Requirements Engineering, SE-4351.

001
Assignment: Evolving Project Plan
Due: Thursday, March 28, 2024

We are doing a project to develop a phone app for the blind to navigate indoor areas called
“InSight”.

Team Name: InSight

Team Members:

• Sayyid Safaoui. UTD_ID: Sjs200008

• Tsion Yigzaw. UTD_ID: Tgy220000
• Vishwa Pandian. UTD_ID: Vrp200001
• Saishrey Bhandare. UTD_ID: Ssb200013

Team website URL: https://softwareengineeringgroup.weebly.com

Tools:

We will be using discord for voice meetings and Microsoft teams if we need to
record something and/or if we need cameras.

Meeting time: TBD

Team Deliverable Date Team Lead Tools

(Rotating)
Preliminary Project Plan February 15, 2024 Tsion Yigzaw • Weebly
(Thursday) • MS Word
• Discord

Project I presentation & March 27, 2024 Sayyid Safaoui • PPT

submission (Tuesday) + • Discord
 (WRS-document + March 29, 2024 • Word
Questionnaire + PPT + (Thursday) • MS Teams
evolving Project Plan) • Statistics Tool
(mentioned on the
syllabus)
Final Project II April 30, 2024 Vishwa • PPT
submission, (Tuesday) Pandian • Discord
presentation and demo + • Word
May 2, 2024 • MS Teams
(Thursday) • Statistics Tool
(mentioned on the
syllabus) (same as
used in Project I)
• Printer

Team Signatures:

Sayyid Safaoui:

Tsion Yigzaw:

Vishwa Pandian:

Saishrey Bhandare:
Objective: Create a user-friendly smartphone app that utilizes the built-in camera(s) and
advanced AI to provide real-time, narrated guidance, helping blind and visually impaired
users navigate indoor environments with greater confidence and autonomy.

Assumptions: World: an indoor environment with moderate foot traffic. Different services
are available for the user within the environment.

We assume that:

• the user is able to hear and understand instructions the app provides
• the user is able to use voice commands to instruct the app
• the system is precise with tracking the user’s location
• the user will have a stable internet connection for accurate real-time data and GPS
functionality
• the user has a device compatible with the system

Project Scope

• Core Features:
o Object Recognition: Accurately detect and identify objects (furniture, doors,
stairs, obstacles).
o Obstacle Warning: Issue warnings about potential obstacles in the user's
path.
o Spatial Mapping: Build a basic understanding of the room's layout.
o Text Recognition: Read signs, labels, and instructions out loud.
o Navigation: Provide turn-by-turn directions and describe hallways or
corridors.
• Additional Features (Potential Expansions):
o Facial Recognition: Identify familiar people in the environment.
 o Location Integration: Integration with indoor mapping systems (if available
in specific buildings).

Stakeholders:

External:

• End users: Blind or visually impaired people are the primary users of the app
• Organizations for blind and visually impaired people: Institutions that provide
support for visually impaired people and promote the app
• Family members and emergency contacts: Individuals who will be contacted in the
case of an emergency

Internal:

• Project team: Team Sight designed and developed the application

WRSPM Model:

World - There is an internet connection stable enough to provide accurate GPS

functionality. The user must be able to provide the app with commands.

Requirements - A system that navigate users through an indoor environment, as well as

alert emergency contacts or services when needed.

Specification - The system will use GPS tracking guide the user to the desired location. It
will also use algorithms that generate the best possible route. It will utilize the device’s
cameras and sensors to identify obstacles.

System (problem, machine) - The system is an app that uses the device’s cameras and
sensors to guide the user, and utilizes microphones to process the user’s instructions as
speech.
Functional Requirements:

• Real-time Object Recognition: The app should accurately identify objects within
the user's surroundings (e.g., chairs, tables, doors, people, vehicles).
• Text Reading: Detect and read aloud text from various sources (e.g., signs, labels,
menus, documents).
• Navigation Assistance: Provide verbal directions to help the user navigate their
surroundings, indicating obstacles, pathways, and points of interest.
• Scene Description: Offer a general description of the environment to give the user a
holistic picture of their location.
• Facial Recognition (Optional): Identify familiar faces stored in the app's database
and announce their presence.
• Customization: Allow the user to adjust the voice narration speed, volume, and
tone.

Non-Functional Requirements:

• Accuracy: The object recognition, text reading, and navigation features should have
a high degree of accuracy to ensure reliability.
• Speed: Information processing and narration should be near real-time to provide a
seamless user experience.
• Offline Functionality: The core features (object recognition, basic navigation)
should work without an internet connection.
• Power Efficiency: The app must be optimized to minimize battery drain during
prolonged use.
• Usability: Have an intuitive interface with easily accessible controls. It may
consider voice commands for added convenience.
 • Privacy: Protect user data and any images captured with strict security measures.
Image data should not be saved or transmitted unless explicitly permitted by the
user.

Project Plan

1. Research & Feasibility Study

• Existing Technologies: Examine current indoor navigation apps and assistive

technology for the blind. Analyze their strengths, weaknesses, and relevant
techniques.
• Image Recognition Algorithms: Research state-of-the-art object detection and
scene understanding algorithms suitable for real-time processing on smartphones.
• Technical Feasibility: Assess computational requirements and battery life
implications for continuous use.

2. Development

• Technology Stack: Choose a suitable programming language (Python, Swift,

Java/Kotlin), development environment, and computer vision libraries (OpenCV,
TensorFlow, etc.).
• Prototyping: Build a basic app prototype showcasing core features (simple object
recognition, verbal feedback).
• User Interface (UI) Design: Prioritize aural interface and voice commands.
Carefully consider non-visual input and output modalities.
• Testing & Feedback: Early and frequent user testing with visually impaired
individuals is crucial for iterative development and ensuring usability.

3. Algorithm Refinement

• Dataset Collection: Gather a large, diverse dataset of indoor images with labeled
objects, environments, and signs to train the models.
 • Model Training & Improvement: Iteratively train and refine the object detection and
scene understanding algorithms for robustness in different lighting conditions and
environments.

4. App Usability & Accessibility

• Voice Commands: Develop a robust, intuitive voice command system for user
control.
• Audio Feedback: Optimize the clarity, timing, and prioritization of the verbal
guidance.
• Customization: Allow users to tailor the level of detail in descriptions and
personalize the app's voice.
• WCAG Compliance: Rigorously adhere to Web Content Accessibility Guidelines
(WCAG).

5. Pilot Testing

• Partner with Organizations: Collaborate with schools for testing and feedback in
real-world use cases.
• Collect User Data: Analyze usage patterns, identify common challenges, and
gather suggestions for improvement.

6. Deployment

• App Stores: Release the app on major app stores (Apple App Store, Google Play
Store).
• Marketing & Outreach: Engage with organizations for the visually impaired to
promote the app and inform the target audience.
• Maintenance & Updates: Provide ongoing support, bug fixes, and updates based
on user feedback.
Important Considerations:

• Privacy: Emphasize user privacy and data security.

• Power Consumption: Optimize algorithms and app design to minimize battery
drain.