Embelias
Shubham Shukla
Dr. Shubham Shukla is an Associate Professor in KIET Group of Institutions and Assistant Research and Development Engineer, with an area of research in the field of Multi Robot Path Planning with Nature Inspired Algorithm.
He acted as the project guide in the development of Anushka Robot.
Atul Kumar
Mr. Atul Kumar is an expert in Power Electronics and Circuit Designing.
He is the hardware and electrical systems incharge of development team of Anushka Robot.
Piyush Khanna
Mr. Piyush Khanna is an experienced programmer with stack ranging from Software System Designing and Web Development to Machine Learning and IoT.
He is the incharge of software and electronic systems of development of Anushka Robot.
Govind Panwar
Mr. Govind Panwar is a software and DBA intern with Team Embelias.
He is responsible for the database retrieval and dataset collection for training custom data for the robot.
Harsh Jain
Mr. Harsh Jain is a hardware and Embedded Systems intern with Team Embelias.
He is responsible for R&D on circuitry and managing the supply of electronic components.
And together,
we
It's all possible
Face Recognition
Anushka excels in face recognition, employing cutting-edge computer vision to memorize faces in real-time with remarkable precision. This capability ensures accurate identification across diverse expressions and features, enabling personalized interactions in dynamic environments.
The implementation incorporates popular deep learning frameworks like TensorFlow and facial landmark detection algorithms to identify key points on a face, enhancing the efficiency and accuracy of face recognition.
Speech Recognition
With a support of over 40 Indian languages and major foreign languages including English, Spanish, Japanese etc, Anushka boasts a huge Natural Language Processing unit.
The NLP module in Anushka employs cutting-edge techniques, including sentiment analysis and contextual understanding, allowing it to grasp the emotional undertones of user queries. This not only enhances the quality of responses but also enables Anushka to tailor its communication style based on the user's emotional state.
Bot Operations Scheduling System
Bot Operations Scheduling System, shortened to BOSS, is the heart of Anushka robot's development. It is a dedicated Pseduo-OS specially designed to merge Linux's fast instruction control and Windows' huge support for python libraries. By acting as a hypervisor, it makes the robot extremely fast and responsive to all its sensing units.
It also acts as a messenger queue between all the modules of the robot and can actively command a module in case one of the vital modules fails.
Object identification and Tracking
Anushka has a dedicatedly developed computer vision program which enables it to approximate distance without a depth-sensing camera.
Whether it's following a person as they move around a room or tracking a specific object, Anushka's Object Tracking enhances its engagement capabilities. This feature is particularly valuable in environments where the robot needs to navigate and interact with its surroundings actively.
Cultural Integration with Vedic Principles
The entire system design and brain's response system of Anushka has been built after a thorough research on the Vedic views of human mind's architecture.
This integration aims to align the robot's decision-making processes with the holistic understanding of the mind presented in Vedic philosophy. Components such as 'Chitta', 'Manas', 'Ahankaram' and 'Buddhi' have been incorporated into the architecture of robot's Intelligence.
IoT and Automation
Anushka can integrate itself seamlessly into smart home, autonomously managing appliances like fans and bulbs based on predefined conditions.
Anushka also has the ability to completely automate the task of application writing and printing by performing print jobs from remote printers.
She can also make phone calls as well as send and recieve Whatsapp messages autonomously.
But, How?
Team
Our beloved team worked tirelessly for 1.5 years to bring Anushka to life.
RoboCup
Anushka, The Humanoid was made to be presented at the RoboCup @Home, '24 Eindhoven.
'24 Robocup @ Home
About RoboCup '24
RoboCup is an annual international robotics competition founded in 1996 by a group of university professors (including Hiroaki Kitano, Manuela M. Veloso, and Minoru Asada). The aim of the competition is to promote robotics and AI research by offering a publicly appealing but formidable challenge.
|
|
Our Novel Approaches
Single Camera Depth Approximation
The robot uses a novel approach for distance approximation using computer vision. In this approach, the robot is able to efficiently approximate the distance from an object or person by analysing its image and comparing it to the size of the respective standard object or body part.
This is helpful in case where there is no other way of depth approximation by ultrasonic waves or Infrared waves sensor system. This technique also enables the robot to tag the person to be followed as well as maintain a formal distance from him/her.
Speech Question Cache-ing
A novel approach for âquestion cache-ingâ allows the robot to store the questions and their answers locally based on their frequency of being asked in normal conversations.
Thus, in case of repeating questions, the robot doesn't need to necessarily scrap answers from web but fetch it from local question cache. By utilizing NLP, Anushka can interpret and extract meaning from spoken or written text, enabling it to understand user commands, generate appropriate responses, and engage in natural language conversations.
Facial Feature Forgetfulness
Novel approach for identifying and prioritizing humans in real-time has been applied to the face recognition system to mimic human behavior. The face recognition system uses deep learning Convolutional Neural Network for first creating face embeddings and storing them in local database.
This database is routinely checked for repeating faces and their mapped names. Priority is given to people based on their number of times meeting the robot and time since last meeting. On a routine-base dumping, faces are âforgottenâ if their priority is less, that is, the robot has met the person only once/twice and long time ago.
Vedic Weave
The entire system design and brain's response system of Anushka has been built after a thorough research on the Vedic views of human mind's architecture.
Novel prioritization techniques have been employed in the robot, segregating each data unit as a portion to be remembered or discarded according to the robot's discretion. This quadriplex architecture comes from the vedic texts, and named 'Chitta', 'Manas', 'Ahankaram', 'Buddhi'.