A Deep Learning Approach for Precision Viticulture, Assessing Grape Maturity via YOLOv7.

In the viticulture sector, robots are being employed more frequently to increase productivity and accuracy in operations such as vineyard mapping, pruning, and harvesting, especially in locations where human labor is in short supply or expensive. This paper presents the development of an algorithm for grape maturity estimation in the framework of vineyard management. An object detection algorithm is proposed based on You Only Look Once (YOLO) v7 and its extensions in order to detect grape maturity in a white variety of grape (Assyrtiko grape variety). The proposed algorithm was trained using images received over a period of six weeks from grapevines in Drama, Greece. Tests on high-quality images have demonstrated that the detection of five grape maturity stages is possible. Furthermore, the proposed approach has been compared against alternative object detection algorithms. The results showed that YOLO v7 outperforms other architectures both in precision and accuracy. This work paves the way for the development of an autonomous robot for grapevine management.

Coordinated Navigation of Two Agricultural Robots in a Vineyard: A Simulation Study.

The development of an effective agricultural robot presents various challenges in actuation, localization, navigation, sensing, etc., depending on the prescribed task. Moreover, when multiple robots are engaged in an agricultural task, this requires appropriate coordination strategies to be developed to ensure safe, effective, and efficient operation. This paper presents a simulation study that demonstrates a robust coordination strategy for the navigation of two heterogeneous robots, where one robot is the expert and the second robot is the helper in a vineyard. The robots are equipped with localization and navigation capabilities so that they can navigate the environment and appropriately position themselves in the work area. A modular collaborative algorithm is proposed for the coordinated navigation of the two robots in the field via a communications module. Furthermore, the robots are also able to position themselves accurately relative to each other using a vision module in order to effectively perform their cooperative tasks. For the experiments, a realistic simulation environment is considered, and the various control mechanisms are described. Experiments were carried out to investigate the robustness of the various algorithms and provide preliminary results before real-life implementation.

Efficacy of a Robot-Assisted Intervention in Improving Learning Performance of Elementary School Children with Specific Learning Disorders.

(1) Background: There has been significant recent interest in the potential role of social robots (SRs) in special education. Specific Learning Disorders (SpLDs) have a high prevalence in the student population, and early intervention with personalized special educational programs is crucial for optimal academic achievement. (2) Methods: We designed an intense special education intervention for children in the third and fourth years of elementary school with a diagnosis of a SpLD. Following confirmation of eligibility and informed consent, the participants were prospectively and randomly allocated to two groups: (a) the SR group, for which the intervention was delivered by the humanoid robot NAO with the assistance of a special education teacher and (b) the control group, for which the intervention was delivered by the special educator. All participants underwent pre- and post-intervention evaluation for outcome measures. (3) Results: 40 children (NAO = 19, control = 21, similar baseline characteristics) were included. Pre- and post-intervention evaluation showed comparable improvements in both groups in cognition skills (decoding, phonological awareness and reading comprehension), while between-group changes favored the NAO group only for some phonological awareness exercises. In total, no significant changes were found in any of the groups regarding the emotional/behavioral secondary outcomes. (4) Conclusion: NAO was efficient as a tutor for a human-supported intervention when compared to the gold-standard intervention for elementary school students with SpLDs.

A Pediatric Patient With Autism Spectrum Disorder and Comorbid Compulsive Behaviors Treated With Robot-Assisted Relaxation: A Case Report.

The nature of autism spectrum disorders (ASDs) presents significant challenges, especially with regard to comorbidities and drug treatments. Parents and caregivers are often hesitant towards psychotropic medications, mainly due to the fear of side effects. Problems arise when comorbid symptomatology reaches extreme levels, leading to functional decline in the patients. We discuss the case of a 13-year-old girl diagnosed with autism disorder who presented with a long history of social, interpersonal, and academic challenges. The patient was hospitalized with the complaint of a persistent, non-painful, and unpleasant sensation in the perineal area that eventually resulted in repetitive and compulsive behaviors. Robot-enhanced relaxation training was introduced to support the patient since she declined to undergo any form of talk therapy. The aim of the intervention was to prevent the irritation from escalating and promote self-regulation skills. The results, based on parent reporting, indicated that the patient acquired relaxation skills, experienced some positive effects on emotional regulation, and showed a decrease in the duration of her disruptive behaviors upon completing the relaxation training. This case report provides evidence that robot-assisted relaxation training may be effective in dealing with ASD-related behavioral disturbances and comorbid anxiety.

Behavioral Data Analysis of Robot-Assisted Autism Spectrum Disorder (ASD) Interventions Based on Lattice Computing Techniques.

Recent years have witnessed the proliferation of social robots in various domains including special education. However, specialized tools to assess their effect on human behavior, as well as to holistically design social robot applications, are often missing. In response, this work presents novel tools for analysis of human behavior data regarding robot-assisted special education. The objectives include, first, an understanding of human behavior in response to an array of robot actions and, second, an improved intervention design based on suitable mathematical instruments. To achieve these objectives, Lattice Computing (LC) models in conjunction with machine learning techniques have been employed to construct a representation of a child's behavioral state. Using data collected during real-world robot-assisted interventions with children diagnosed with Autism Spectrum Disorder (ASD) and the aforementioned behavioral state representation, time series of behavioral states were constructed. The paper then investigates the causal relationship between specific robot actions and the observed child behavioral states in order to determine how the different interaction modalities of the social robot affected the child's behavior.

Machine Vision Systems in Precision Agriculture for Crop Farming.

Machine vision for precision agriculture has attracted considerable research interest in recent years. The aim of this paper is to review the most recent work in the application of machine vision to agriculture, mainly for crop farming. This study can serve as a research guide for the researcher and practitioner alike in applying cognitive technology to agriculture. Studies of different agricultural activities that support crop harvesting are reviewed, such as fruit grading, fruit counting, and yield estimation. Moreover, plant health monitoring approaches are addressed, including weed, insect, and disease detection. Finally, recent research efforts considering vehicle guidance systems and agricultural harvesting robots are also reviewed.

Lattice computing extension of the FAM neural classifier for human facial expression recognition.

This paper proposes a fundamentally novel extension, namely, flrFAM, of the fuzzy ARTMAP (FAM) neural classifier for incremental real-time learning and generalization based on fuzzy lattice reasoning techniques. FAM is enhanced first by a parameter optimization training (sub)phase, and then by a capacity to process partially ordered (non)numeric data including information granules. The interest here focuses on intervals' numbers (INs) data, where an IN represents a distribution of data samples. We describe the proposed flrFAM classifier as a fuzzy neural network that can induce descriptive as well as flexible (i.e., tunable) decision-making knowledge (rules) from the data. We demonstrate the capacity of the flrFAM classifier for human facial expression recognition on benchmark datasets. The novel feature extraction as well as knowledge-representation is based on orthogonal moments. The reported experimental results compare well with the results by alternative classifiers from the literature. The far-reaching potential of fuzzy lattice reasoning in human-machine interaction applications is discussed.

Granular self-organizing map (grSOM) for structure identification.

This work presents a useful extension of Kohonen's Self-Organizing Map (KSOM) for structure identification in linguistic (fuzzy) system modeling applications. More specifically the granular SOM neural model is presented for inducing a distribution of nonparametric fuzzy interval numbers (FINs) from the data. A FIN can represent a local probability distribution function and/or a conventional fuzzy set; moreover, a FIN is interpreted as an information granule. Learning is based on a novel metric distance d(K)(.,.) between FINs. The metric d(K)(.,.) can be tuned nonlinearly by a mass function m(x), the latter attaches a weight of significance to a real number 'x' in a data dimension. Rigorous analysis is based on mathematical lattice theory. A grSOM can cope with ambiguity by processing linguistic (fuzzy) input data and/or intervals. This work presents a simple grSOM variant, namely greedy grSOM, for classification. A genetic algorithm (GA) introduces tunable nonlinearities during training. Extensive comparisons are shown with related work from the literature. The practical effectiveness of the greedy grSOM is demonstrated comparatively in three benchmark classification problems. Statistical evidence strongly suggests that the proposed techniques improve classification performance. In addition, the greedy grSOM induces descriptive decision-making knowledge (fuzzy rules) from the training data.