João Guerreiro's 2020/2021 Thesis Proposals

This page outlines a set of research proposals you can pursue in the year 2020/2021 under my supervision (co-supervised by other professors and researchers). My research focuses on designing, implementing, and evaluating interactive systems that help people with different abilities overcoming accessibility barriers. To that end, I engage with users early on to understand how technology can have a positive (and real) impact on their lives according to their needs and preferences.

My prior work includes a broad set of technologies and research areas, ranging from Human-Computer Interaction and Accessible Computing, to Wearables, Ubiquitous Computing, Sensors, Virtual Reality, Human-Robot Interaction, among others. To see examples of my previous research you can check my publications page in my website or Google Scholar. If you have a research idea that fits my expertise, feel free to contact me so we can draft a new proposal for you!

You can also check the two videos below concerning two projects I have worked on recently at Carnegie Mellon University.

#1 Virtual Reality for Blind People

Advisors

João Guerreiro and André Rodrigues

Motivation

Virtual reality is an emerging technology that is slowly becoming available to the masses at affordable prices. VR is currently used in a variety of contexts: gaming, education, shopping, social spaces, employee training, to name a few. As with any emerging technology, it is fundamental we ensure its accessibility among people with different abilities. One of the major challenges blind people face in virtual environments is to navigate/move in the virtual space. While prior work has focused on mimicking real-world techniques, such as a virtual white cane (due to user familiarity), in virtual reality there are many locomotion techniques that vary greatly from application to application (e.g. free teleportation, walk in place, analog stick, directional dashes, waypoint navigation). In addition, blind users in virtual environments will not have the same restrictions as in the real world, nor the restrictions sighted people have due to a lack VR sickness (similar to motion sickness due to visual stimuli). We argue that this combination provides an opportunity to explore novel/fantastical mobility methods that are not possible otherwise.

What you will do

In this thesis you will be challenged to design, develop and evaluate novel navigation techniques in VR for blind people. You will conduct user studies early on to engage participants in co-design sessions ensuring user engagement and representation. This work will conclude with a user study evaluating the developed set of navigation techniques.

#2 Guide Dog (H)as a Sensor: Understanding the Owner-Dog Dynamics using Sensors

Advisors

João Guerreiro and Joana Campos

Motivation

Guide dogs are able to actively guide a blind person in order to avoid obstacles and dangerous situations. The blind person, however, is the one responsible to know the route to their destination and to communicate the actions the dog should perform. In addition to a few verbal commands, it is the constant owner-dog interaction through the harness that allows the pair to communicate and feel each other’s movements. This mutual understanding results in increased mobility, independence, and confidence for guide dog users.
Orientation and Mobility (O&M) training is essential for future guide dog users to learn how to interact with and provide commands to the dog. Still, the assessment of correct/incorrect exercises is exclusively made through observation by O&M specialists. Our goal is to use a sensor-based approach to detect specific actions of both the blind person and the guide dog, in order to support the O&M instructor by identifying correct and incorrect behaviour automatically. Such knowledge can be used both during and after O&M classes.

What you will do

You will design, develop, and evaluate a sensor-based solution (e.g., sensors in the harness, collar, person’s hand, or other options) that is able to detect specific movements made by the guide dog and/or the user. You will conduct user studies early on through co-design sessions with O&M specialists to understand the relevant communication mechanisms between a blind person and a guide dog. This work will conclude with a user study evaluating your solution for automatic detection of communication movements between the owner-dog pair.

#3 Teaching Orientation and Mobility Skills to Children with Vision Impairments

Advisors

João Guerreiro and Ana Pires

Motivation

Orientation and Mobility (O&M) can be defined as a set of concepts, skills and techniques that enable people with vision impairments to travel an environment safely and independently. Orientation refers to people's ability to position themselves in the environment, reflected in their awareness of where they are and where they want to go, while mobility refers to people's ability to move independently from one place to another in a safe, effective and efficient manner. These two interlinked concepts play a very important role in the lives of people with vision impairments in general and are extremely important to children, as the ability to travel independently provides access to a wide range of activities that enable people to participate in society.

What you will do

In this thesis, we aim to investigate current methodologies for O&M training and to design, develop and evaluate novel technological solutions to improve its effectiveness and engagement. We see opportunities for technology to further support O&M training activities both during and after classes with O&M specialists/teachers. Potential areas of research (you may suggest others) that can be explored in this thesis are (one or a combination of): You will conduct user studies early on to engage O&M specialists and blind children in co-design sessions ensuring user engagement and representation. This work will conclude with a user study evaluating the technological solution developed.

#4 I Can’t See You! Coping with Physical Distance in a ‘Social Distance’ Era when you’re Blind

Advisors

João Guerreiro and Tiago Guerreiro

Motivation

Blind people can navigate independently by using their orientation and mobility skills as well as their travel aids - the white cane or the guide dog. The white cane allows its users to identify obstacles/objects and open pathways in order to navigate safely, while the guide dog can follow the user's instructions while avoiding obstacles and dangerous situations. While extremely helpful, these two mobility aids fall short on supporting blind people when current instructions are to keep physical distance greater than two meters from each other: the white cane can only detect obstacles (or people!) at close distance, and a guide dog is completely unaware of social distancing guidelines.

What you will do

In this thesis, we aim to investigate the challenges that blind people face due to the COVID-19 pandemic, with a focus on their difficulties to keep physical distancing. You will conduct user studies early on to understand users’ needs and preferences in order to co-design technological solutions allowing them to keep physical distance while maintaining their autonomy and independence. Then, you will design, develop, and evaluate a technological solution that supports the needs of blind people in a Social Distancing Era.

#5 iRobot - Inclusive robots for children

Advisors

João Guerreiro and Ana Pires

Motivation

Robots have an enormous potential to support educational activities with children with mixed-abilities due to the sensing capabilities they can afford. Its predominant focus on visual feedback, however, limits its applicability for children with visual impairments. Our prior work with children with visual impairments and their educators identified the characteristics a Robot should have in order to be used by children with mixed-abilities, including: be easily heard and guide children throughout the activities; tactile cues to indicate head direction and sensors’ location; detect and give feedback on maps and objects; rewarding children’s’ actions; among others. The goal of this thesis is to design and develop an inclusive robot (which may be based on existing robots) by allowing it to be felt, listened, inspected and followed.

What you will do

You will design, develop, and evaluate an inclusive robot for educational activities of children with mixed-abilities. You will conduct user studies early on through co-design sessions with children and educators in order to define requirements, aesthetics, tactile and auditory cues. Co-design sessions may include techniques such as crafting, role-play, fiction, among others. This work will conclude with a user study evaluating the robot and its functionalities.

#6 Using Mobility Data to Improve Navigation Assistance for People with Vision Impairments

Advisors

João Guerreiro and Joana Campos

Motivation

People with vision impairments are able to navigate independently by using their Orientation and Mobility (O&M) skills and their travel aids - white cane or guide dog. However, navigating independently in unfamiliar and/or complex locations is still a main challenge and therefore blind people are often assisted by sighted people in such scenarios. While navigation technologies such as those based on GPS (e.g., Google Maps) can help, their accuracy is still too low (e.g., around 5 meters) to fully support blind users when navigating in unknown locations. On the other hand, indoor locations do not support GPS and generally do not have a navigation system installed.

What you will do

Our goal is to use data from the crowd (other people who walked the same areas before) to learn more about the environment and be able to provide additional instructions to blind users. A possible approach is to use smartphone sensors to estimate possible paths after an individual enters a building. While the GPS may inform us that users are entering/inside a building, large amounts of data (crowd-based) from smartphone sensors may give us information about possible paths and obstacles to instruct blind users. For instance, one may learn that after the entrance there is a path going left and after approximately 10 meters there are stairs to go up one floor; and another path that goes forward (and so on). While the data from a single user may be erroneous, we plan to use data from many users to make such an approach more robust.