Internet of Things (IoT) systems are increasingly becoming important in several applications, ranging from mission-critical to consumer applications. In particular, they have reached a remarkable importance in sport services, where data are used as support for athletic performance improvement for single player or the overall team, or for studying new tactics and schemes. For these systems, an accurate localization is fundamental. The goal of this thesis is to design a localization system based on the UltraWide Band (UWB) technology to provide precise and accurate positioning information as well as an easy installation process for the final users. The targeted users for this system are amateur or semi-professional teams that desire an improvement of their performances. During the localization phase, users wear UWB devices called tags which are localized by reference UWB devices called anchors. Before the localization phase, the anchors deployed on the field and their positions are estimated from the system itself. The first contribution of this thesis consists in designing algorithms for achieving a precise localization of the anchors. In particular, issues regarding filtering and compensation of the antenna delay introduced by the UWB modules are addressed using calibration and outlier removal algorithms. Furthermore, a new auto-localization approach is developed to estimate the anchors’ positions from the ranging measurements provided by the UWB modules. For studying the optimal anchor configuration, a optimization procedure is devised in order to select the most suited anchors configuration and to maximize the tag positioning accuracy. The second contribution of the thesis regards the development of tag localization solutions. A number of positioning algorithms are devised and assessed. Considering the intrinsic dynamics present in sport applications, Bayesian filtering techniques show to be superior compared to snapshot least square ones. The experimental assessment of all proposed techniques are performed using synthetic data as well as real data extracted from an experimental campaign executed in the surroundings of the laboratory IoT Lab of Politecnico di Milano. The results shows that the proposed approach is able to satisfy the required level of accuracy for the considered application without compromising its simplicity of use.
I sistemi Internet of Things (IOT) sono diventati importanti in numerose applicazioni, sia missioni critiche che applicazioni quotidiane. In particolare hanno raggiunto notevole importanza in ambito sportivo, dove i dati collezionati sono utilizzati come supporto per migliorare le prestazioni del singolo giocatore o della squadra, o per studiare nuove tattiche e schemi. In questi sistemi, una localizzazione accurata `e fondamentale. L’obiettivo principale di questa tesi e’ di sviluppare un sistema di localizzazione basato sulla tecnologia Ultra WideBand (UWB) preciso e accurato ma anche semplice da installare per gli utenti finali. Gli utenti considerati dal sistema solo prevalentemente squadre dilettanti o semi-professionistiche che desiderano un miglioramento delle loro prestazioni. Durante ogni registrazione, gli utenti indossano dispositivi UWB chiamati tag che vengono localizzati dai dispositivi UWB detti ancore. Prima della registrazione, le ancore vengono disposte sul campo e le loro posizioni stimate dal sistema stesso. Il primo contributo di questa tesi consiste nello sviluppo di algoritmi che permettano una precisa localizzazione del sistema di ancore. In particolare, le problematiche relative al filtraggio e alla compensazione del ritardo introdotto dai moduli UWB vengono risolte attraverso l’utilizzo di metodi di calibrazione e rimozione degli outlier. Inoltre, un nuovo algoritmo di auto-localizzazione viene proposto per stimare le posizioni delle ancore dalle misure di distanze fornite dai moduli UWB. Per studiare la configurazione ottimale delle ancore, una procedura di ottimizzazione viene presentata che permette di trovare la migliore della configurazione delle ancore e massimizzare l’accuratezza della posizione dei tag. Il secondo contributo riguarda la localizzazione dei tag vera e propria. Diversi algoritmi di posizionamento sono stati sviluppati e comparati per selezionare quello con le migliori prestazioni. Considerando la dinamica intrinseca dei sistemi sportivi, i filtri Bayesiani si dimostrano superiori rispetto ai metodi snapshot. La valutazione sperimentale di tutte le tecniche proposte nella tesi `e stata eseguita usando sia dati simulati che dati estratti da una campagna sperimentale svolta nei pressi del laboratorio IoT Lab del Politecnico di Milano. I risultati ottenuti dal sistema finale mostrano che l’approccio proposto e’ in grado di soddisfare i livelli di accuratezza richiesti per l’applicazione considerata senza compromettere la sua semplicita’.
Ultra-wide band localization system for sport applications
Piavanini, Marco
2020/2021
Abstract
Internet of Things (IoT) systems are increasingly becoming important in several applications, ranging from mission-critical to consumer applications. In particular, they have reached a remarkable importance in sport services, where data are used as support for athletic performance improvement for single player or the overall team, or for studying new tactics and schemes. For these systems, an accurate localization is fundamental. The goal of this thesis is to design a localization system based on the UltraWide Band (UWB) technology to provide precise and accurate positioning information as well as an easy installation process for the final users. The targeted users for this system are amateur or semi-professional teams that desire an improvement of their performances. During the localization phase, users wear UWB devices called tags which are localized by reference UWB devices called anchors. Before the localization phase, the anchors deployed on the field and their positions are estimated from the system itself. The first contribution of this thesis consists in designing algorithms for achieving a precise localization of the anchors. In particular, issues regarding filtering and compensation of the antenna delay introduced by the UWB modules are addressed using calibration and outlier removal algorithms. Furthermore, a new auto-localization approach is developed to estimate the anchors’ positions from the ranging measurements provided by the UWB modules. For studying the optimal anchor configuration, a optimization procedure is devised in order to select the most suited anchors configuration and to maximize the tag positioning accuracy. The second contribution of the thesis regards the development of tag localization solutions. A number of positioning algorithms are devised and assessed. Considering the intrinsic dynamics present in sport applications, Bayesian filtering techniques show to be superior compared to snapshot least square ones. The experimental assessment of all proposed techniques are performed using synthetic data as well as real data extracted from an experimental campaign executed in the surroundings of the laboratory IoT Lab of Politecnico di Milano. The results shows that the proposed approach is able to satisfy the required level of accuracy for the considered application without compromising its simplicity of use.| File | Dimensione | Formato | |
|---|---|---|---|
|
thesis_piavanini.pdf
non accessibile
Dimensione
14.25 MB
Formato
Adobe PDF
|
14.25 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/173955