Deep Learning Aided Human-Machine Interaction Applications




Mundlamuri, Rahul

Journal Title

Journal ISSN

Volume Title



In an era dominated by rapid technological advancements, the interface between humans and machines is increasingly becoming a critical area of research. This doctoral dissertation delves into into machine learning-aided human-machine interaction contexts: ambient awareness data collection from sensors and other machine-originated sources and direct human-originated data from feedback, dialogs, etc. Machine-learning approaches apply to both these contexts and their integrations. More specifically, t he d issertation f ocuses o n human l ocalization t opics f or ambient awareness contexts and human-machine dialogs called chatbots for direct data contexts. The study is segmented into two main sections, each focusing on distinct technological advancements and their integration into everyday life. The first s ection of t he dissertation investigates human localization techniques, which have evolved significantly due to advancements in signal processing and machine learning. This study focuses on indoor localization methods, particularly WLAN fingerprinting and multipath a nalysis, by using channel s tate information(CSI), channel impulse response, and received signal strength(RSS) measurements, which have shown promising results in enhancing the accuracy and reliability of tracking individuals within complex indoor environments.

The second section explores human interaction technologies. Messaging technology limitations, including delays, message length, and encoding issues, constrain conversations' simplicity through most automated systems. These constraints prevent the development of more sophisticated interactive systems. This research seeks to overcome these barriers by developing strategies to enhance the design, implementation, testing, and scalability of automated text messaging systems that accommodate these limitations. The project's primary aim is to create a functional prototype of the system core that supports non-technical protocol programming for intervention projects. This will be achieved through a distributed state-machine architecture with autonomously programmed nodes and personalized user timers, utilizing both SMS and instant messaging channels for broader reach and functionality. This research also targets defining system requirements, architecture, and preliminary and critical design specifications. Establishing suitable data structures for different message types, such as broadcast, poll, and on-demand messages. Developing logic for triggering conditions to automate the sending of messages based on specific algorithms. Implementing a tailored, smart timer for each participant that schedules messages according to these trigger conditions. Connecting the complete messaging chain with messaging channel providers for seamless message transmission. Integrating the platform with framework-based chatbots to utilize their Natural Language Processing (NLP) capabilities through APIs, enhancing the interaction quality and user experience.


The full text of this item is not available at this time because the author has placed this item under an embargo until May 16, 2027.




Electrical and Computer Engineering