KAVIKULGURU INSTITUTE OF TECHNOLOGY AND

SCIENCE
Ramtek, Nagpur, Maharashtra, India
(Managed by Vodithala Education Society, Hyderabad)
UGC AUTONOMOUS INSTITUTE

Project Preliminary Investigation Report (PPIR)

Name of Department:
INFORMATION TECHNOLOGY

Name of Project Guide:

Mrs. Saroj Shambharkar

Name of Project Co - Guide (if any):

------

Students Details:
Roll No. Name of Student Email ID Mobile No.
IT23015 Harshal Bhujade harshalbhujade13@[Link] 87889 41942
IT23010 Aryan Dhabale aryandhabale11@[Link] 78220 40752
ITD23071 Rohit Dhande rohitdhande79@[Link] 93222 82611
IT23002 Ansh Khadase anshkhadase21@[Link] 86983 98964
IT22015 Mohit mohitkamdi11@[Link] 93221 88133

Title of the Project:

A Voice-Controlled Intelligent Assistant for Speech Recognition using Natural Language
Processing

Area of Project Work:

Natural Language Processing
Problem Statement:
Traditional computer systems require manual interaction through keyboards, mouse, or touch-
based interfaces, design a lightweight, customizable voice assistant is required for academic
and practical use.

Art ( Patent Search) :

Patent Title of Patent Existing Solutions
Application No. (Abstract of Patent)
KEYWORD Topics ofpotential interest to a user, useful for
14/447,487
DETERMINATIONS purposes such as targeted advertising and product
FROM VOICE DATA recommendations, can be extracted from voice
content produced by a user. A computing device
can capture voice content, such as when a user
speaks into or near the device, One or more
sniffer algorithms or processes can attenpt to
ideti 的 iggerwords in the voice content, which
can indicate a level of interest of the user. For
each identified potential trigger word, the device
can capture adjacent audio that can be analyzed,
onthe device or remotely, to attempt to determine
one or more keywords associated with that
trigger word. The identified keywords can be
stored and/or transmitted to an appropriate
location acsessible to entities such as advertisers
or content providers who can use the keywords to
attempt to select or customize content that is
likely relevant to the user.

18/073,555 SYSTEM AND A method for converting speech in one of a

METHOD FOR plurality of input languages into text using
SPEECH machine transliteration and trans-fer learning is
RECOGNITION disclosed. The method includes a training stage.
USING MACHINE The training stage includes receiving a training
TRANSLITERATIO set of a plurality of audio files and an input text
N AND TRANSFER corresponding to the audio input in any input
LEARNING language using the speech recognition engine;
transliterating the training set to trans-form the
input text into transliterated text that includes
characters of a base language and training
acoustic model with the plurality of audio files
and corresponding translit-erated text using
transfer learning. The method further includes an
inference stage. The inference stage includes
performing decoding on output of the trained
acoustic model to generate text includes
characters of the base language at inference and
transliterating the generated text to output text
includes characters in input language using
 reverse translit-eration.

11/733,108 SPEECH A speech-enabled internet based computing system

RECOGNITION includes a configurable speech recognition engine
SYSTEM FOR which allows Sup
CLIENT DEVICES port for client devices having differing computing
HAVING OFFERING capabilities. Natural language operations can also be
COMPUTING supported as
CAPABILITIES desired.

5,32,867 MULTI-LANGUAGE Input speech of an arbitrary speaker is automatically

SPEECH transcribed into one of many pre-determined spoken
RECOGNITION languages by determining, with the frequency
SYSTEM discrimination and frequency response of human
hearing the radiated spectrum of a speech input
signal, identifying continuously from that spectrum
the phones in the signal. aggregating those phones
into phonemes, and translating the phonemic string
into the pre-determined spoken language for output.

09/049,771 SPEECH TO TEXT A speech-to-text conversion system is

CONVERSION provided which com-prises at least one user
terminal for recording speech, at least one
automatic speech recognition processor to
generate text from a recorded speech file, and
communication means operative to return a
corresponding text file to a user, in which
said at least one user terminal is remote from
said at least one automatic speech recognition
processor, and a server is provided remote
from said at least one user terminal to control
the transfer of recorded speech files to a
selected automatic speech recognition
processor.

Literature Review:

Title of Paper Publication Details Literature Identified for Project

AI-Powered Intelligent Zhang, Z. (2025). AI-
• The paper reviews the historical
Speech Processing: Powered Intelligent Speech
evolution of speech recognition, synthesis,
Evolution, Applications Processing: Evolution,
and processing technologies, highlighting
and Future Directions Applications and Future
the shift from statistical models to deep
Directions. International
learning-based models, which have
Journal of Advanced
significantly improved accuracy and
Computer Science &
naturalness in speech outputs.
Applications, 16(2).
• It discusses three key learning
paradigms in speech recognition:
supervised, self-supervised, and semi-
supervised learning, emphasizing their roles
in enhancing generalization and reducing
reliance on labeled datasets.
• The literature indicates a wide range
of applications for AI-driven speech
processing, including smart homes,
healthcare, and finance, showcasing the
transformative impact of AI technology on
various industries.
• Current challenges in the field are
addressed, such as performance issues in
noisy environments, data scarcity for low-
resource languages, and concerns regarding
data privacy and algorithmic bias.
• The paper anticipates future
advancements in speech processing
technology, focusing on personalized
services, real-time processing, and
multilingual support, while acknowledging
the ongoing technical and ethical challenges.
Using AI-Powered Dennis, N. K. (2024). Using
Speech Recognition AI-Powered Speech • The literature review highlights the
Technology to Improve Recognition Technology to effectiveness of AI-powered speech
English Pronunciation Improve English recognition technology in enhancing
and Speaking Skills Pronunciation and Speaking pronunciation and speaking skills among
Skills. IAFOR Journal of EFL learners, with several studies reporting
Education, 12(2), 107-126. improvements in spoken English abilities.
• It emphasizes the need for further
exploration of AI's potential drawbacks,
such as reduced human engagement and
accessibility issues, indicating a gap in
current research.
• The review calls for detailed pedagogical
strategies to effectively integrate AI
technology into language learning,
suggesting that existing methodologies
require critical examination.
• It notes the infancy of research in this
domain, pointing to a vast potential for
future studies to bridge knowledge gaps and
improve application in educational settings.
• The review also discusses the importance of
understanding learner variables, such as
motivation and prior language proficiency,
which can influence the efficacy of AI
interventions.

Voice Assistant Rawat, S., Gupta, P., & • The literature review examines emotion
Using Automated Kumar, P. (2014, recognition in conversational AI, focusing
Speech Recognition November). Digital life on methods for analyzing emotional tones in
assistant using automated virtual assistants and the effectiveness of
speech recognition. In 2014 various machine learning techniques.
Innovative Applications of • It highlights the importance of feature
Computational Intelligence extraction, contextual understanding, and
on Power, Energy and semantic integration in natural language
Controls with their impact processing for identifying emotions from
on Humanity text.
(CIPECH) (pp. 43-47). • The review discusses the impact of
IEEE. emotion recognition on customer service and
mental health, emphasizing the need for
advancements in accuracy and adaptability
across languages and contexts.
• It covers vocal feature analysis, including
pitch and tone, and addresses challenges
such as variability in emotional expression
and cultural factors.
• The review suggests future research
directions to enhance the robustness and
accuracy of emotion recognition systems in
applications like social robotics and
healthcare.
Voice Assitant Using Sahu, A., Jha, A., • The literature review discusses the
Artificial Intelligence Bhargava, R., Priya, P., functionality of virtual voice assistants,
emphasizing their ability to access and
& Kumari, R. (2022, manage user data through voice commands.
March). Voice assistant • It highlights the integration of various
using artificial APIs, such as gTTs and Gmail, to enhance
intelligence. In the capabilities of these assistants, allowing
Proceedings of the for tasks like sending emails without typing.
international • The review notes the evolution of virtual
assistants from software to more interactive
conference on systems that can perform a range of tasks,
innovative computing including controlling devices and providing
& communication information.
(ICICC). • It mentions the importance of user trust
and interaction quality in the adoption of AI-
based technologies, suggesting that these
factors significantly influence user intentions
to utilize virtual assistants.
• The review also points out the potential for
future developments in virtual assistant
technology, indicating a trend towards more
personalized and accessible user
experiences.

Voice Assistants and Terzopoulos, G., & • The literature review highlights the limited
Artificial Intelligence Satratzemi, M. (2019, research on voice assistants and smart
in Education September). Voice speakers, particularly since their rise in
assistants and artificial popularity in 2017.
intelligence in education. In • It identifies key research questions
Proceedings of the 9th regarding user interactions, security
Balkan Conference on concerns, and beneficial use cases in
Informatics (pp. 1-6). education for these technologies.
• A consensus emerges around significant
privacy and security concerns among users,
with suggestions for features like incognito
modes and visual indicators for data
collection.
• The review discusses the contrasting
perceptions of smart speaker users and non-
users, particularly regarding trust and utility.
• It emphasizes the need for further
exploration of the use of voice assistants in
special education and the implications for
users with disabilities.
Current Limitations :

1. Dependence on microphone quality and background noise

2. Internet dependency for advanced AI responses

3. Limited understanding of highly complex natural language commands

Proposed Solution:

The proposed system is a Python-based intelligent voice assistant that accepts voice
commands using speech recognition, processes the commands using AI and logical
decision modules, and provides responses using text-to-speech synthesis. The system
can perform system automation tasks, answer general queries, and interact naturally
with the user through voice.

Aim and Objectives :

[Link] implement speech-to-text conversion for voice input

2. To process commands using AI and NLP techniques

3. To provide voice-based responses using text-to-speech

4. To perform basic system and web automation tasks

5. To demonstrate practical application of AI in human-computer interaction

Feasibility Study:

Technical Feasibility

• Python programming language

• SpeechRecognition and pyttsx3 libraries

• AI language model APIs

• Standard laptop with microphone

Economic Feasibility

• Open-source tools used

• No additional hardware cost

• Estimated budget: Minimal / Nil

I. Expected Outcomes of the Project

• A working prototype of a voice-based virtual assistant

• Improved understanding of AI, NLP, and speech technologies

• Hands-free interaction with system and applications

II. Innovation Potential

The project provides a customizable, lightweight academic prototype of a voice
assistant that demonstrates integration of modern AI concepts such as speech
recognition and language models, with scope for future expansion into smart systems
and IoT.

III. Task Involved

 1. Requirement analysis

2. System design

3. Speech recognition implementation

4. AI response integration

5. Text-to-speech output

6. Testing and validation

IV. Expertise Required

1. Inhouse Expertise:
• Python programming
• Basic AI and NLP knowledge

2. External Expertise : Not Required

V. Facilities Required
1. Inhouse Facilities:

2. External Facilities:
Time Plan:
Task JULY AUG SEP OCT NOV DEC JAN FEB MAR APR
2025 2025 2025 2025 2025 2025 2026 2026 2026 2026

Conceptual
Design
Detailed
design
Design
Design
Modifications
Final Design

Procurement
(If any)
Prototyping
Develop

Modifications

Testing and
Validation
Final
Modifications
Delivery
IPR / patent
draft
Thesis and
Poster

Name and Signature of Project Guide Signature of HOD