There is a growing interest in recent years, both from a researchers perspective and from a consumers point of view, to provide more natural, human-centred means of interacting with smart computing devices, such as cellphones, speakers, watches and IoT devices. Many different types of input are being tested and implemented: from vision-based body gesture recognition, to voice recognition, to proximity hand gesture recognition. In this project, we explore how the human-computer interaction can be expanded in the case of a relatively simple device that is already present in most homes: a lamp. We have first designed, then assembled and implemented a prototype of smart lamp that is capable of performing smart tasks such as: localising and connecting with other lamps which are placed in its vicinity; display many different illumination effects depending on the state of the lamp itself and of the whole network of lamps and recognising touch-less hand gestures that are performed on top of the lamp via an IR-based sensor. In this thesis, we will first present an overview of the state of the art on smart lamp devices and on the different implementations in the literature of various indoor localization algorithms and gesture recognition algorithms, with a particular focus on touch-less IR-based solutions. Afterwards we will describe our proposed solution, both the hardware (the external structure, internal components and design), and the software (the main flow chart, the various software managers and the most useful algorithms). Finally, we will test our solution in a real case scenario using two prototypes of smart lamp in a laboratory environment. The two lamp prototypes are capable of performing the localization and gesture recognition tasks that we planned to implement. In particular, the accuracy of the gesture recognition reaches 96.00% on the test dataset. Each gesture input can trigger various illumination effects, which include both the input lamp and the network of lamps, depending on the type of gesture.
Negli ultimi anni si è assistito ad un crescente interesse, sia da parte dei ricercatori che da parte dei consumatori, nella ricerca di modi più intuitivi e naturali di interazione uomo-macchina. Recentemente ne sono stati testati ed implementati molti tipi differenti. I principali che possiamo trovare, sia in commercio che in letteratura, appartengono principalmente alle seguenti categorie: riconoscimento visivo di movimenti effettuati con il corpo, riconoscimento vocale o riconoscimento di gesti eseguiti a distanza con le mani. In questo progetto abbiamo esaminato come espandere l’interazione che una persona può avere con un prodotto che è già presente nella nostra vita quotidiana: una lampada. Per prima cosa abbiamo progettato, poi assemblato ed implementato un prototipo di lampada intelligente in grado di: localizzare e connettersi a lampade dello stesso tipo nelle vicinanze in maniera del tutto autonoma; eseguire vari effetti luminosi a seconda del suo stato o di quello della rete di lampade e di riconoscere gesture a distanza, grazie ad un sensore IR posizionato nella base superiore. In questa tesi per prima cosa presentiamo una panoramica sullo stato dell’arte sulle lampade intelligenti che possono essere trovate in commercio e sulle differenti implementazioni di algoritmi di localizzazione in ambienti chiusi e per il riconoscimento di gesture a distanza, prestando particolare attenzione alle soluzioni con dispositivi IR. Successivamente abbiamo descritto sia la parte hardware (struttura della lampada e componenti elettroniche utilizzate) sia quella software della soluzione da noi implementata. Infine abbiamo testato il nostro lavoro in laboratorio utilizzando due prototipi. I due prototipi di lampada intelligente sono in grado di localizzarsi e di riconosce gesture come da progetto. In particolare, abbiamo ottenuto un’accuratezza di riconoscimento delle gestures del 96.00% sul dataset di test. Ogni gesture che viene eseguita può attivare svariati effetti luminosi che possono avere effetto sia sulla lampada sia sull’intero gruppo di lampade.
Study of a network of connected gesture-driven smart lamps
BOSSI, LUCA
2018/2019
Abstract
There is a growing interest in recent years, both from a researchers perspective and from a consumers point of view, to provide more natural, human-centred means of interacting with smart computing devices, such as cellphones, speakers, watches and IoT devices. Many different types of input are being tested and implemented: from vision-based body gesture recognition, to voice recognition, to proximity hand gesture recognition. In this project, we explore how the human-computer interaction can be expanded in the case of a relatively simple device that is already present in most homes: a lamp. We have first designed, then assembled and implemented a prototype of smart lamp that is capable of performing smart tasks such as: localising and connecting with other lamps which are placed in its vicinity; display many different illumination effects depending on the state of the lamp itself and of the whole network of lamps and recognising touch-less hand gestures that are performed on top of the lamp via an IR-based sensor. In this thesis, we will first present an overview of the state of the art on smart lamp devices and on the different implementations in the literature of various indoor localization algorithms and gesture recognition algorithms, with a particular focus on touch-less IR-based solutions. Afterwards we will describe our proposed solution, both the hardware (the external structure, internal components and design), and the software (the main flow chart, the various software managers and the most useful algorithms). Finally, we will test our solution in a real case scenario using two prototypes of smart lamp in a laboratory environment. The two lamp prototypes are capable of performing the localization and gesture recognition tasks that we planned to implement. In particular, the accuracy of the gesture recognition reaches 96.00% on the test dataset. Each gesture input can trigger various illumination effects, which include both the input lamp and the network of lamps, depending on the type of gesture.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi_definitiva.pdf
non accessibile
Descrizione: Testo della tesi
Dimensione
9.69 MB
Formato
Adobe PDF
|
9.69 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/10589/152267