an accurate reference CO sensor for field calibration. The low cost sensors’ data shows strong correlation with the data collected from the reference sensor. Offset and gain calibration further improves the quality of the sensor data.

The extrusion system is an integral part of any fused deposition style 3D printing technique. However, the extruder designs found in commercial and hobbyist printers are mostly suitable for materials in filament form. While printing with a filament is not a problem per se, the printing of materials that may not be readily available in the filament form or not commercially viable remains untapped, e.g., biopolymers and material blends. This is particularly an issue in the research and hobbyist space where the capability of printing a variety of materials or materials recycled from already printed parts may be of utmost importance. This paper presents a pellet-based extrusion system for the 3D printing of biopolymers. The system has been designed from the first principles and therefore can be extended to other materials with parameter adjustments or slight hardware modifications. A robust mechatronic design has been realized using an unconventional yet simplistic approach. The extrusion system uses a series of control factors to generate a consistent output of material over the course of a print. The platform and surrounding processes are set up so that software can be used to define the printing parameters; this allows a simpler adaption to different materials. The utility of the extruder is demonstrated through extensive printing and testing of the printed parts.

Plant diseases affect the growth of their respective species, therefore their early identification is very important. Many Machine Learning (ML) models have been employed for the detection and classification of plant diseases but, after the advancements in a subset of ML, that is, Deep Learning (DL), this area of research appears to have great potential in terms of increased accuracy. Many developed/modified DL architectures are implemented along with several visualization techniques to detect and classify the symptoms of plant diseases. Moreover, several performance metrics are used for the evaluation of these architectures/techniques. This review provides a comprehensive explanation of DL models used to visualize various plant diseases. In addition, some research gaps are identified from which to obtain greater transparency for detecting diseases in plants, even before their symptoms appear clearly.

Industrial progression of Additive Manufacturing awaits novel quality management strategies to verify 3D printed parts are fit for purpose. This paper presents a review of testing technologies: tensile testing and Scanning Electron Microscopy (SEM). The experiment performed aimed to provide a comparative study, of the testing technologies, to better understand current quality management processes in the 3D printing space. Additionally, this study was performed to establish a good fundamental groundwork on which future Non-Destructive Testing research can be conducted. The study revealed the advantages of tensile and SEM testing methods, whilst highlighting a need for better technologies to cover the disadvantages.

Mapping and localization is a well studied area in mobile robotics as all robots need to position themselves in the environment correctly to do their tasks. However, most maps lack any information about the surface they operate on. If a surface map is available, it can be used for robust control of the drive system as well as for planning of tasks such as cleaning, polishing, inspection, mowing, etc. This paper presents an investigation into the use of a robot-mounted 2D laser scanner for floor surface point cloud generation. The robot wheel odometry information is used for pose and orientation tracking. The system reliably produces point cloud surfaces of three types of flooring; carpet, coated asphalt, and asphalt. The devised method of surface profiling is effective for a fast overview of the floor and can be improved through more robust localisation, and greater laser scanner accuracy.

Medical exoskeletons are robotic devices that provide paraplegics with the ability to ambulate again. There are four current market leaders in the industry: REX, by Rex Bionics; ReWalk, by Argo Medical Technologies Ltd; Ekso GT, by Ekso Bionics; and Indego, by Parker Hannifin. A weighted table-based method was used to evaluate the capability of these four devices, which was then compared to their maximum speed. The Ekso GT and the ReWalk devices achieved high speeds but had low capability scores due to their instability and minimal safety features. The REX and Indego devices achieved the highest capability scores but for contrasting reasons; the REX devices are inherently stable but have a slow, static gait, while the Indego device relies on walking aids for stability and achieves a faster speed. Both devices scored high in the safety criteria. It was derived from the results that there is an indirectly proportional relationship between speed and stability, and none of the exoskeletons achieve a balance between the two. This study found that there is an opportunity to enhance the speed, mobility, stability, and cost of medical exoskeletons, which is presented below.

The predominance of conventional polymer processes (CPP) is undoubtedly formidable regarding monetary benefits gained in lieu of mass production in the global market. In this research, profitable commercial status of CPP is critically questioned in case of a specific large-scale application in low quantities, considering the following three factors: cost, time, and complexity. This research paper reports the subjective analysis-based exclusion of conventional processes (injection moulding) that brings the ball in the court of additive manufacturing (AM) technologies which are superior to injection moulding in low volume production. However, the AM techniques are limited to print only small-scale products in low quantity. This paper proposes a novel application of large scale products in low quantities by extruding polymer pellets from conceptual extrusion based deposition modeling setup. The requirement of a suitable material for conceptual extrusion setup entails the analysis of various factors to verify the feasibility of current fused deposition modeling materials for the conceptual large extrusion based deposition modeling setup. Therefore, this research contributes through providing a feasibility report for the CPP, AM techniques and their materials to print large-scale products.

3D Printing additive manufacturing is a rapidly developing form of technology. Currently able to manipulate many polymers (both synthetic and organic) this technique is quickly becoming an integral part of biotechnological developments. This paper highlights the fundamentals of this technology namely the mechanisms employed in standard 3D printing it then introduces tissue engineering a field in which current versions of this technology have been employed as bioprinting. The limitations with respect to tissue engineering are discussed outlining the current technologies inability to produce nanofibre based structures common in tissue such as tendon cartilage and cornea. From this requirement for nanofibre production electrospinning is introduced as a potential pathway for future tissue engineering 3D printing technologies and finally the current combination of this technology with 3D Printing is discussed yielding current limitations in retaining required nano-resolutions.

Consumer-grade RGB-D cameras capture RGB images along with per-pixel depth information, and because of their limited cost and ability to measure distances at a high frame rate, have been used in robotics and computer vision application. However, drawbacks include the repeatability and accuracy of RGB-D cameras for object detection and localization. This paper investigates and compares RGB-D cameras' performance in terms of depth image quality, depth clouds distribution, etc. performance and configuration methods of frequently used cameras, e.g. PrimeSense, Kinect V1 and Kinect V2, in order to provide useful advice when choosing a camera for robotic applications. Experimental and Point Cloud Library (PCL)-based methods are introduced for point-to-plane distance detection. Based on the obtained results, a relationship between measurements and ground truth is built.

Additive manufacturing (AM) allows designers to build 3D parts of almost any geometry and streamline the process of moving from initial design to prototype in a shortened timeframe. However the existing technologies are limited to small size parts due only to high resolution of the generally produced designs. If a very large part with typical layer resolutions of 0.2mm is produced, the time required for making such a part will be prohibitive. In addition, the material costs will excessive and the machine size will need to be gigantic. In large scale printing the material and the high costs restrict testing on the actual structures and lends itself to the use of numerical or analytical studies of the structures before actual printing. In this paper we analyse the current state of the 3D printing within the boundaries of large scale printing and present a possible solution with numerical simulations.

A leading area of research in Additive Manufacturing, studies the quality of 3D prints using several testing strategies. This paper reviews two testing technologies: tensile testing (mechanical testing) and Scanning Electron Microscopy (SEM). The aim of this experiment was to provide a comparative study to better understand quality management processes in the 3D printing space. Additionally, this research study was performed to establish a good fundamental groundwork on which future Non-Destructive Testing research can be conducted. The study revealed the advantages of the tensile and SEM testing methods, whilst highlighting a need for better technologies to cover the disadvantages.

Low cost and accurate 3D surface profiling can help in numerous industry applications including inspection tasks, cleaning, minimizing bumps in navigation of non-uniform terrain, aid navigation, and road/pavement condition analysis. However, most of the available systems are costly or inaccessible for widespread use. This research presents investigation into the capability of cheap and accessible sensors to capture the floor surface profile information. A differential drive robotic platform has been developed to perform testing and conduct the research. 2D localization methods are extrapolated into 3D for the floor capturing process. Two different types of sensors, a 2D laser scanner and an RGB-D camera, are used for comparison of the floor profile capture ability. The robotic system is able to successfully capture the floor surface profile of a number of different type floors such as carpet, asphalt, and a coated floor. A key finding is that the surface itself is a significant factor on the measured profile, i.e. dirt or differing materials can cause false height measurements. Overall the methodology has proved a successful real time solution for creating a point cloud of the floor surface.

Matsuoka neuronal oscillator is proposed to control the traffic signals of an isolated four-phase signalized intersection. The oscillator is a model of central pattern generator (CPG) and has seen various applications in humanoid robots. Matsuoka oscillator was chosen for the traffic signal control because of its stable and predictable rhythmic outputs that exploit autonomously the dynamics of the road system. In this paper, the dynamics of Matsuoka oscillator was described in a set of first-order differential equations and simluated in an agent-based modelling environment. This novel signal control algorithm was validated in a Application Programming Interface (API) function by AIMSUN (Advanced Interactive Microscopic Simulator for Urban and Non-Urban Networks). The results were compared to the performance of the existing traffic system and have shown the potential capability of the proposed algorithm in reductions of vehicle delay time and queues.

This paper examines additive manufacturing technologies in the context of their potential use in flexible manufacturing systems. It reviews which current technologies are capable of producing full-strength production parts. It also examines which technologies might be applicable to FMS and how they might be implemented as part of a hybrid manufacturing cell.

A masticatory robot refers to a robot that can perform at least some defined human masticatory functions. This paper briefly reviews the masticatory system, masticatory measurements and computational models of mastication that are relevant to masticatory robotics. Also critically reviewed is the state of the art of the robotics research in engineering of the jaw system. The masticatory system has two rigid components: a fixed maxillary (upper) jaw and a mobile mandibular (lower) jaw, which are joined by two temporomandibular joints (TMJ). Unique features of the TMJ are described. The role of muscles of mastication is explained with regard to their role in rhythmic opening and closing of the mandible in three-dimensional space. Because the breakdown of food is performed directly by the teeth; the functionalities of the incisor, pre-molar and rear molar are presented. Two computational models of the masticatory system are presented in which Hill-type muscle models are used. We also describe masticatory robots developed for dental training, jaw simulation, food texture and breakdown analysis, and speech therapy with regard to muscle modelling, TMJ models, masticatory biomechanics and controls of actuation. Finally, we discuss the major accomplishments and challenges in masticatory modelling and robotics; and we compare a number of such robots in the light of relevant biomechanical aspects of the mastication system.

Mastication is a complex process influenced by numerous factors including those associated with an individual and the ingested food. Human chewing behaviour can be characterised by measuring mandibular movements and muscular activities during a masticatory sequence or by measuring the particle size distribution and rheological characteristics of the swallowed food mass. To constructively understand the mastication process and assess the mastication performance, a formal description of the chewing behaviour is proposed in this paper. An object-oriented model is developed and described in Unified Modelling Language (UML). The chewing behaviour model is composed of three objects, one for the jaw's physiological apparatus, one for the properties defining the mastication process and foods being chewed, and a further one for the association of the properties. A complete representation of the chewing behaviour is achieved by linking three object models via an additional class for chewing data that is collected experimentally. With the object model, the chewing behaviour is further instantiated by discovering knowledge hidden in the chewing database by data mining. A case study is presented to show the procedure of how the hidden knowledge is discovered and the data mining results are interpreted in the context of food science.

One of the issues associated with programming a VEX Robotics Competition (VRC) robot for the autonomous period is providing it with enough information regarding its environment so that it can move about the field intelligently. As such, an objective of our research was to develop a series of machine vision tools so that a VRC robot could identify VRC field elements, other robots, and field perimeters; responding appropriately. We have carried out a review of relevant literature and identified a number of algorithms, image processing tools, and control paradigms, as well as, developed our own approaches, which we have implemented in C++ using the OpenCV library. Here we present the results of our initial efforts, namely our implemented colour identification, connected-object separation, and multiple connected-objects separation methodologies.

A jaw exoskeleton is proposed in this paper to assist the exercise for the purpose of practical rehabilitation of temporomandibular disorder (TMD). The jaw, attached to the skull by muscles and pivoted at the condyle via the temporomandibular joint (TMJ), can be simplified as moving in the two-dimensional sagittal plane. Based on the in-vivo recording the jaw movement from the healthy subject, the motion pattern is justified to be the primary specification to design the jaw exoskeleton. A planar four-bar linkage is synthesized to reproduce the specified normal jaw motion in terms of incisor and condyle trajectory on the coupler point to meet the kinematic specification. Adjustable lengths of the links are used to achieve a group of trajectories of any possibility.

Recent stereolithography rapid prototyping advancements have been in the area of microstereolithography. Microstereolithography incorporates an integral curing approach to the creation of layers. Integral curing has been made possible by the advent of dynamic masking generators such as LCD and DMD devices. Much of the theory for this new layering process has only been applied on a micro-scale and awaits further development for application on medium to large sized parts. This paper presents some recent research on the topic of the integral curing process and analyses a mathematical model used for process planning of integral curing using a microstereolithography apparatus. Aspects of the mathematical model are enhanced to prepare the way for its application to medium and large sized parts, and two methods for achieving even irradiance for each model layer are proposed.

Most current industrial robot arms require a dedicated controller for the actuating systems. This can be a disadvantage when trying to integrate several robots into an agile manufacturing environment. More flexible and adaptive modular plug-and-play controllers can highly enhance the use of these robots and eases their integration into modern, agile manufacturing environments. Interfacing automated machines can then be done at a PC level. At this level, “plug-and-play” becomes the benchmark for new devices being added to the system, allowing ease of operation and increased flexibility for agile manufacturing. The modular mechatronic control system described in this paper was used to operate a Unimate PUMA 560 series industrial robotic arm.