Programming a Purpose Based Humanoid Robot

Project Title

Programming a Purpose Based Humanoid Robot

Project Area of Specialization Artificial IntelligenceProject Summary

The projected chosen to be done by us is titled “Programming a purpose based Humanoid Robot”. The project targets the demands of the public and brings it to life. The main idea is to prepare a robot that has a human outlook and program it in such a way that it can be used as an office assistant. The programming options aren’t limited and the robot can be programmed to perform multiple and diverse range of tasks according to need. The tasks that are ought to be done by our robot are as follows

1 – Capable enough of having a full on conversation with employees, staff members or clients of the officer. For example, greeting everyone coming to the person’s office and answering questions like when the person inside is free, his meeting timings and more specific details etc.

2 – Telling the officer who is outside to meet him by pronouncing his name and showing the image of the person outside to the officer.

3 – Movement from one side of the officer’s room to the other, carrying files etc.

4 – Virtual assistant tasks like opening browser, Wikipedia, Facebook, sending emails etc.

5 – Can also be used for security purpose (as a security guard). The executive/officer can deploy it in his office when he’s going out and the humanoid robot will send him an email notification when an unidentified face is captured by the camera.

Combined with Artificial Intelligence and Machine Learning, the project involves different fields of science and technology through which the learning opportunities are innumerable. These fields will open up a lot of gates for us as students in the future, enabling us to work at reputed companies only through the things learnt via this project.

In short, the project will turn out to be a revolutionary product that will attract people from all sorts of different fields. The most important factor lies in the simplicity of the project and the robot will be able to do everything whilst looking simple.

Project Objectives

The main objective of this project is to play a part in the emerging field of Artificial Intelligence, explore various aspects linked to it providing efficient solutions. The main aim is to develop a modern day humanoid robot to carry out different human tasks and occupy different roles in the employment sector. Robotics and AI amplify human potentials, increasing productivity and moving from simple reasoning to human like capabilities. With increased capabilities and sophistication of AI systems, this project can be used in diverse ranges of employment sector. The main goal of this project is not to replace humans in their jobs, but to make all processes safer and more efficient. It’s not a confrontation, but rather a beneficial collaboration between automated robotics and humans. Especially in situations where human intervention might be dangerous, a humanoid robot with the same human capabilities and accurate outputs. It could also withstand harsh environments in addition to providing as much work time as required.

As the name of the Project suggests itself initially it looks to focus on four deliverables. Communication, assistance, security and automated movement. These services are often essential in the workplace sector and have huge demand. The Communication is to be implemented via National language processing (NLP). With the help of NLP, the robot will be able to communicate like a normal human which would help it understand the query and provide the required assistance such as scheduling meetings, sending emails etc. It also would offer security via computer vision which enables it to detect known and unknown faces, therefore providing the service of alerting the designated authorities. Movement is a very important aspect that we look to provide in the robot as in an office constant displacement from one place to another is required. The unique thing we look to implement here is its auto movement where avoiding it would navigate and avoid obstacles completely on its own using computer vision. This project’s objective is to provide maximum efficiency and greater productivity under any circumstances. This would prove to be a great solution in a situation like the recent pandemic where manpower had to be reduced due to quarantine thus effecting production and daily life work but a humanoid robot in this situation can prove to beneficial carrying out the required tasks without any threat.

Project Implementation Method

The implementation of the project involves both, hardware and software.

For hardware we require a humanoid robot body which was available to us in the Chinese market but due to the current circumstances involving Covid-19 we decided that this wasn’t a feasible option and settled for another approach which is building a humanoid robot body and integrating necessary items in it which involves:

1-Raspberry Pi: This will be the focal point of the humanoid robot, like the heart or the brain. Every action, answer or command will result due to code which will be running on Raspberry Pi.

2-Raspberry Pi 5-inch LCD Display: A LCD display which will be connected to Raspberry Pi via HDMI cable for completion of certain virtual assistant tasks like opening different social media accounts which will be viewed on the mini screen and also viewing database created using Computer Vision techniques of people who visited the office in the officer’s absence.

3-Mini Speaker along with a Mic: A mini speaker and a mic connected to Raspberry Pi so we can give our input/query to the humanoid robot and hear the humanoid robot’s response using Natural Language Processing Techniques.  

4-External Camera(Optional): If the Raspberry Pi camera isn’t good enough for face detection and isn’t giving a high accuracy, we will opt for an external camera for better resolution.

5-Electronic Devices: Electronic devices such as sensor and motors for movement of humanoid robot.

For software we are using python as the programming language. The reason being that it has standard and updated libraries like NumPy, Pandas, Pybrain, OpenCV, NLTK and SciPy that help in the development of Artificial Intelligence and its components such as NLP and Computer Vision.

Benefits of the Project

Every project to be initiated has a long back story. Everything is brought into perspective when doing a specific project, things like reasons for doing the project which involves all the basis took into account when the first idea of the project was created, but mainly the most fundamental thing which is considered before starting the project is the contribution it will make towards the betterment of the society and the beneficial qualities it possesses. Similar to this, a lot of thought was put into the project we opted for, and to say the least it’s nothing short of a technically amazing project aimed at aiding the advancement of the world.

The project involves a humanoid robot working as an office assistant. Considering the widespread of the COVID-19 pandemic, a robot cannot get infected nor can it infect anyone else. This is a huge benefit for big corporations as well as small offices, a robot working as an office peon will definitely aid in reducing the spread of the virus inside and the outside the office. The benefits that it provides as a project itself range from providing security to entertainment and not to show off; this genuinely looks cool and is a step towards the technical advancement of a country. In any big office or a workplace, there are more than a few top man/executive/heads, so they require personal assistants too. Instead of paying 30-35k to each personal assistant every month, buying this humanoid robot would save the company/office a lot of money in the long run, secondly the security and virtual assistant tasks makes it even more worthy, efficient and useful for the workplace. Apart from this, this project also benefits us as students in the following way

Implementing this project involves research and execution of different fields and subsets of Artificial Intelligence, Natural Language Processing and Computer Vision.

Language processing and Computer Vision are the two major subsets of Artificial Intelligence, using both in the same project and having theoretical and practical knowledge of both fields opens up a lot of doors for the future and has a lot of academic worth in our opinion.

In conclusion, after extensive pondering and research was the decision made to do this project. The benefits that it will provide are quite evident; the project might stir up popularity which may lead to excessive production of the project which in turn will create job and work opportunities. With the proper funding the project will flourish into a one of a kind project.

Technical Details of Final Deliverable

The project basically has four major deliverables which merge to constitute a final deliverable.                    

1-Natural Language Processing Communication: The first deliverable is using NLP techniques to make a closed domain chatbot which can answer queries of the users, have a conversation with employees and set up meetings for the specified officer. For example, if someone asks about meeting timings available for a particular office, the chatbot can answer on the basis of the data it’s been trained on. The NLP techniques used are:

Tokenization: Tokenization is essentially splitting a phrase, sentence, paragraph, or an entire text document into smaller units, such as individual words or terms. Each of these smaller units are called tokens. Tokenization is one of the most common tasks when it comes to working with text data.          

Stemming: Stemming is basically removing the suffix from a word to reduce it to its root word.

Bag of Words: The output of Tokenization and Stemming acts as an input to the technique “bag of words”. For training data, we have to train our deep learning model but we can’t simply put string or sentences into our model so we convert the strings into vectors which contains numbers. For this we use a concept called bag of words.

Neural Network: We train our data through artificial neural network called feed forward neural network. We input the binary coded vector in the neural network, it passes through certain number of hidden layers that we define in our code and in the end the output is predicted through the final layer called Softmax activation function.

2-Face Recognition Using Computer Vision: The second major deliverable is using Computer Vision techniques to detect faces, differentiate between known and unknown faces. If the robot is in security mode send an email to the authorized person upon detecting an unknown face and if the robot is in office mode, then make a database of the people visiting the office. For this purpose, we use a Computer Vision technique called histogram of oriented gradients. The main idea behind Histogram of Oriented Gradients is to extract features into a vector and feed to a classification algorithm that will tell if the face is present in a system or not. In the backend, histogram of each training image is calculated by filtering the image with kernels and it is compared to all the new faces/data coming in. If the histogram matches this means the face matches too. This is done by using OpenCV.

3-Virtual Assistant Tasks Using Selenium: The third deliverable is execution of virtual assistant tasks such as sending emails, opening social media accounts or playing songs etc. For this purpose, we are using a tool/library called Selenium. We will connect with the browser using Selenium and carry out the tasks the user requested through voice command.

4-Movement Using Lidar Sensor: Movement from one side of the office to the other to carry things.

Final Deliverable of the Project HW/SW integrated systemCore Industry ITOther Industries Security Core Technology Artificial Intelligence(AI)Other Technologies RoboticsSustainable Development Goals Industry, Innovation and Infrastructure, Life on LandRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	48500
Raspberry Pi 4 Model B Starter Kit	Equipment	1	11000	11000
Raspberry Pi 8mp module V2 Camera	Equipment	1	4500	4500
Raspberry Pi 5-inch LCD Display	Equipment	1	6500	6500
Raspberry Pi 3 Power Supply and HDMI cable	Equipment	1	1500	1500
Lidar Sensor	Equipment	1	15000	15000
Humanoid Robot Body	Equipment	1	10000	10000