Ripeness estimation of fruits and vegetables is a key factor for the optimization of field management and the harvesting of the desired product quality. Typical ripeness estimation involves multiple manual samplings before harvest followed by chemical analyses. Machine vision has paved the way for agricultural automation by introducing quicker, cost-effective, and non-destructive methods. This work comprehensively surveys the most recent applications of machine vision techniques for ripeness estimation. Due to the broad area of machine vision applications in agriculture, this review is limited only to the most recent techniques related to grapes. The aim of this work is to provide an overview of the state-of-the-art algorithms by covering a wide range of applications. The potential of current machine vision techniques for specific viticulture applications is also analyzed. Problems, limitations of each technique, and future trends are discussed. Moreover, the integration of machine vision algorithms in grape harvesting robots for real-time in-field maturity assessment is additionally examined.

Our interest is in time series classification regarding cyber–physical systems (CPSs) with emphasis in human-robot interaction. We propose an extension of the k nearest neighbor (kNN) classifier to time-series classification using intervals’ numbers (INs). More specifically, we partition a time-series into windows of equal length and from each window data we induce a distribution which is represented by an IN. This preserves the time dimension in the representation. All-order data statistics, represented by an IN, are employed implicitly as features; moreover, parametric non-linearities are introduced in order to tune the geometrical relationship (i.e., the distance) between signals and consequently tune classification performance. In conclusion, we introduce the windowed IN kNN (WINkNN) classifier whose application is demonstrated comparatively in two benchmark datasets regarding, first, electroencephalography (EEG) signals and, second, audio signals. The results by WINkNN are superior in both problems; in addition, no ad-hoc data preprocessing is required. Potential future work is discussed.

No abstract

In their verbal interactions, humans are often afforded with language barriers and communication problems and disabilities. This problem is even more serious in the fields of education and health care for children with special needs. The use of robotic agents, notably humanoids integrated within human groups, is a very important option to face these limitations. Many scientific research projects attempt to provide solutions to these communication problems by integrating intelligent robotic agents with natural language communication abilities. These agents will thus be able to help children suffering from verbal communication disorders, more particularly in the fields of education and medicine. In addition, the introduction of robotic agents into the child's environment creates stimulating effects for more verbal interaction. Such stimulation may improve their ability to interact with pairs. In this paper, we propose a new approach for the human-robot multilingual verbal interaction based on hybridization of recent and performant approach on translation machine system consisting of neural network model reinforced by a large distributed domain-ontology knowledge database. We have constructed this ontology by crawling a large number of educational web sites providing multilingual parallel texts and speeches. Furthermore, we present the design of augmented LSTM neural Network models and their implementation to permit, in learning context, communication between robots and children using multiple natural languages. The model of a general ontology for multilingual verbal communication is produced to describe a set of linguistic and semantic entities, their properties and relationships. This model is used as an ontological knowledge base representing the verbal communication of robots with children.

No abstract