Nowadays, the most adopted technique to address infertility problems is in vitro fertilisation (IVF). However, its success rate is limited, and the associated procedures, known as assisted reproduction technology (ART), suffer from a lack of objectivity at the laboratory level and in clinical practice. This paper deals with applications of Artificial Intelligence (AI) techniques to IVF procedures. Artificial intelligence is considered a promising tool for ascertaining the quality of embryos, a critical step in IVF. Since the oocyte quality influences the final embryo quality, we present a systematic review of the literature on AI-based techniques used to assess oocyte quality; we analyse its results and discuss several promising research directions. In particular, we highlight how AI-based techniques can support the IVF process and examine their current applications as presented in the literature. Then, we discuss the challenges research must face in fully deploying AI-based solutions in current medical practice. Among them, the availability of high-quality data sets as well as standardised imaging protocols and data formats, the use of physics-informed simulation and machine learning techniques, the study of informative, descriptive yet observable features, and, above all, studies of the quality of oocytes and embryos, specifically about their live birth potential. An improved understanding of determinants for oocyte quality can improve success rates while reducing costs, risks for long-term embryo cultures, and bioethical concerns.

This article presents an open-source architecture for conveying robots’ intentions to human teammates using Mixed Reality and Head-Mounted Displays. The architecture has been developed focusing on its modularity and re-usability aspects. Both binaries and source code are available, enabling researchers and companies to adopt the proposed architecture as a standalone solution or to integrate it in more comprehensive implementations. Due to its scalability, the proposed architecture can be easily employed to develop shared Mixed Reality experiences involving multiple robots and human teammates in complex collaborative scenarios.

The problem of autonomous transportation in industrial scenarios is receiving a renewed interest due to the way it can revolutionise internal logistics, especially in unstructured environments. This paper presents a novel architecture allowing a robot to detect, localise, and track (possibly multiple) pallets using machine learning techniques based on an on-board 2D laser rangefinder only. The architecture is composed of two main components: the first stage is a pallet detector employing a Faster Region-Based Convolutional Neural Network (Faster R-CNN) detector cascaded with a CNN-based classifier; the second stage is a Kalman filter for localising and tracking detected pallets, which we also use to defer commitment to a pallet detected in the first stage until sufficient confidence has been acquired via a sequential data acquisition process. For fine-tuning the CNNs, the architecture has been systematically evaluated using a real-world dataset containing 340 labelled 2D scans, which have been made freely available in an online repository. Detection performance has been assessed on the basis of the average accuracy over k-fold cross-validation, and it scored 99.58% in our tests. Concerning pallet localisation and tracking, experiments have been performed in a scenario where the robot is approaching the pallet to fork. Although data have been originally acquired by considering only one pallet as per specification of the use case we consider, artificial data have been generated as well to mimic the presence of multiple pallets in the robot workspace. Our experimental results confirm that the system is capable of identifying, localising and tracking pallets with a high success rate while being robust to false positives.

A common way to organize a high number of robots, both when moving autonomously and when controlled by a human operator, is to let them move in formation. This is a principle that takes inspiration from the nature, that maximizes the possibility of monitoring the environment and therefore of anticipating risks and finding targets. In robotics, alongside these reasons, the organization of a robot team in a formation allows a human operator to deal with a high number of agents in a simpler way, moving the swarm as a single entity. In this context, the typology of visual feedback is fundamental for a correct situational awareness, but in common practice having an optimal camera configuration is not always possible. Usually human operators use cameras on board the multirotors, with an egocentric point of view, while it is known that in mobile robotics overall awareness and pattern recognition are optimized by exocentric views. In this article we present an analysis of the performance achieved by human operators controlling a swarm of UAVs in formation, accomplishing different tasks and using different point of views. The control architecture is implemented in a ROS framework and interfaced with a 3D simulation environment. Experimental tests show a degradation of performance while using egocentric cameras with respect of an exocentric point of view, although cameras on board the robots allow to satisfactorily accomplish simple tasks.

The automatic assessment of the level of independence of a person, based on the recognition of a set of Activities of Daily Living, is among the most challenging research fields in Ambient Intelligence. The article proposes a framework for the recognition of motion primitives, relying on Gaussian Mixture Modeling and Gaussian Mixture Regression for the creation of activity models. A recognition procedure based on Dynamic Time Warping and Mahalanobis distance is found to: (i) ensure good classification results; (ii) exploit the properties of GMM and GMR modeling to allow for an easy run-time recognition; (iii) enhance the consistency of the recognition via the use of a classifier allowing unknown as an answer.

This paper describes a system for context awareness in smart environments, which is based on an ontology expressed in description logic and implemented in OWL 2 EL, which is a subset of the Web Ontology Language that allows for reasoning in polynomial time. The approach is different from all other works in the literature since the proposed system requires only the basic reasoning mechanisms of description logic, i.e., subsumption and instance checking, without any additional external reasoning engine. Experiments performed with data collected in three different scenarios are described, i.e., the CASAS Project at Washington State University, the assisted living facility Villa Basilea in Genoa, and the Merry Porter mobile robot at the Polyclinic of Modena.

The article describes a system for describing and recognizing spatial and temporal patterns of events. The system is based on an ontology described through the Description Logics formalism and implemented in OWL DL. The approach is different from all other works in the literature since the system does not require an external reasoning engine, but relies only on the base mechanism for ontology classification. Experiments performed in two different scenarios are described, i.e., a Smart Home and a mobile robot for autonomous transportation operating within a partially automated building.

Humanoid robots operating in the real world must exhibit very complex behaviors, such as object manipulation or interaction with people. Such capabilities pose the problem of being able to reason on a huge number of different objects, places and actions to carry out, each one relevant for achieving robot goals. This article proposes a functional representation of objects, places and actions described in terms of affordances and capabilities. Everyday problems can be efficiently dealt with by decomposing the reasoning process in two phases, namely problem awareness (which is the focus of this article) and action selection.

This article introduces a model for representing motion primitives for entertainment humanoid robots using Generalized Hierarchical AND/OR graphs. On the one hand, the goal is to drive robots with scripts, as if they were on a stage. On the other hand, the approach allows for storing a minimum amount of behaviours, thereby reducing on-board memory and computational requirements. Standard ontology-based reasoning mechanisms are used to operate on such a representation, in a fully hierarchical fashion. Experimental validation has been assessed using toy Kondo robots.

A new feedback control model is provided that allows a wheeled vehicle to follow a prescribed path. Differently from all other methods in the literature, the method that is proposed neither requires the computation of a projection of the robot position on the path, nor does it need to consider a moving virtual target to be tracked. Nevertheless, it guarantees asymptotic convergence to a generic 2-D curve which can be represented through its implicit equation in the formf(x,y)=0, and it puts no bounds on the initial position of the vehicle, provided that ∇f≠ 0 .

The focus of the present work is on daily activity planning, i.e., representations and algorithms able to produce a course of action to deal efficiently with problems of daily living. To achieve this, the article proposes a functional representation of everyday objects, places and actions described in terms of affordances. Its contributions are two--fold: (i) it proposes to represent affordances and capabilities as regions in a proper affordance and capability space, and to describe such regions using neural maps; (ii) it proposes to decompose the planning process into different activities, by introducing a phase referred to as Problem Awareness preceding Action Planning, which allows to reduce the planning space in order to tackle large-scale planning problems.

The article proposes a feedback control system for path following in presence of obstacles that is an extension of previous work and is made of two components: (i) a sensor-based, real-time model that generates and periodically updates the path on-line in order to avoid both known and unforeseen obstacles, and (ii) a feedback-control model that is capable of driving a unicycle vehicle along the collision free path. The system has some unique characteristics, among which it requires very few computational resources as a consequence of its extreme simplicity.

The article proposes a new feedback control model which is suited for path following in a 3 Dimensional Cartesian space. Differently from other methods in literature, the method proposed neither requires to compute a projection of the robot's position on the path, nor it needs considering a moving virtual target. In spite of this: i) it guarantees asymptotic stability for every 3D curve which can be represented through a couple of intersecting surfaces f1(X, Y, Z) = 0, f2(X, Y, Z) = 0; ii) it does not put any bounds on the initial position of the vehicle depending on the path's curvature.

This paper presents an architecture for context-aware Ubiquitous Robotics applications, where mobile robots cooperate with intelligent environments to fulfill their tasks. Specifically, the work is focused on distributed knowledge representation issues and context assessment strategies, and introduces a technique for on-line context recognition in highly dynamic environments. Experimental validation, performed in a civillian hospital building, is described and discussed.

A knowledge representation system is introduced that allows the recognition of temporal patterns of events in context-aware environments. The system is based on standard frameworks, such as an ontology, an inference mechanism, and relational operators acting on numerical quantities. The paper describes how knowledge is managed, then introduces a collection of temporal operators that are inspired by the Allen's interval algebra, and then details a situation recognition algorithm to assess knowledge semantics. An example is reported to describe the approach.

The article proposes a feedback control system for real-time navigation and obstacle avoidance that is made of two components: (i) a sensor-based, real-time model that generates and periodically updates the path on-line in order to avoid both known and unforeseen obstacles, and (ii) a feedback-control model that is capable of driving a unicycle vehicle along the collision free path. The system has some unique characteristics, among which it requires very few computational resources as a consequence of its extreme simplicity. In spite of this, it is formally demonstrated to be asymptotically stable, as well as computationally efficient to be implemented in real-world scenarios where obstacles are not known, and possibly move in the environment.

The article proposes a new feedback control model that allows to track a generic curve in the Cartesian Space expressed through its implicit equation, and has minimal requirements in terms of measurement and computation capabilities. The model measures only the distance between the vehicle and the path, whereas it ignores the vehicle's orientation. In spite of this, it allows to regulate to zero both the distance to the path and the difference between the vehicle's orientation and the tangent to the curve, and it is asymptotically stable.

This paper presents muNav, a novel approach to navigation which, with minimal requirements in terms of onboard sensory, memory, and computational power, exhibits way-finding behaviors in very complex environments. The algorithm is intrinsically robust, since it does not require any internal geometrical representation or self-localization capabilities. Experimental results, performed with both simulated and real robots, validate the proposed theoretical approach.

A framework is introduced that is aimed at integrating ontology and logic approaches for context-awareness, suitable for use in Ambient Intelligence (AmI) scenarios. In particular, the context description language CDL is described, which allows to easily specify patterns of events which occurrences must be monitored by actual systems. As long as systems evolve, symbolic data originating from heterogeneous sources are first aggregated and then classified according to formulas described in CDL. Experimental results performed in a Smart Home environment are presented and discussed.

This paper describes ANSER, a system designed to perform surveillance in civilian airports and similar wide outdoor areas. Whereas an intelligent system - possibly controlled by a human supervisor - is able to integrate the information originating from different sources (i.e., fixed devices and sensors distributed throughout the environment) and to coordinate their behaviors in case of anomalies, the mobile robot is a significant part of the overall system: its main subsystems, i.e., autonomous surveillance, localization (performed using only a non-differential GPS and a laser rangefinder) and navigation are investigated in depth. Experimental results validate the robustness and reliability of the approach.

The paper focuses on the localization subsystem of ANSER, a mobile robot for autonomous surveillance in civilian airports and similar wide outdoor areas. ANSER localization subsystem is composed of a non-differential GPS unit and a laser rangefinder for landmark-based localization (inertial sensors are absent). An augmented state vector approach and an Extended Kalman filter are successfully employed to estimate the colored components in GPS noise, thus getting closer to the conditions for the EKF to be applicable.

The paper deals with the role of line features in mobile robot self-localization, when an extended Kalman filter is adopted for position tracking. First, a theoretical analysis is introduced, showing how the "length" of each extracted line (i.e., the number of the contributing range measurements) affects the localization accuracy. Second, a novel approach that takes into account the main findings of the theoretical analysis is considered. Finally, experimental results are used to validate the system.

This paper presents a novel navigation approach which, with minimal requirements in terms of on-board sensory, memory, and computational power, exhibits way-finding behaviors in very complex environments. The algorithm does not require any internal spatial representation, nor self-localization abilities: however, since it relies on heuristics to find a path to the goal, completeness is not guaranteed. The paper shows that this is the price to pay for augmenting the robustness of the system in presence of incomplete information and measurement noise

We propose a multiagent, distributed approach to autonomous mobile robotics which is an alternative to most existing systems in literature: robots are thought of as mobile units within an intelligent environment where they coexist and co-operate with fixed, intelligent devices that are assigned different roles: helping the robot to localize itself, controlling automated doors and elevators, detecting emergency situations, etc. To achieve this, intelligent sensors and actuators (i.e. physical agents) are distributed both onboard the robot and throughout the environment, and they are handled by Real-Time software agents which exchange information on a distributed message board. The paper outlines the benefits of the approach in terms of efficiency and Real-Time responsiveness.

In this paper we present a mobile system for visual tracking, i.e., a mobile platform equipped with a TV camera, which is capable of following a human leader along complex trajectories. Differently from existing systems, which mainly rely on the detection of colour blobs or particular features/markers, our system is based on the detection of motion through optical flow computation, thus being implicitly able to follow both persons and other robots with different characteristics in terms of colour, shape, etc.

Psychologists' debates on the role of knowledge to control actions in living beings have strongly influenced the research in the field of artificial intelligence and, consequently, of robotics. In both fields a focus of the debate has been the relevance (or even the presence) of a mental/internal representation in driving the course of actions of biological/artificial beings. In this paper we show that even very complex navigation tasks within maze-like environments can be carried out by an agent which needs to store just 'a handful of bytes' of internal representation about the world in which it is moving. The approach is called micronavigation, since it aims to capture the problem of mobile robot navigation in its entirety but with a minimalist approach.

We address the problem of autonomous navigation and localization in indoor environments, referring in particular to the specific scenario of service mobile robotics applications. The localization system uses active beacons (i.e. active transponders distributed throughout the building) as reference points; the estimate of the position of the robot and its uncertainty, both retrieved by correcting the estimate provided by odometry through an extended Kalman filter, are fed to the navigation system in order to help the robot to plan and execute target-oriented navigation tasks while showing a reactive behavior to handle the unpredictability of the environment.

Autonomous and semi-autonomous mobile robots have to perform a multiplicity of concurrent activities in order to carry out useful tasks in unstructured human-populated environments. Even if it is commonly accepted that a successful accomplishment of assigned tasks requires some sort of real time capability to quickly react and adapt to environmental changes, it is not clear which operating system support is best suited for the scheduling and synchronizing of concurrent activities with different timing requirements. This paper discusses this problem, comparing two different real time scheduling policies for autonomous robot applications: pre-emptive rate monotonic and non pre-emptive Earliest Deadline First (EDF). Experimental results are presented and evaluated.

Autonomous mobile robots have to perform a multiplicity of concurrent activities to carry out useful tasks while quickly reacting to sensorial inputs in a dynamic, partially unknown environment. The paper discusses the operating system requirements of a mobile robotic system, by focusing on the timing and communication requirements of the involved tasks. A distributed software architecture is proposed which implements a hybrid (pre-emptive/non-pre-emptive) task scheduling policy and a dedicated inter-task communication protocol, offering an efficient programming interface for the development of soft real-time robotic applications.

In this paper we discuss the architecture a reactive agent that copes with autonomous robot navigation problems, with partial knowledge of the environment, and capable of accepting problems due to limit cycles or deadlocks. The agent is part of a general multilevel cognitive framework and we focus here on the model of the agent and the navigation problems it is well suited for, with particular attention to those ones which require a knowledge-based approach.

In recent years, many different architectures have been proposed for autonomous robots that have to operate in a dynamic environment. These architectures address the problem of handling real time worlds for which classic paradigms have failed. However, little effort has been made to illustrate how different autonomous robots are supposed to cooperate in such an environment, despite the fact that most applications involving mobile autonomous vehicles require the activity of more than one entity. The paper addresses this particular problem. It presents an architecture for autonomous robots that have to or may collaborate to perform certain tasks. A practical system is also illustrated. It is in the advanced implementation phase, with prototypical vehicles operating in our department area, capable of collaborating and moving fluently even in crowded areas.

In this paper we propose a distributed architecture for intelligent robotic systems. The architecture is specific to this domain because it intends to provide support for the type of applications that share particular functional requirements: concurrent perception and action, task and plan execution, reasoning, "intelligent behaviours". The architecture also aims to improve re-usability and integration of different software components. It allows the transparent distribution of processes on different computers in a network to take advantage of the increased computational power and so overcome the vehicles on-board limitations. We focus on the related cognitive model on the internal structure of the architecture and of its components. We also examine in detail the EIE protocol we defined to exchange information within the distributed system. Finally we indicate some experimental results.

In this paper we propose a real-time architecture for autonomous mobile robots. We describe the single components of the architecture that are responsible of data acquisition and representation navigation and obstacle avoidance. In particular we focus on the integration of the components in the architecture and on the mechanisms that allow the system to operate in real time. Some experimental results are illustrated.

Active localization is the best cost-effective technique for mobile robotics in the real world: it allows accurate positioning in different environments with great flexibility and minimal environmental impact. In this paper different techniques to reach an accurate positioning with active localization systems are presented. Static and dynamic localization paradigms have been considered separately and solutions for both cases are presented. Simulation and experimental results in real environments are also discussed.

The traditional method for controlling the trajectories of an AGV in industrial environments is based on wire-guide systems. This paper describes a multi level architecture for mobile robots that integrates a high level mission planner, a real-time trajectory generator, and real-time motion control. The positioning accuracy is guaranteed by an active localization system that integrates in real-time the odometry measures giving position and orientation with high level resolution. This architecture allows the removal of the wire guidance system in AGV technology, and the introduction of an intelligent trajectory generation system at shop floor factory level.

A classification tool, based on the SOC (self-organizing classifier) neural model, is presented as an alternative solution to the problem of world modeling, aimed at navigation planning of an autonomous mobile robot. Starting from rough sensorial data, the knowledge about the explored environment of a mobile robot can be incrementally organized by means of self-organizing maps and a set of heuristic rules, avoiding the computational overhead due to classical geometric approaches to world modeling. The classification strategy realized, called SON (self-organizing navigation), allows to map neural information into symbols: the authors called such emergent symbols 'navigation situations'. The prototype has been successfully tested both with simulated and real data.

Strong deficiencies are present in symbolic models for action representation and planning, regarding mainly the difficulty of coping with real, complex environments. These deficiencies can be attributed to several problems, such as the inadequacy in coping with incompletely structured situations, the difficulty of interacting with visual and motorial aspects, the difficulty in representing low-level knowledge, the need to specify the problem at a high level of detail, and so on. Besides the purely symbolic approaches, several nonsymbolic models have been developed, such as the recent class of subsym-bolic techniques. A promising paradigm for the modeling of reasoning, which combines features of both symbolic and analogical approaches, is based on the construction of analogical models of the reference for the internal representations, as introduced by Johnson-Laird. In this work, we propose a similar approach to the problem of knowledge representation and reasoning about actions and plans. We propose a hybrid approach, symbolic and analogical, in which the inferences are partially devolved to measurements on analogical models generated starting from the symbolic representation. the interaction between the symbolic and the analogical level is due to the fact that procedures are connected to some symbols, allowing generating, updating, and verifying the mental model. the hybrid model utilizes, for the symbolic component, a representation system based on the distinction between terminological and assertional knowledge. the terminological component adopts a SI-Net formalism, extended by temporal primitives. the assertional component is a subset of first-order logics. the analogical representation is a set of concurrent procedures modeling parts of the world, action processes, simulations, and metaphors based on force fields concepts. A particular case study, regarding the problem of the assembly of a complex object from parts, is taken as an experimental paradigm. © 1993 John Wiley Sons, Inc.

One of the main problems in robotic research is planning. Different approaches have been considered, ranging from global to local planning. In the authors' approach, the key idea is that of planning movements of redundant robots, which involves both robotic and AI aspects. Pushing is used as a case study and sensorial feedback is considered in order to provide the planner with updated information on the dynamic evolution of the scene and to be able to deal with a priori unknown objects.>

The animation of actors and objects is considered in the framework of knowledge representation, with particular regard to motor knowledge. In our opinion, a formal model for motor knowledge should fill a gap between action planning and low level computer animation graphics; such model should be able to "virtualize" the actor and the interaction with the environment so that the planner could produce (and rely on) high level abstract actions, characterized by high autonomy and naturality. The paper discusses some general aspects about actions, actors, and scenes, and finally describes the NEM language, for the representation and animation of humanoids in a scene, which is meant to provide a software laboratory for experimenting with action schemas.

The aim of this paper is to outline a model of the eye-head system and some of its control strategies. To formulate such a model, the eye-head system was stimulated with random or periodic visual targets. Coupling between the eye and head motor commands was evident in the experiments performed, and this was also found to be the case for acoustic and tactile stimuli. Further experiments were performed to investigate whether the central nervous system (CNS) would be able to overcome the relative lack offlexibility shown by the model of the eye-head control system initially proposed. These experiments, in which it was required to execute independent eye-head movements towards two simultaneous targets (visual and acoustic), showed an impaired performance, indicating that an overcoming action by the CNS is not always possible. The findings are interpreted in terms of an interaction between a "hardware" low-level controller and a "software" high-level decisionmaker.