Bengaluru start-up launches AI autopilot system for Indian roads
What's the story
Bengaluru-based autonomous mobility start-up Minus Zero has launched a unique autopilot system designed specifically for India's complex traffic conditions. The vision-centric solution leverages foundational AI models to navigate on unstructured roads with little or no lane markings. It can also deal with unpredictable obstacles like animals, pushcarts, and two-wheelers.
Innovative approach
Minus Zero adopts self-supervised learning for autopilot system
The company has also adopted a self-supervised learning strategy, moving away from traditional rule-based algorithms. This unique approach enables the autopilot system to train itself with tons of raw data without needing human-labeled inputs. Plus, Minus Zero's autopilot system also does away with expensive LiDAR sensors and HD maps. It uses a minimal hardware vision-first stack, which is more economical and easier to implement.
Testing phase
Minus Zero's autopilot system tested on Bengaluru roads
The autopilot system has been tested on challenging routes in Bengaluru, marking a major milestone as the first application of foundational models to Indian roads for autonomous driving. The platform is designed to adapt to new obstacles and unseen environments, making it suitable for rapid scaling across different geographies and vehicle types.
Collaboration
Minus Zero collaborates with OEMs for market launch
Currently, Minus Zero is working with a number of original equipment manufacturers (OEMs) to commercialize the autopilot system within two years. This puts the company at the forefront of India's growing autonomous driving space. By building globally competitive deep-tech solutions from the ground up, Minus Zero hopes to transform autonomous mobility in India.