Wipro and IISc working together to develop driverless cars

Wipro and IISc first announced their strategic partnership earlier this year; under the purview of this tie-up, the two will work on autonomous technology, vehicle-to-everything communication, and more

Autonomous car, Driverless car, Indian Institute of Science, Wipro, Automation, Robotics, Machine learning, Artificial Intelligence, Neuro morphic systems, Dr Ramachandra Budihal, Bengaluru, Wipro IISc Research Innovation Network

Building an autonomous or driverless car for Indian roads may seem like an impossibility — given the volume of traffic in cities, the unpredictability of road conditions, the behaviour of people, accident rates and even the presence of stray cattle.

However, scientists and researchers from the Indian Institute of Science (IISc) and IT major Wipro have joined hands to create a driverless car for Indian conditions with a targeted rollout date of March 2020.

The collaboration is also an effort to take IISc and Wipro to the frontiers of new technology — automation, robotics, machine learning, artificial intelligence and neuro morphic systems. The nearly one-year collaboration project to build an autonomous car for Indian conditions is an extension of a nearly three-year effort at Wipro to build an autonomous car using data for roads in Europe and the US.

The efforts at Wipro have resulted in “a grounds-up autonomous vehicle” but the real challenge now is to build a car by factoring in the unpredictability of Indian roads, say the IISc scientists and Wipro researchers engaged in developing the car at IISc.

“We are working on some cutting edge algorithms for Indian conditions. Earlier whatever we worked on was user data from Europe and America but we want to do something for India. We cannot use the systems that we built for the other data sets because the conditions here are unique,’’ said Dr Ramachandra Budihal, Wipro’s head for AI, autonomous systems and robotics, who is based on the IISc campus with a team of 18 engineers to build the car.

“It is a greenfield design — everything has been designed from scratch. We are designing our own electric vehicle and then putting all the systems inside. It is a concept autonomous vehicle,’’ Budihal said.

A dozen professors drawn from six IISc departments — aerospace engineering, electrical engineering, electrical communication engineering, electronic systems engineering, electronic design and technology, computer science and automation — are involved in the project, which has nearly 200 participants from IISc and Wipro Ltd.

Over the past six months, investigators have been using a car equipped with multiple sensors to collect raw data on aspects of navigating the roads in Bengaluru and have culled out around 100 hours of data, out of nearly 1000 hours collected, to build algorithms and systems for the driverless car.

The plan is to have about 28 sensors on the autonomous car, say the scientists.

One of the key challenges in building a driverless car for Indian road conditions is the amount of processing power that will be needed on board the car to process the enormous amount of data coming in from the road environment, say the scientists involved in the project.

“The processing power needed for this data is far in excess of what is seen in more organised environments,’’ said Prof Debasish Ghose of IISc’s aerospace engineering department.

“The autonomous system design has a lot of spin offs. This is a platform and it has many technology components. This is a certain use case — there will be a certain core technology in terms of algorithms, sensing, computing — all these things are being developed and tested so innovation can happen there. Autonomous vehicles are an excuse to innovate,’’ said Prof K V S Hari from the electrical communication engineering department.

The car project is part of a Wipro IISc Research Innovation Network created recently — Wipro will invest Rs 9.6 crore in collaborative research in cutting edge technology areas like “artificial intelligence, machine learning, computer vision, and human machine interaction, related to autonomous systems and robotics”.


Most Popular

Digital Edition

October 2019
From the magazine

Subscribe Now