32 C
Mumbai
Sunday, June 16, 2024

Minus Zero: This Bengaluru startup built AI tech to help self-driving cars move around better

Autonomous mobility startup Minus Zero in India is proposing a nature-inspired approach to artificial intelligence to address issues such as traffic laws, safety, and complex traffic management. The company focuses on representation, prediction, and adaptation to train AI to become more human-like…

Autonomous mobility is touted as the next big thing in mobility, especially for the shared one, But autonomous mobility has endless problems.

Even in the West with traffic laws, safety, dynamic road navigation, anticipatory human behavior prediction, complex traffic management, and other issues have slowed autonomous vehicle adoption and development.

In India, where traffic laws are considered’suggestions’, Minus Zero, an autonomous mobility startup based in Bengaluru, proposes a nature-inspired approach to artificial intelligence to address these issues.

Minus Zero believes it can expand autonomous vehicles by making AI more humanlike.

“For AI to work for humans, it must think like a human,” says Minus Zero CEO and Co-founder Gagandeep Reehal.

The easiest way to understand nature-focused AI is to imagine how a Bengaluru resident would navigate traffic from all sides. AI hates unpredictability, especially vehicle chaos at all angles. People manage to escape the mess daily.

Reehal says Minus Zero trains its AI models using this cognitive approach of being able to make a decision based on multiple variables [i.e., vehicles moving in different directions].

Reehal suggests focusing on representation, prediction, and adaption to train AI to become human.

His nature-inspired AI hypothesis is that the human brain spends more time correlating data points than storing it.

If you slip on a wet surface and hurt yourself, your brain will remember that water can cause slipping and potential harm rather than the date, time, location, liquid type, or clothing.

Artificial intelligence naturally stores data. By training it to look for correlations, it can learn new behaviors, such as observing other drivers or assessing its own actions.

Reehal’s ‘representation’ and ‘prediction’ thesis involves identifying correlations and applying them in similar situations.

Next is adaptation, which he illustrates with a hatchback driver who switched to an SUV. The brain doesn’t treat these two actions as unrelated or new because driving crossfades. It learns to adapt to new information, such as the SUV’s wider width or softer clutch pedal.

This could mean drawing from past AI experiences and acknowledging their impact on future applications.

In simulation testing using AI based on representation, prediction, and adaptation, our autonomous vehicles outperformed existing AI models and navigated real-world scenarios almost as well as humans. AI made safer decisions where humans would have made risky ones “Reehal.

More AVs will be able to signal each other and share traffic conditions, road closures, and driving lessons, he says.

Minus Zero’s AVs use this nature-inspired AI for autonomous mobility and safety.

Technology and software company Minus Zero works with manufacturers to build AV fleets. A computer inside the vehicle runs AI-powered software, and dashcams and sensors help the vehicle navigate.

Reehal says a simple mobile phone camera would suffice without the camera requirements of some Western countries.

Reehal did not estimate the average price of a two- or four-wheeler AV. He said the startup focused on shared mobility and commercial fleet operations, where redundancies and inefficiencies cost the country $20 billion a year.

The startup secured $1.7 million in funding from Chiratae Ventures, IIT Mandi Catalyst, VM Financial Services, and Jaineo Avishk Singhi Capital Finance. Another funding round is underway.

It plans to launch in India but will manufacture AVs wherever demand is.

Minus Zero demonstrated its zPod, an all-terrain, all-weather vehicle that can reach Level 5 autonomy—the highest level of autonomy without human intervention—last month.

Minus Zero faces competition from global brands like Waymo, Cruise, Aurora, and Tesla.

Conclusion

Autonomous mobility is gaining momentum, but challenges such as traffic laws, safety, and complex traffic management have slowed its adoption. In India, startup Minus Zero proposes a nature-inspired approach to artificial intelligence to address these issues. The company focuses on representation, prediction, and adaptation to train AI to become more human-like. The human brain spends more time correlating data points than storing it, and AI naturally stores data. By training it to look for correlations, it can learn new behaviors, such as observing other drivers or assessing its own actions. Minus Zero’s autonomous vehicles outperform existing models in simulation testing, making safer decisions where humans would have made risky ones. The company works with manufacturers to build autonomous vehicle fleets, with a computer inside the vehicle running AI-powered software and dashcams and sensors helping the vehicle navigate. The startup secured $1.7 million in funding and plans to launch in India but will manufacture AVs wherever demand is.

Newsdesk

Related Articles

Latest Articles