Worried about the country that accounts for over 11% of global road accident deaths, equating to nearly five lakh road accidents claiming 1.5 lakh lives annually, despite having only 1% of total vehicles in the world, a team of students of Chandigarh University began sketching their dreams to find a solution.
Discussions and deliberations led to the decision to take an attempt that can help curb mishaps on the road and reduce deaths caused by accidents. They zeroed in on developing an autonomous car powered by Artificial Intelligence.
Birth Of NOMAD
Speaking to Mobility Outlook, Dr Harjot Singh Gill, Associate Director, Engineering, Chandigarh University and Mentor of NOMAD Project, said, in 2018, the team came up with an idea for a prototype. However, he asked why couldn’t they work on an actual project. Therefore, they took the shell of Tata Nano, as it was economical, removed the petrol powertrain and related aggregates and powered with a 1.5kW motor at the rear axle, powered by four Nickel-Cadmium batteries, which is used in the e-rickshaws.
Meticulously putting their heads together, the students of the Mechatronics Engineering Department have created an autonomous car -NOMAD modified from Tata Nano. Inspired by the free bird, the vehicle was initially conceived by the students of 2020 Batch, including the team leader Satyam Sharma, KVS Mohan Vamsi, Navjot Singh and Sarabsukh Surya. It combines a variety of sensors to perceive their surroundings, such as Radar, Lidar, Sonar, GPS, Odometry and Inertial Measurement units. In addition, its advanced control systems interpret sensory information to identify appropriate navigation paths and obstacles and relevant signage.
The autonomous car can sense its environment and move with little or no human input, thus curbing the possibilities of road accidents caused majorly by human error. In addition, it provides better accuracy even in foggy weather, enabling a low cost of operation.
ML, AI Algorithms
The NOMAD is equipped with sensors and cameras to diagnose traffic conditions and act accordingly based on a proper road-legal base structure. As the vehicle runs on the road, it collects the data, enabling Machine Learning and Artificial Intelligence algorithms to train the system for various road traffic conditions. In addition, the vehicle is integrated with a 15-inch touch screen for automatic transmission and provides on-screen, touch-enabled buttons for headlights, windows, and indicators.
Reducing Accidents
The car aims to reduce accidents through Pedestrian and Obstacle Detection and Avoidance technology, including Automatic Emergency Braking. It has the capability to distinguish a human/moving body from an object. The car would apply brakes and stop if it comes across an object, while if it comes across a moving body, it will resist the contact by first slowing down and taking a turn. However, if there is insufficient space to make the turn, the car would apply emergency brakes.
Satyam Sharma says the NOMAD uses Deep neural networks to detect objects and the road in each individual frame captured by the camera sensor. At the same time, it drives and assesses the situations to make the driving decisions like acceleration, braking, left or right turn etc. The vehicle uses the data from GPS to know the path the car has to follow, and then neural networks are used to make real-time decisions.
'The recurrent neural network will work as the heart of the system as it has the ability to process sequential data and back-propagation through time. This ability of back-propagation through time allows to make a decision, basis the current as well as the previous position (captured frame) of the car allowing it to detect and differentiate between slow moving and fast-moving objects and their direction of travel to avoid collision efficiently,' Sharma says. More than 200 programming codes in Python have been developed through machine learning.
The students developed Phyton programming independently, with assistance from experts and the industry. It took about a year and a half to finalise the codes and run the car on the university campus. “We have a pre-defined path in the university, and the car is able to move in the autonomous mode,” Gill said.
Unenamoured But Efficient Eyes
Two main cameras provide the primary feedback to the image processing system, supplemented by 360° sensors all around the car. The vehicle has a 2-speed transmission, and is capable of achieving a maximum speed of 25 kmph, and can be charged in 4-5 hours. The car has a solar panel to augment battery charging, and it takes about ten hours to get the battery fully charged. Besides, the batteries can also be charged through inverters in about two hours. So far, INR 3.80 lakh has been spent on this project, supported by the varsity.
NOMAD has now been equipped with many different sensors to accomplish the task of self-parking. The front and back cameras detect a suitable parking space using basic image processing. Once the parking space has been selected, the vehicle uses its rearview camera and short-range ultrasonic sensors to detect the objects surrounding it and slowly start parking it with respect to the dynamic environmental changes. The vehicle calculates every movement of the tyres made by the car's motors. It uses the rearview camera and ultrasonic sensors as feedback mechanisms to constantly correct the car's position and slowly park it.
The car was tested every day for about two months running about 20 kms a day, cumulatively clocking over 1,200 kms without any glitch.
Next Step
Asked about the next step for the team, Gill said, at the moment, the car is based on an AI algorithm, and it will learn as much as it runs on the road. It will run considering the road condition, pedestrian movement and others. The next part is that the students from the next batch are taking it in a different shell and will be working on independent motors as well, with improved speed, stability and control.
Sanjeet Singh, Dean, Research, Chandigarh University, said the varsity has a Corporate Advisory Board, and the autonomous car was demonstrated to about 400 CEOs of different companies, including a top official from Tata Motors.
The innovation-driven university has a technology business incubator funded by the Department of Science and Technology and the Ministry of MSME, Government of India. Of the INR 20 crore funding, the varsity will plough about a crore of rupees into the autonomous vehicle project. The first project will not be commercialised; instead, it will be given to the rural areas. Clarifying the objective, Singh said the varsity has already developed a couple of projects for rural areas – an automated sanitiser and mask manufacturing machine, a small lab bench for schools in rural areas that can enable students to carry out over 200 experiments, including those relating to AI. These were given to schools in the border areas. Similarly, the varsity is planning to make three or four cars and deploy them in different applications in the rural area to get feedback and think about commercialisation, he added.
Sharpening Focus On Research
S Satnam Singh Sandhu, Chancellor of Chandigarh University, says that ever since the inception of the varsity, it has laid focused and pointed emphasis on inculcating the spirit of research and innovation amongst its faculty and students alike. The top performance of the varsity in the field of research and innovation is an outcome of that.
'Till date, the students and faculty of Chandigarh University have filed 1,273 patents. It fills me with pride to note that as many as 683 patents have been published, while 128 patents are in the first examination stage. Apart from this, 120 patents of the university have been approved. For the last three years, Chandigarh University holds the number one position among the private universities of the countries to file the maximum number of patents, while it is ranked third in the number of patents published and granted,' he adds.
The University has 20 laboratories under industry collaboration, 30 state-of-art research centres, and more than 60 research groups, where over 800 research scholars work on innovative research and innovation in various fields.