Making Our Roads Safer: Research Engineer Isht Dwivedi Is Leading Honda’s Vehicle Technology To Enhance Road Safety

Isht Dwivedi
Isht Dwivedi

An estimated 85% of US workers commute to work using road transport. According to the World Health Organization, road traffic injuries are the leading cause of death in children and young adults aged 5-29 years. In the US alone, the National Highway Traffic Safety Administration (NHTSA) estimates that motor vehicle traffic crashes caused almost 43,000 deaths in 2021, which is not only a 10.5% increase from 2020 but also the largest annual percentage increase in the Fatality Analysis Reporting System's history.

This ongoing loss of lives due to road accidents demands urgent action. Honda has taken an initiative to bring about meaningful change by setting a mission to reduce road fatalities by 50% before 2030, with the ultimate goal of completely eliminating them by 2050. The goal is ambitious, but, as Bill Bradley once said: "Ambition is the path to success. Persistence is the vehicle you arrive in".

Isht Dwivedi - ambitious and persistent

From a very young age, Isht Dwivedi was fascinated by the possibility of teaching computers to learn to perceive the world around us the way humans do. He pursued engineering, earning his Bachelor of Technology from India's premier research and engineering institute, IIT Guwahati. It was his first opportunity to formally study subjects such as computer vision, machine learning, image processing, and pattern recognition. For one project, he developed a wirelessly controlled snake robot and, for his undergrad thesis, he simultaneously worked with computer vision on a writer identification tool - an important application in the forensics field. With a major in electrical and electronics engineering and a minor in computer science and engineering, he was laser-focused on developing his career as quickly as possible.

To augment his knowledge and expertise, Dwivedi pursued a master's degree in computer science at Columbia University. This opportunity provided him with a chance to work with renowned leaders in the field of computer vision, including Professor Shih-Fu Chang, and learn some of the most advanced technology available in the US. Following his graduation in 2018, Dwivedi began working at the Honda Research Institute USA, where he focuses on developing cutting-edge computer vision and smart mobility technologies to minimize road accidents.

Dwivedi has worked on some groundbreaking projects, primarily related to developing computer vision algorithms to improve road safety and road scene understanding using video captured from cameras mounted on cars. He developed a novel Bird's Eye View (BEV) semantic segmentation algorithm, which creates a map of the car's surroundings. This novel algorithm also needed a new type of dataset to be created. He explains the process: "For creating [the required] dataset, I had to create a large number of 3D maps using only videos collected from the car, which was difficult. Typically, people use additional sensors such as LiDAR if they want to build maps, but in this project, we were not using this sensor which made the task more challenging. These 3D maps, obtained solely from videos, were then annotated with semantic road information indicating lanes, intersections, construction zones, crosswalks, etc. This dataset was [then] used to create our new real-time algorithm to predict BEV segmentation". Dwivedi has also published video classification algorithms at reputed conferences such as the International Conference on Robotics and Automation (ICRA). These algorithms analyze information about the road scene. This technology can provide insight into road surface conditions, the type of road, intersection details, and other relevant information about the scene, such as construction zones, road merges, or highway exits. Dwivedi's work has not only been published but has also received four patents.

Honda is creating a safer future

The team at Honda is currently focused on incrementally enhancing road safety through new developments in Advanced Driver Assistance Systems (ADAS), while simultaneously working to develop a fully independent ADS (Autonomous Driving System) commonly known as a "self-driving" car. The idea is that, over time, their Driving Assistance Systems will become more intelligent and eventually lead to the evolution of fully autonomous vehicles.

Dwivedi's work has led to some novel patented inventions. He is a key member of the Honda team working on algorithms to understand and localize risky elements on the road. He has developed algorithms to create assistive captions that a driver will hear inside the car, providing information about upcoming situations. For eg., it can also be used to alert the driver if a risky scenario such as an imminent collision is predicted. Dwivedi is also developing a road condition monitoring system that automatically identifies locations with poor road infrastructure. This road condition monitoring system is installed in a car and is connected to the internet. It logs infrastructure quality information in different areas automatically and uploads it to a central server. Without the availability of such a system, road operators now have to physically visit different areas and log areas with poor infrastructure manually. This process is slow and expensive. In the future, Dwivedi's road condition monitoring system will be installed in production cars which will allow road operators to easily identify areas with degraded infrastructure quality and perform maintenance more efficiently.

Despite making impressive strides, Honda realizes the urgency in achieving their ambitious targets and is therefore taking prompt action. Currently, Honda has multiple exciting projects in progress to ensure timely progress toward their goals. In addition to his other projects, Dwivedi has also researched approaches to do future trajectory prediction and path planning inside intersections for autonomous driving (AD) systems. There are other companies developing similar AD systems, but they are based on extremely expensive High Definition (HD) maps which are annotated tediously by humans, making Honda's approach much more economical.

Although these projects are currently operating separately, the overall development of Honda's ADAS and AD systems inspires great hope for a much safer future on our roads. The grim reality is that road accidents not only cause significant losses to individuals and families but also have a severe economic impact on the entire nation. According to the WHO, these accidents can cost countries up to 3% of their GDP. However, the most crucial aspect of this progress is that it has the potential to save countless lives. Isht Dwivedi is an essential player in turning Honda's bold dreams into a reality.

ⓒ 2024 TECHTIMES.com All rights reserved. Do not reproduce without permission.
Join the Discussion
Real Time Analytics