Alberto Quattrini Li

PoliSpell: An Adaptive Spellchecker and Predictor for People with Dyslexia

People with dyslexia often face huge writing difficulties. Spellcheckers/predictors can help, but the current systems are not appropriate for them, because of the assumptions behind the models and because of heavy-to-use interfaces. This paper presents a system for spellchecking/predicting words, which can adapt both its model and its interface according to the individual behavior. The model takes into account typical errors made by people with dyslexia, such as boundary errors, and the context for correcting real-word errors. The interface aims at reducing interaction with the user. The model and the interface are easily adaptable to general use.

How Much Worth Is Coordination of Mobile Robots for Exploration in Search and Rescue

Exploration of unknown environments is an enabling task for several applications, including map building and search and rescue. It is widely recognized that several benefits can be derived from deploying multiple mobile robots in exploration, including increased robustness and efficiency. Two main issues of multirobot exploration are the exploration strategy employed to select the most convenient observation locations the robots should reach in a partially known environment and the coordination method employed to manage the interferences between the actions performed by robots. From the literature, it is difficult to assess the relative effects of these two issues on the system performance. In this paper, we contribute to filling this gap by studying a search and rescue setting in which different coordination methods and exploration strategies are implemented and their contributions to an efficient exploration of indoor environments are comparatively evaluated. Although preliminary, our experimental data lead to the following results: the role of exploration strategies dominates that of coordination methods in determining the performance of an exploring multirobot system in a highly structured indoor environment, while the situation is reversed in a less structured indoor environment.

A System for Building Semantic Maps of Indoor Environments Exploiting the Concept of Building Typology

Semantic mapping of indoor environments refers to the task of building representations of these environments that associate spatial concepts with spatial entities. In particular, semantic labels, like ‘rooms’ and ‘corridors’ are associated to portions of an underlying metric map, to allow robots or humans to exploit this additional knowledge. Usually, the classifiers that build semantic maps process data coming from laser range scanners and cameras and do not consider the specific type of the mapped building. However, in architecture it is well known that each building has a specific typology. The concept of building typology denotes the set of buildings that have the same function (e.g., being a school building) and that share the same structural features. In this paper, we exploit the concept of building typology to build semantic maps of indoor environments. The proposed system uses only data from laser range scanners and creates a specific classifier for each building typology, showing good classification accuracy.

Evaluating the impact of perception and decision timing on autonomous robotic exploration

Autonomous robotic exploration of initially unknown environments is at the basis of several applications, including map building and search and rescue. Despite the many recent works on robotic exploration, an issue that has not been adequately addressed in the literature so far is the evaluation of the impact of the perception (for map update) and decision (about where to go next) timing on the behavior of an exploring robotic system. In this paper, we contribute to fill this gap by providing a quantitative experimental analysis of how frequencies of perception and decision influence the performance of an exploring mobile robot. Results, obtained with an experimental simulation framework (implemented and made publicly available) based on ROS and Stage, confirm the intuitive idea that the best performance is obtained with fast-paced perceptions and decisions, but also suggest some tradeoffs for the values of perception and decision frequencies in some settings.

Asynchronous multirobot exploration under recurrent connectivity constraints

In multirobot exploration under centralized control, communication plays an important role in constraining the team exploration strategy. Recurrent connectivity is a way to define communication constraints for which robots must connect to a base station only when making new observations. This paper studies effective multirobot exploration strategies under recurrent connectivity by considering a centralized and asynchronous planning framework. We formalize the problem of selecting the optimal set of locations robots should reach, provide an exact formulation to solve it, and devise an approximation algorithm to obtain efficient solutions with a bounded loss of optimality. Experiments in simulation and on real robots evaluate our approach in a number of settings.

Experimental Comparison of Open Source Vision-Based State Estimation Algorithms

The problem of state estimation using primarily visual data has received a lot of attention in the last decade. Several open source packages have appeared addressing the problem, each supported by impressive demonstrations. Applying any of these packages on a new dataset however, has been proven extremely challenging. Suboptimal performance, loss of localization, and challenges in customization have not produced a clear winner. Several other research groups have presented superb performance without releasing the code, sometimes materializing as commercial products. In this paper, ten of the most promising open source packages are evaluated, by cross validating them on the datasets provided for each package and by testing them on eight different datasets collected over the years in our laboratory. Indoor and outdoor, terrestrial and flying vehicles, in addition to underwater robots, cameras, and buoys were used to collect data. An analysis on the motions required for the different approaches and an evaluation of their performance is presented.

Shallow coral reef surveying by inexpensive drifters

Coral reefs exhibit the highest biodiversity in the ocean and are an extremely vulnerable ecosystem. Monitoring the state of the reefs is a tedious process performed by human divers which can be automated. This paper presents the use of several inexpensive drifting sensor nodes in order to reconstruct a visual mosaic of a shallow coral reef. The drifters produce geo-referenced visual data from a downward facing camera while floating above a shallow-water coral reef. The vision is augmented with inertial data enabling the recovery of the drifters attitude. A brief description of the drifters together with a framework to produce visual mosaics are discussed. Experimental results from a deployment over the Folkestone Marine Reserve in Barbados demonstrating the utility of our approach are presented.

The CATS Project

University students with learning or sensorial disability often face huge difficulties in accessing campus facilities and, specifically, lectures. Many universities over a wide range of support services to overcome such issues, but this is not always enough. This paper presents CATS, an ongoing research project involving three Italian universities, aiming to design and test technological solutions directed towards a better support to accessible lectures. By providing students with a set of experimental, advanced tools, the aim of the project is also to foster inclusive practices. The solutions described here share the principle of being adaptable to the real needs of the students, which are measured using ICF*, an adapted version of the WHO ICF model.

Moving From ‘How to go There?’ to ‘Where to go?’: Towards Increased Autonomy of Mobile Robots

Autonomous mobile robots have seen a widespread development in recent years, due to their possible applications (e.g., surveillance and search and rescue). Several techniques have been proposed for solving the path planning problem, in which a user specifies spatial targets and the robots autonomously decide how to go there. In contrast, the problem of where to go next, in which the targets themselves are autonomously decided by the robots, is largely unexplored and lacking an assessed theoretical basis. In this work, we make a step towards a framework for casting and addressing this problem. The framework includes the following dimensions: the amount of knowledge about the environment the robots have, the kind of that knowledge, the criteria used to evaluate the success of the decisions, the number of decision makers, and the possible adversarial nature of the settings. We focus on applications relative to exploration and patrolling.

Strategies for coordinated multirobot exploration with recurrent connectivity constraints

During several applications, such as search and rescue, robots must discover new information about the environment and, at the same time, share operational knowledge with a base station through an ad hoc network. In this paper, we design exploration strategies that allow robots to coordinate with teammates to form such a network in order to satisfy recurrent connectivity constraints—that is, data must be shared with the base station when making new observations at the assigned locations. Current approaches lack in flexibility due to the assumptions made about the communication model. Furthermore, they are sometimes inefficient because of the synchronous way they work: new plans are issued only once all robots have reached their goals. This paper introduces two novel asynchronous strategies that work with arbitrary communication models. In this paper, ‘asynchronous’ means that it is possible to issue new plans to subgroups of robots, when they are ready to receive them. First, we propose a single-stage strategy based on Integer Linear Programming for selecting and assigning robots to locations. Second, we design a two-stage strategy to improve computational efficiency, by separating the problem of locations’ selection from that of robot-location assignments. Extensive testing both in simulation and with real robots show that the proposed strategies provide good situation awareness at the base station while efficiently exploring the environment.