Tecniche automatiche di mapping e localizzazione in ambienti indoor

Il mapping di grandezze fisiche risulta estremamente importante, essendo in grado di fornire un adeguato supporto per la localizzazione e il monitoraggio di parametri ambientali sensibili. Nel caso indoor, in assenza di un sistema di localizzazione di riferimento analogo al GPS per il caso outdoor, sfruttando appieno le potenzialità della sensoristica a bordo degli smartphone, si è fatto progressivamente strada il mapping di grandezze fisiche quali, ad esempio, il segnale Wi-Fi e il campo magnetico terrestre. In questo caso il mapping, senza richiedere alcuna infrastruttura e coadiuvato dall'utilizzo di dispositivi portatili largamente diffusi ad uso quotidiano, rappresenta una soluzione relativamente recente ridefinibile come Mobile Crowd Sensing. Il MCS rappresenta un nuovo paradigma di servizio, volto a sfruttare l'interconnettività tra dispositivi portatili per effettuare misurazioni di caratteristiche ambientali in maniera automatizzata, aggregandole in un sistema cloud usufruibile ad una vasta comunità. Tuttavia , il considerevole flusso di dati generato, la variabilità temporale delle grandezze di interesse e il rumore insito nelle misurazioni costituiscono problematiche fondamentali per l'utilizzo e la gestione delle misurazioni effettuate. Per tali motivi l'attività di tesi ha previsto i seguenti obiettivi: (i) fornire una panoramica delle principali tecniche e tecnologie di localizzazione volta a motivare l'importanza del mapping di grandezze fisiche ambientali; (ii) individuazione di grandezze fisiche appetibili per la creazione di mappe affidabili e realizzabili nei contesti applicativi più disparati, sfruttando risorse già presenti nell'ambiente; (iii) sviluppo di un algoritmo statistico in grado di fornire una stima accurata dell'andamento spaziale della grandezza di interesse attraverso un numero limitato di misurazioni, mantenendo la compatibilità con processi MCS e una bassa complessità computazionale. L’algoritmo sviluppato è stato validato attraverso simulazioni e misurazioni svolte in ambienti reali. In particolare, prove sperimentali sono state effettuate nell’arena Vicon nei laboratori DEI dell’Università di Bologna, sede Cesena, concepita dal gruppo di ricerca Casy.

Mapping of indoor environments by robots using low-cost vision sensors

For robots to operate in human environments they must be able to make their own maps because it is unrealistic to expect a user to enter a map into the robot’s memory; existing floorplans are often incorrect; and human environments tend to change. Traditionally robots have used sonar, infra-red or laser range finders to perform the mapping task. Digital cameras have become very cheap in recent years and they have opened up new possibilities as a sensor for robot perception. Any robot that must interact with humans can reasonably be expected to have a camera for tasks such as face recognition, so it makes sense to also use the camera for navigation. Cameras have advantages over other sensors such as colour information (not available with any other sensor), better immunity to noise (compared to sonar), and not being restricted to operating in a plane (like laser range finders). However, there are disadvantages too, with the principal one being the effect of perspective. This research investigated ways to use a single colour camera as a range sensor to guide an autonomous robot and allow it to build a map of its environment, a process referred to as Simultaneous Localization and Mapping (SLAM). An experimental system was built using a robot controlled via a wireless network connection. Using the on-board camera as the only sensor, the robot successfully explored and mapped indoor office environments. The quality of the resulting maps is comparable to those that have been reported in the literature for sonar or infra-red sensors. Although the maps are not as accurate as ones created with a laser range finder, the solution using a camera is significantly cheaper and is more appropriate for toys and early domestic robots.

Progettazione e implementazione di una piattaforma per la localizzazione indoor basata su tecniche di radio fingerprinting e data-crowdsourcing

Questa tesi ha come scopo principale l'analisi delle diverse tecnologie di localizzazione in ambito indoor, analizzando in particolare l'utilizzo del Wifi RSS Fingerprinting. La tecnica del Wifi RSS Fingerprinting è una tecnica per la localizzazione all'interno di ambienti chiusi, che consiste nella definizione di un 'impronta'(fingerprint) in un punto preciso dell'ambiente(definito reference point), andando a inserire in un database i valori di potenza del segnale ricevuto(RSS) da ogni access point rilevato all'interno di quel determinato reference point. Per l'implementazione di questa tecnica è stato sviluppato un applicativo con un architettura client-server. Il client è stato sviluppato in ambiente Android, realizzando una applicazione per la gestione della fase di salvataggio di nuovi fingerprint e per la fase di localizzazione della posizione corrente, tramite l'utilizzo dei vari fingerprint precedentemente inseriti all'interno del DB. Il server, sviluppato in Node.js(framework Javascript), gestirà le diverse richieste ricevute dal client tramite delle chiamate AJAX, prelevando le informazioni richieste direttamente dal database. All'interno delle applicativo sono stati implementati diversi algoritmi per la localizzazione indoor, in modo da poter verificare l'applicabilità di questo sistema in un ambito reale. Questi algoritmi sono stati in seguito testati per valutare l'accuratezza e la precisione di ciascuno, andando ad individuare gli algoritmi migliori da utilizzare in base a scenari diversi.

Una piattaforma per la localizzazione indoor tramite crowdsensing

Questo elaborato studia e analizza il comportamento di tre algoritmi di base per quanto riguarda la localizzazione indoor e in particolare la tecnica del fingerprint. L'elaborato include l'analisi di come l'eterogeneita dei dispositivi possa influenzare gli algoritmi e la loro accuratezza nel produrre il risultato. Si include inoltre l'analisi dello stato dell'arte la progettazione e lo sviluppo di un'applicazione Android e di un web service. L'illustrazione dei test effettuati e le considerazioni finali concludono la tesi.

Gaussian process models for sensor-centric robot localisation

This paper presents an approach to building an observation likelihood function from a set of sparse, noisy training observations taken from known locations by a sensor with no obvious geometric model. The basic approach is to fit an interpolant to the training data, representing the expected observation, and to assume additive sensor noise. This paper takes a Bayesian view of the problem, maintaining a posterior over interpolants rather than simply the maximum-likelihood interpolant, giving a measure of uncertainty in the map at any point. This is done using a Gaussian process framework. To validate the approach experimentally, a model of an environment is built using observations from an omni-directional camera. After a model has been built from the training data, a particle filter is used to localise while traversing this environment

Laser-radar data fusion with Gaussian Process Implicit Surfaces

This work considers the problem of building high-fidelity 3D representations of the environment from sensor data acquired by mobile robots. Multi-sensor data fusion allows for more complete and accurate representations, and for more reliable perception, especially when different sensing modalities are used. In this paper, we propose a thorough experimental analysis of the performance of 3D surface reconstruction from laser and mm-wave radar data using Gaussian Process Implicit Surfaces (GPIS), in a realistic field robotics scenario. We first analyse the performance of GPIS using raw laser data alone and raw radar data alone, respectively, with different choices of covariance matrices and different resolutions of the input data. We then evaluate and compare the performance of two different GPIS fusion approaches. The first, state-of-the-art approach directly fuses raw data from laser and radar. The alternative approach proposed in this paper first computes an initial estimate of the surface from each single source of data, and then fuses these two estimates. We show that this method outperforms the state of the art, especially in situations where the sensors react differently to the targets they perceive.

Traversability estimation for a planetary rover via experimental kernel learning in a Gaussian process framework

A critical requirement for safe autonomous navigation of a planetary rover is the ability to accurately estimate the traversability of the terrain. This work considers the problem of predicting the attitude and configuration angles of the platform from terrain representations that are often incomplete due to occlusions and sensor limitations. Using Gaussian Processes (GP) and exteroceptive data as training input, we can provide a continuous and complete representation of terrain traversability, with uncertainty in the output estimates. In this paper, we propose a novel method that focuses on exploiting the explicit correlation in vehicle attitude and configuration during operation by learning a kernel function from vehicle experience to perform GP regression. We provide an extensive experimental validation of the proposed method on a planetary rover. We show significant improvement in the accuracy of our estimation compared with results obtained using standard kernels (Squared Exponential and Neural Network), and compared to traversability estimation made over terrain models built using state-of-the-art GP techniques.

Robust multiple-sensing-modality data fusion using Gaussian Process Implicit Surfaces

The ability to build high-fidelity 3D representations of the environment from sensor data is critical for autonomous robots. Multi-sensor data fusion allows for more complete and accurate representations. Furthermore, using distinct sensing modalities (i.e. sensors using a different physical process and/or operating at different electromagnetic frequencies) usually leads to more reliable perception, especially in challenging environments, as modalities may complement each other. However, they may react differently to certain materials or environmental conditions, leading to catastrophic fusion. In this paper, we propose a new method to reliably fuse data from multiple sensing modalities, including in situations where they detect different targets. We first compute distinct continuous surface representations for each sensing modality, with uncertainty, using Gaussian Process Implicit Surfaces (GPIS). Second, we perform a local consistency test between these representations, to separate consistent data (i.e. data corresponding to the detection of the same target by the sensors) from inconsistent data. The consistent data can then be fused together, using another GPIS process, and the rest of the data can be combined as appropriate. The approach is first validated using synthetic data. We then demonstrate its benefit using a mobile robot, equipped with a laser scanner and a radar, which operates in an outdoor environment in the presence of large clouds of airborne dust and smoke.

Validation-Based Sparse Gaussian Process Classifier Design

Gaussian processes (GPs) are promising Bayesian methods for classification and regression problems. Design of a GP classifier and making predictions using it is, however, computationally demanding, especially when the training set size is large. Sparse GP classifiers are known to overcome this limitation. In this letter, we propose and study a validation-based method for sparse GP classifier design. The proposed method uses a negative log predictive (NLP) loss measure, which is easy to compute for GP models. We use this measure for both basis vector selection and hyperparameter adaptation. The experimental results on several real-world benchmark data sets show better orcomparable generalization performance over existing methods.

Semi-Supervised Classification Using Sparse Gaussian Process Regression

Gaussian Processes (GPs) are promising Bayesian methods for classification and regression problems. They have also been used for semi-supervised learning tasks. In this paper, we propose a new algorithm for solving semi-supervised binary classification problem using sparse GP regression (GPR) models. It is closely related to semi-supervised learning based on support vector regression (SVR) and maximum margin clustering. The proposed algorithm is simple and easy to implement. It gives a sparse solution directly unlike the SVR based algorithm. Also, the hyperparameters are estimated easily without resorting to expensive cross-validation technique. Use of sparse GPR model helps in making the proposed algorithm scalable. Preliminary results on synthetic and real-world data sets demonstrate the efficacy of the new algorithm.

Tractable nonparametric bayesian inference in poisson processes with Gaussian process intensities

The inhomogeneous Poisson process is a point process that has varying intensity across its domain (usually time or space). For nonparametric Bayesian modeling, the Gaussian process is a useful way to place a prior distribution on this intensity. The combination of a Poisson process and GP is known as a Gaussian Cox process, or doubly-stochastic Poisson process. Likelihood-based inference in these models requires an intractable integral over an infinite-dimensional random function. In this paper we present the first approach to Gaussian Cox processes in which it is possible to perform inference without introducing approximations or finitedimensional proxy distributions. We call our method the Sigmoidal Gaussian Cox Process, which uses a generative model for Poisson data to enable tractable inference via Markov chain Monte Carlo. We compare our methods to competing methods on synthetic data and apply it to several real-world data sets. Copyright 2009.

Tractable nonparametric Bayesian inference in Poisson processes with Gaussian process intensities

The inhomogeneous Poisson process is a point process that has varying intensity across its domain (usually time or space). For nonparametric Bayesian modeling, the Gaussian process is a useful way to place a prior distribution on this intensity. The combination of a Poisson process and GP is known as a Gaussian Cox process, or doubly-stochastic Poisson process. Likelihood-based inference in these models requires an intractable integral over an infinite-dimensional random function. In this paper we present the first approach to Gaussian Cox processes in which it is possible to perform inference without introducing approximations or finite-dimensional proxy distributions. We call our method the Sigmoidal Gaussian Cox Process, which uses a generative model for Poisson data to enable tractable inference via Markov chain Monte Carlo. We compare our methods to competing methods on synthetic data and apply it to several real-world data sets.