top of page
11th CavinKare-MMA ChinniKrishnan Innovation Awards 2022
< Back
CKIA-MMA/2021/12/1626087625122
12 Jul 2021
Last Name
Email
Mobile
PAN card image (Front)
Image
Max size: 25 MB
Registration Certificate
Image
Max size: 25 MB
Address proof
Image
Max size: 25 MB
Bank / IT Statement
Select DOC / PDF
Max size: 25 MB
Upload product pix
Image
Size: 100 KB
Innovation Info
Has the innovation been patented?
Yes
No
Applied for a patent?
Yes
No
Patent / Reference ID
URL of the product picture or video
Has this innovation been recognised / awarded by any other organisation?
Yes
No
Organisation that has recognised your innovation
Additional Info
Your innovation in brief
We, at Minus Zero, became the first company in the world to test a Self Driving Vehicle (driverless vehicle) live on unregulated Indian record (with Level 3 autonomy). This was done in record span of 4 months and budget of only $700 by retro-fitting a rented electric three-wheeler rickshaw. We are able to perform all the maneuvers that could be performed by Tesla's Autopilot (Level 3) without spending billions in R&D.
Describe about your innovation
We are in process of filing patent for two sub-technologies powering this innovation. We have different AI & mathematical algorithmic approaches clubbed with special hardware configurations that together helped us achieve this. 1. Our proprietary multi-modal nature-inspired AI which can mimic the cognitive intuition found in human brain, making it less dependent on data, i.e. it is able to gain maximum insights from minimal amount of data. With this in place, our entire autonomy stack would need 83.2% lesser data, requiring just 20,000 miles of driving to be able to achieve Level 5 autonomy compared to 20 million miles of data used by Waymo to achieve Level 4 autonomy. 2. We eliminate the use of expensive depth sensors like LiDARs entirely, decreasing capex on single vehicle relying of camera sensor suite along with cheaper RADARs/Ultrasonics. This was made possible by our proprietary algorithm (patent filing in process) that can extract 3D information like depth, velocity and acceleration of surrounding objects from a 2D video feed of monocular camera only. 3. Our algorithms use 10x lesser GFLOPs per frame reducing latency and enabling faster computation, which make us capable to run our stack on comparatively cheaper processing hardware. 4. Uses cognitive predictive based motion-planning reducing bias in decision making. This was the solution to two major problem faced by Self-Driving Industry-. 1. High dependency on data which causes AI models to fail in unseen scenarios leading to crashes. So companies are spending millions on data acquisition only. But we are able to replicate human intuition artificially. Best analogy would be "You do now show a baby 20,000 images of a dog to make him know there is a dog in front of him". 2. Need of expensive computation and sensory hardware (high-end processors on vehicle, costly LiDAR sensors and supercomputers for model training) is eliminated with our proprietary nature-inspired AI that is less dependent on data.
Approval Status
arrow&v
Submit
Notes
Your content has been submitted
An error occurred. Try again later
Additional File
Select DOC / PDF
Size: 25B
Personal Info
First Name
Business Info
Organisation Name
Date of Incorporation
Organisation Type
Private limited company
Partnership
Proprietary
Other
Firm's PAN Card Number
Street Name
State
City
Pin Code
Upload Info
bottom of page
window.lintrk('track', { conversion_id: 9253641 });
%20(1).jpg)