Bengaluru startup Minus Zero has launched a completely integrated autopilot solution designed specifically to address the intricacies of Indian traffic. According to the company, its system has been successfully tested on Bengaluru’s crowded roads, successfully dealing with situations that include the lack of lane marking and the presence of unknown factors such as animals, handcarts, and two-wheelers. While other systems rely on high-definition maps or LiDAR technology, this solution employs a vision-based setup.
The firm’s technology is founded on core AI models trained in a self-supervised learning manner from large unlabelled raw data. This differs from the traditional rule-based system that needs human input and the mapping of the environment in detail. Through experience in the real world, the AI adapts to new roads and unexpected roadblocks. At present, the autopilot system operates under a ‘Hands OFF & Eyes ON’ category that needs the presence of a safety driver during use.
Addressing Emerging Markets with Scalable Technology
Minus Zero perceives huge prospects for its technology in emerging economies such as India, where road infrastructure is predominantly unstructured and the infrastructure differs greatly. While autonomous technologies like Tesla’s Full Self-Driving (FSD) and Mercedes’ Drive Pilot operate in heavily regulated environments of developed nations, India offers special challenges. The firm indicates that while the majority of ADAS are deployed in developed countries, a significant proportion of global road crashes happen in developing markets—emphasizing the necessity for solutions in regional context. Minus Zero is already collaborating with leading Indian carmakers and in talks with others, with a two-year target for production rollout in partnership with OEMs.
Key Technological Features
Methodology shift: The autopilot relies on foundation AI rather than classic supervised learning or rule-based methodologies.
Efficient sensors: The system prioritizes a vision-centric approach and does not rely on expensive sensor suites.
Map-free driving: It is free from HD maps, which are not yet ubiquitous in India.
Scalable and adaptive: The AI models themselves are adapted to update constantly, making the technology adaptable to different terrains and vehicle types.
End-to-end development: Minus Zero developed its own onboard deployment software, simulation platforms, and generative AI tools.
“Following LLMs and AI agents in the virtual space, next-generation base models in the physical space such as autonomous driving, robotics, etc. are picking up steam worldwide,” the company said in a statement. “With India being very strong in establishing a place in the AI race, this is one space where we can develop in India for the world.”