Voice control as a potential method of human-computer interaction is steadilyincreasing, most recently with watches, cars and household appliances. Theelderly should be in focus for more digital development since they aresensitive to bad design. The purpose of the study therefore intends to includethis group in a survey of how a voice control system for beds can be formedfor the residents of a retirement home.  The aim was to study selected components of a system that the elderly couldinteract with using their voice to control their adjustable beds. Limitationswere conducted mainly in the technical aspects of this system by excluding theuse of a technical prototype. The survey was performed using two methods.First, residents of a retirement home were individually interviewed to confirmand strenghten the purpose of the study, that they would benefit of having avoice controlled system to complement the current handheld controller ofeach of their beds. The results of the interviews lay the foundation for thesecond part of the study: user tests of a simulated system. In these tests,selected parts of a hypothetical voice control system were evaluated forusefullness. The results indicate that participants valued simplicity over complexfunctions in a voice control system. In the feedback test, participantspreferred voice over sound as information-carrying. Thus, voice control ofbeds should be developed with a simple design that is conceptually the same,or very similar, to the participants mental model of the handheld controller.

With the rapid development of computers and technology, virtual assistants (VA) are becoming more and more common and intelligent. However, virtual assistants, such as Apple's Siri, Amazon's Alexa, and Google Assistant, do not currently have any physical functions. As an important part of the internet of things (IoT), the field of robotics has become a new trend in the usage of VA. In this project, a mobile robot, Fetch, is connected with the Amazon Echo Dot through the Amazon web service (AWS) and a local robot operation system (ROS) bridge server. We demonstrated that the robot could be controlled by voice commands through an Amazon Alexa. Given certain commands, Fetch was able to move in a desired direction as well as track and follow a target object. The follow model was also learned by Neural Network training, which allows for the target position to be predicted in future maps. / Master of Science / Nowadays, virtual personalized assistants (VPAs) exist everywhere around us. For example, Siri or android VPAs exist on every smartphone. More and more people are getting household Virtual Assistants, such as Amazon Alexa, Google Assistant, and Microsoft's Cortana. If the virtual assistants can connect with objects which have physical functions like an actual robot, they will be able to provide better services and more functions for humans. In this project, a mobile robot, Fetch, is connected with the Echo dot from Amazon. This connection allows us to control the robot by voice command. You can ask the robot to move in a given direction or track and follow a certain object. In order to let the robot learn how to predict the position of the target when the target is lost, a map is built as an influence factor. Since a designed algorithm of target position prediction is difficult to implement, we opted to use a machine learning method instead. Therefore, a machine learning algorithm was tested on the following model.

Individuals with upper limb motor deficits due to neurological conditions, such as stroke and traumatic brain injury, may exhibit hypertonia and spasticity, which makes it difficult for these individuals to open their hand. The Hand Orthosis with Powered Extension (HOPE) Hand was created in 2018. The performance of the HOPE Hand was evaluated by conducting a Box and Blocks test with an impaired subject. Improvements were identified and the HOPE Hand was mechanically redesigned to increase the functionality in performing grasps. The original motor configuration was reorganized to include active thumb flexion and extension, as well as thumb abduction/adduction. An Electromyography (EMG) study was conducted on 19 individuals (10 healthy, 9 impaired) to evaluate the viability of EMG device control for the specified user group. EMG control, voice control, and manual control were implemented with the HOPE Hand 2.0 and the exoskeleton system was tested for usability during a second Box and Blocks test.

Master of Science / Department of Computer Science / Mitchell L. Neilsen / Smart glasses, like smart phones, have separate operating systems, and can execute many different kinds of software and games. Smart glasses can be used to add a schedule, map navigation, interact with friends, take photos and videos, and make video calls with friends through voice control. They can support wireless network access through a mobile communication network. Bluetooth is a radio technology that supports short-range communication between of the devices. It can exchange information between devices including mobile phones, wireless headsets, laptops, etc. Bluetooth technology can effectively simplify the communication between mobile devices. This thesis focuses on smart glasses applications for high-throughput phenotyping which requires a data monitor, data synchronization, Bluetooth service, and voice control between devices. On the Android side, the application, which is extended, is called Field Book. The new software called Field Book AR, includes a data monitor module and a Bluetooth server module to achieve data exchange with smart glasses. On the smart glasses side, the application is called DataReceiver. It receives voice commands from users and controls the actions of Field Book AR. Also, when Field Book detects data changing, it accepts new data and shows changes to the users.

GATA is a web-application used by truckdrivers transporting timber from assigned pick-up places in the forest to specific receiving locations. GATA is used to facilitate and ease the navigation, communication and planning behind each delivery. This create situations where GATA needs to be used while driving, this is both dangerous and illegal. To solve this problem a voice-controlled proof-of-concept application for the most common features used during driving was developed and given the name GATA-Assistenten. The application was designed with a concept of having a few basic features built upon a stable platform. Dialogflow was chosen as tool to create the AI-voice control and to get it integrated with Google Assistant. Two tracks were created, one for drivers to change the estimated time of arrival and one to get the status on the receiving sites. According to statics at least 0,5% of all traffic related accidents can be directly contributed to interacting with communication devices, this is without taking the unrecorded cases into account. The conclusion based on these statistics is that GATA-Assistenten can not only help avoid accidents but also save lives. If a company wants to invest in safety, it is paramount to use voice-control, however it is also important to do research and development on new technologies to find out and fix the causes behind accidents. / GATA är en webbapplikation som används av lastbilschaufförer vid frakt av timmer från skogarna till mottagningsplatser för att lättare navigera, kommunicera och planera körningar på uppdrag av SCA. Det innebär att chaufförerna ibland behöver använda GATA under färd, något som inte bara kan vara en trafikfara utan även är olagligt. För att lösa detta problem skall ett proof-of-concept på röststyrning utvecklas för några av de vanligaste funktionerna som används under färd i GATA. Röststyrningsapplikationen som fick namnet GATA-Assistenten designades utifrån ett koncept om att ha få grundläggande funktioner byggt på en stabil plattform. Dialogflow valdes som verktyg för att skapa en AI-baserad röststyrning med integration via Google Assistant. GATA- Assistenten bestod slutligen av två huvudspår, det ena för att ändra den av chauffören angivna ankomsttiden till mottagningsplatsen och det andra för att ta reda på aviseringsläge på mottagningsplatserna. Enligt statistik är minst 0,5% av alla olyckor direkt orsakade av en förares interaktion med någon form av kommunikationsutrustning, detta utan att ta mörkertalet i beaktande. Slutsatsen som kan tas av detta är att GATA-Assistenten kan hjälpa till att undvika olyckor och rädda liv. Vill ett företag satsa på säkerhet är röststyrning ett måste, om det finns en nollvision kring olyckor i trafiken måste dock mer forskning och utveckling läggas på ny teknik för att ta reda på samt åtgärda orsakerna bakom alla olyckor.

Lindningsstationen på ABB Power Transformers i Ludvika är idag styrd med fotpedal, som på senare år fått stor kritik kring dess ergonomiska utformning och risk för belastningsskador. I en studie år 2011 testades ett system med röststyrning som bemöttes positivt inom det ergonomiska perspektivet, men negativt inom dess prestanda och bristande säkerhetsklassning. Strax efteråt lades konceptet på is, till vilket företaget nu villee se vart tekniken kring röststyrning och assisterande lösningar ligger idag. Syftet med denna rapport är att skapa ett jämförande underlag för vidare investeringsbeslut av olika lösningar relaterade till industriell röststyrning och assisterande lösningar, vilket inkluderar hur de kan appliceras på den nuvarande lindningsmaskinen. Genom en litteraturstudie och intervjuer skapades 1) en lista med dem kundbehov som lösningen skulle uppfylla samt en lista med kundbehov som lösningen var tvungen att uppfylla samt 2) produktkrav som en trådlös nödstoppslösning var tvungen att följa. Dessa ingångsparametrar användes sedan i en marknadsundersökning där olika företag och system kring röststyrning, assisterande lösningar och nödstopp jämfördes.,  där tre Tre förslag på röststyrning och två nödstopp gick vidare för analys kring konceptuell implementering. Slutsatsen av arbetet blev att tekniken kring industriell röststyrning fortfarande är i en utvecklande fas, där en lösning idag kräver en kombination av flera mindre komponenter. De nödstoppslösningar som identifierades kan dock inköpas och implementeras idag. Om företaget väljer att investera i en lösning inom en snar framtid, rekommenderas en uppgradering av den tidigare lösningens hårdvara från Honeywell. Är företaget villiga att lägga lite mer tid och resurser på lösningen rekommenderas Voice INTER Connect:s lösning. Alternativt avvakta tills ITSpeex kan erbjuda en fullständig lösning. / The winding station at ABB Power Transformers in Ludvika is today controlled using a foot pedal, which in recent years has received a lot of criticism for its poor ergonomic design and risk for repetitive strain injuries. In a study performed in 2011, a voice control system was tested, and positively received from an ergonomic perspective, but negatively when it came to both performance and proper safety. Shortly afterwards, the concept was put on hold, to which the company now wants to see where the technology surrounding voice control and assistive solutions are today. The purpose of this report is to create a comparative foundation for future investment decisions, about different solutions related to industrial voice control and assistive solutions, which includes how the solutions are applied at the current winding station. Through a literature study and interviews,1) a list of customer needs was made, which the voice solution had to fulfill, and 2) product requirements which the wireless emergency stop had to follow. These were later used in a market research where different companies and systems involving voice control, assistive solutions and emergency stops were compared. Three proposals for voice control and two for emergency stop moved on to further analysis about conceptual implementation. The conclusion drawn from the study was that the industrial voice control technology is still under development, where a contemporary solution requires a combination of several smaller components. The identified emergency stops solutions can, however, be purchased and installed today. If the company chooses to invest in a solution within the near future, an upgrade of the previous solution from Honeywell is recommended. With time and resources available, a solution from Voice INTER Connect is the one to pick. Alternatively hold off until ITSpeex can offer a complete solution.

To order pick-up food by using your computer or phone is nothing new. Food delivery companies such as FoodHero and Uber Eats along with many other around the world base their entire company idea around the food order and delivery process. For a company to stand out in such a vast market can sometimes be quite tricky. Sometimes your company needs a niche to stand out in the crowd. This project aims to create such a niche in an order system prototype based on voice-controlled systems and conversation. This prototype allows users to place food orders through only the use of natural speech and a voice assistant. The prototype utilizes products and services from both Amazon and Google to create the order system structure. The ordering system also takes advantage of the serverless architecture that both Amazon and Google provide. The end result of this project is a simple, convenient, and user-friendly prototype

The field of Artificial Intelligence (AI) has seen much development in recent years. This is mostly because of new technologies and computers becoming faster. With actors like Microsoft releasing software to reduce the complexity of the development of AI, the barrier to entry is now lower than ever. Due to this lower barrier, one area that has gained much traction is chat bots, being able to act as first line support for companies. While the technology is evolving fast and updates to Microsoft’s tools are released at an impressive rate, it appears to be difficult for other areas like documentation and ease of use to keep up with this pace. This thesis will explore some tools from Microsoft used in AI development. Areas covered are the Cognitive Services with a focus on speech, the Microsoft Azure Bot Service, the QnA Maker and the bot framework’s web-based client. These tools will be evaluated in the context of a chat bot, taking into account their functionality and development. The result is a chat bot that uses a knowledge base for data storage & answer logic, and a web interface for chat functionalities using the bot framework web-client.

Register and pitch are two distinct perceptual entities of the human voice. Without clear evidence for the use of the terminology, sources have begun to refer to lighter, or "falsetto", register as being "cricothyroid dominant" and heavier, or "chest", register as being "thyroarytenoid dominant" (Hirano, 1987; Miller, 1996; McCoy, 2004; Henrich, 2006; Spivey, 2008; Edwin, 2008; Phillips, Williams, & Edwin, 2012). The same intrinsic laryngeal musculature (i.e. the cricothyroid, CT, and thyroarytenoid, TA, muscles) play a role in the control of both register and pitch. Higher-pitched phonation, typically used to produce falsetto register, is mediated primarily by the cricothyroid (CT) muscle. The thyroarytenoid (TA) muscle plays a larger role in controlling lower-pitched voicing, the pitch range in which chest register tends to be used (Titze, 1989b; Shipp and McGlone, 1971). Despite their frequent co-occurrence, high and low pitched phonation are not controlled in the same way as light and heavy register productions. The purpose of this study was to examine the ratio of CT and TA muscular activity in the control of chest and falsetto registers. Data were collected from untrained voice users: four females and one male. Hooked-wire electrodes were inserted into both the CT and TA muscles of each participant in order to collect electromyographic (EMG) data during glissando from low to high pitch on the vowel /i/ twice per subject, and tasks eliciting maximal activation of CT and TA muscles. A trained singing instructor with 17 years of experience determined and recorded the occurrence of register transition during each glissando. CT and TA EMG activity data from the glissando were normalized relative to maximum elicited CT and TA EMG activity, and were then retrospectively analyzed. CT muscular dominance was defined as a ratio of percentage of maximum CT EMG activity to percentage of maximum TA EMG activity greater than 1 (i.e. CT:TA greater than 1). TA dominance is a ratio of CT:TA activity less than 1 (i.e. CT:TA less than 1). During glissando, all subjects experienced register transition from chest to falsetto register. In all subjects, the majority of chest register, and all of falsetto register, was produced with CT muscular dominance. Only the 3-4 lowest semitones, on average, in chest register were TA dominant. The transition from chest to falsetto register consistently did occur when the CT muscle was dominant, however, register transition did not occur as CT muscle activity began to dominate TA muscle activity. Results of the study showed that CT muscular dominance did not define falsetto register, nor was chest register defined by the TA muscular dominance.

Magistro darbe pristatoma sukurta programa, realizuojanti interneto naršyklės valdymą balsu. Ši programa papildo atskirų žodžių prototipinę atpažinimo sistemą, pagrįstą paslėptaisiais Markovo modeliais (PMM). Šios dvi dalys ir sudaro interneto naršyklės valdymo balsu prototipą, kuris gali atpažinti 71 komandą (vienas arba du žodžiai) lietuvių kalba: 1 komandą, skirtą naršyklės atvėrimui, 54 naršyklės valdymo komandas, 16 komandų, atveriančių konkrečius iš anksto sistemai nurodytus tinklalapius. Darbe aprašytas lietuvių kalbos atskirų žodžių atpažinimo sistemos akustinių modelių, grįstų paslėptaisiais Markovo modeliais, rinkinių eksperimentinis tyrimas. Atsižvelgiant į įvairius atpažinimui turinčius įtakos veiksnius (mokymo duomenų kiekį, mišinio komponenčių skaičių, kalbėtojo lytį, skirtingos techninės įrangos naudojimą atpažinime), buvo sukurti skirtingi balso komandų akustinių modelių rinkiniai. Eksperimentinio tyrimo metu buvo tiriama šių rinkinių panaudojimo atpažinimo sistemoje įtaka sistemos atpažinimo tikslumui. Eksperimentinio tyrimo rezultatai parodė, kad interneto naršyklės valdymo balsu sistemos prototipo atpažinimo tikslumas siekia 98%. Sistema gali būti naudojama kaip vaizdinė priemonė vyresniųjų klasių moksleiviams informacinių technologijų, fizikos, psichologijos, matematikos pamokose. / The thesis presents a prototype of the software (system) for Web browser control by voice. The prototype consists of two parts: the Hidden Markov Models based word recognition system and the program, which implements browser control by voice commands and is integrated in the word recognition system. The prototype is a speaker-independent Lithuanian word (voice commands) recognition system and can recognize 71 voice commands: 1 command is intended to run browser, 54 commands – for browser control, and 16 commands – to open various user predefined websites. Taking into account various factors (amount of training data, number of Gaussian mixture components, gender of speaker, use of different hardware for recognition) which have impact on recognition, different sets of acoustic models of Lithuanian voice commands were created and trained. An experimental investigation of the influence of the sets usage in Lithuanian word recognition system on the word recognition accuracy was performed. The results of the experimental investigation showed that created prototype system achieves 98% word recognition accuracy. The prototype system can be used at secondary school as a visual speech recognition learning tool in the informatics, physics, psychology, and mathematics lessons for the pupils of senior classes.