In this thesis, we summarize our empirical research as well as findings towards an implementation and perception of a live feedback posture correction tool. Our goal is to develop a tool that could help a sedentary user to maintain correct posture without having to evaluate by themselves if their sitting position is ergonomical. The focus of the thesis lies in the identification of a posture recognizing tool, how to make use of the posture recognizing tool to evaluate posture status, and how the user interface should be presented towards the user. We also, based on user experience, focus on the usability in terms of effectiveness, efficiency, satisfaction and user friendliness of said posture correction tool.We chose to evaluate how well 3 different posture detection algorithms would fit our implementation based upon a certain number of criteria. We used the algorithm to get joint position data and drew out vectors between every joint. With our vectors we could calculate the angles between limbs in the human body and by doing so, evaluate if the user has good or bad posture. We conducted experiments on 7 participants towards our system, 3 of them are UX experts and 4 of them had no previous experience in UX design, after each experiment we provided the participant with a survey to share their experiences. The survey was focused on usability and user experiences. We then drew an analysis from our empirical research to form a conclusion.We managed to create a fully functional prototype which accurately could determine the participants posture. Out of our 3 different posture detection algorithms, we selected Openpose. We used their unity plugin and were able to calculate angles between limbs with unity’s built in functions towards vectors. From our testing, we got a lot of useful critique towards how the tool waspresented and explained towards the user, most of the feedback was towards the guidelines of the camera setup and its lack of guidance in that phase. We also received very positive feedback towards the tool and functionality itself, how responsive and accurate the system was for example.We have concluded that even though the tool we developed has some faults from a user perspective, we have come a long way to create something innovative and interesting. If there would be more time and resources to be spent towards the tool, this tool could definitely become big in the ergonomic world.

The Internet of Things (IoTs) has triggered rapid advances in sensors, surveillance devices, wearables and body area networks with advanced Human-Computer Interfaces (HCI). Neural Networks optimized algorithmically for high accuracy and high representation power are very deep and require tremendous storage and processing capabilities leading to higher area and power costs. For developing smart front-ends for ‘always on’ sensor nodes we need to optimize for power and area. This requires considering trade-offs with respect to various entities such as resource utilization, processing time, area, power, accuracy etc. Our experimental results show that there is presence of a network configuration with minimum energy given the input constraints of an application in consideration. This presents the need for a hardware-software co-design approach. We present a highly parameterized hardware design on an FPGA allowing re-configurability and the ability to evaluate different design choices in a short amount of time. We also describe the capability of extending our design to offer run time configurability. This allows the design to be altered for different applications based on need and also allows the design to be used as a cascaded classifier beneficial for continuous sensing for low power applications. This thesis aims to evaluate the use of Restricted Boltzmann Machines for building such re-configurable low power front ends. We develop the hardware architecture for such a system and provide experimental results obtained for the case study of Posture detection for body worn cameras used for law enforcement.

Med en ökad användning av digitala enheter i samhället ser vi också att fler drabbas av hållningsrelaterade hälsobekymmer. Ungdomar är särskilt drabbade även om de i många fall är omedvetna om det, men knappt ingen forskning fokuserar på den här målgruppen och deras behov. Den forskning som tagits fram är inte heller överens hur man på bästa sätt löser problematiken med hållning. Syftet med studien är därför att göra ungdomar mer medvetna om sin kroppshållning samt ta reda på hur digital design på bästa sätt kan motivera ungdomar att förbättra sin kroppshållning. För att ta reda på detta ställdes följande fråga: ”Hur kan en digital designlösning utformas för att skapa medvetenhet kring ungdomars kroppshållning, samt hjälpa att förbättra deras kroppshållning på ett underhållande och icke störande sätt?”. Studien genomfördes med metoden Konceptdriven Designforskning som inkluderar skapandet av en konceptuell prototyp förankrad i tidigare forskning. Prototypen består av en bärbar sensor som mäter hållning och en applikation i form av ett spel där användaren kan se sin hållning i realtid samt aktivt träna sin hållning. Prototypen värderades sedan genom en digital enkätundersökning. I resultatet framgick att designlösningen kunde tänkas användas av deltagarna, samt hjälpa att förbättra deras hållning eftersom den gjorde användaren mer medveten om sin kroppshållning. Många deltagare trodde att spelet kunde bidra med motivation och göra hållningsträningen mer underhållande men att fler spel-funktioner krävs för att behålla användarens intresse under en längre tid. Deltagarna ansåg att konceptet i sin helhet inte var störande men att mer tydlig feedback krävs för att indikera bra/dålig hållning. Utifrån detta resultat kan slutsatsen dras att konceptet i teorin kan skapa medvetenhet kring ungdomars kroppshållning och hjälpa de att träna den på ett underhållande sätt som inte är störande, men att mer studier och designarbete krävs för att komma fram till det bästa sättet att ge användaren feedback om sin hållning samt göra konceptet mer attraktivt att använda ur ett långsiktigt perspektiv. / With the increased use of digital devices in society, we can see that more people are affected by posture related health issues. Adolescents are especially affected although in many cases they are unaware of it, but hardly any research focuses on this group and their needs. The research that has been done also does not agree on how to best solve the problem with posture. The purpose of this study is therefore to make young people more aware of their posture and to find out how digital design can be used in the best way to motivate young people to improve their posture. To find this out, the following question was formed: “How can a digital design solution be formed to create awareness about young people's posture and help improve their posture in an entertaining and non-disturbing way”. The study was carried out with the method Concept driven design research which includes the creation of a conceptual prototype grounded in previous research. The prototype consists of a portable sensor that measures posture and an application in the form of a game where the user can see their posture in real time and actively train their posture. The prototype was then evaluated through a digital survey. The results showed that the participants could consider using the design solution and that it helped to improve their posture as it made the user more aware of their posture. Many participants thought that the game could contribute with motivation and make posture training more entertaining, but that more gaming features are required to keep the user's interest for a longer period. The participants felt that the concept as a whole was not disruptive, but that more clear feedback is required to indicate good/bad posture. Based on this result, it can be concluded that the concept in theory can create awareness of young people's posture and help them to exercise it in an entertaining way that is not disturbing, but that further studies and design work is needed to come up with the best way to give the user feedback on their posture and to make the concept more attractive to use from a long-term perspective.