Swaayatt Robots

11/06/2024

An autonomous driving business called Swaayatt Robots, located in Bhopal raised $4 million from angel investors in the United States.

11/06/2024

We will ensure that we scale autonomous driving to new heights going forward, with the recent funding of $4M, which is a part of larger investment from global investors. We will build the scalable L4 autonomous driving tech architecture by the end of this year, fuelled by our L5 autonomous driving R&D, with the support of a great team!!

09/06/2024

Funding Alert 🚨

Swaayatt Robots (स्वायत्त रोबोट्स), a Bhopal-based autonomous driving startup, has raised $4 million from a US-based angel investor.

Sanjeev Sharma, Founder of Swaayatt Robots, took to LinkedIn to announce that the funding is part of a much larger round the startup is raising.

The round was closed at a valuation of $151 million.

He also mentioned that the startup will raise the remainder of the funds at a valuation of about $175 million.

The funding will be used to invent new AI capabilities to solve the problem of autonomous general navigation.

Read more here 👇
https://yourstory.com/2024/06/swaayatt-robots-autonomous-driving-startup-raised-4million-funding

09/06/2024

Autonomous driving startup Swaayatt Robots has bagged $4 Mn in funding from unnamed US-based investors at a valuation of around $151 Mn

03/04/2024

Mr. Anand Mahindra, Chairman of Mahindra Group, praised our cutting-edge research at Swaayatt Robots, towards enabling Level-5 autonomous driving! :)

Jai Sri Ram 🙏 Jai Ma Kali 🙏

07/05/2020

Swaayatt Robots' education venture, DeepEigen, has launched its first course:

Introduction to Robotics & Visual Navigation.

This course covers Robotic perception, planning, localization, controls and dynamics. The course also presents simulations of some algorithms in CARLA simulator, and is ideal for people wanting to make a career in Industry 4.0 and Robotics.

DeepEigen courses are typically much broader and deeper when compared to other popular online platforms offering micro or nano degrees. Furthermore, our fee is also substantially lower as compared to the humongous fee charged by other platforms (who offer micro / nano degrees) for content that is of little to no practical use.

Course Link: http://www.deepeigen.swaayatt-robots.com/introductiontorobotics_coursedetails

Website Link: www.deepeigen.swaayatt-robots.com

26/04/2020

Sanjeev Sharma was interviewed by Eddie Avil of Chang I am Possible for Swaayatt's work in autonomous driving technology.

In this interview, Sanjeev extensively discussed the use of reinforcement learning and some of key concepts like Homotopy maps, and other mathematical topics to enable autonomous driving in environments with traffic dynamics as stochastic and as complex as the Indian traffic dynamics.

He also discussed how we developed the technology superior to other Multi-Billion Dollar companies like Waymo, , , Uber and the likes. He also discussed how we have build robust tech when compared to heavily funded startups like Zoox, Wayve etc.

Sanjeev also covered how Swaayatt Robots is developing the technology that makes the requirement of high-fidelity maps for autonomous driving totally redundant.

The interview is mathematical at times, but is easily understandable, and covers the superiority and novelty of algorithms developed by Swaayatt in Planning and Perception, as compared to the competitors abroad.

The goal of Swaayatt Robots is to developed autonomous driving technology that is tested in the Indian environments, so that it can ensure the safety of the vehicle and the environment when deployed in the developed countries.

Sanjeev Sharma is the Founder-Ceo of an Auto...

20/04/2020

Our DGN-I working on Indian roads, for use in Autonomous Driving and ADAS applications.

Computations: 15 GFlops

This makes it the fastest DNN that performs two tasks simultaneously:
- Semantic Segmentation for Road Detection
- Bounding Box Proposal for Obstacle Detection

We are working on DGN-RI, a recurrent version, which should set a new benchmark, if DGN-I doesn't already.

PMO India Narendra Modi our autonomous driving technology can be helpful for Civil and Defence applications, including for Covid-19 Fight.

MYogiAdityanath Ministry of Home Affairs, Government of India DRDO





Sanjeev Sharma Shani Sharma

10/04/2020

Sample demo of our "Deep Geometric Network" (DGN) with Integrated Obstacles' Bounding Box Proposal Pipeline (DGN-I).

DGN is our generalized semantic segmentation network which was developed in 2016.

DGN does 12.47 GFlops for segmentation for road detection for autonomous driving. ADAS version is rated at 5.2-10.6 GFlops.

This demo presents our DGN with Integrated Obstacles' Bounding Box Proposal Pipeline (DGN-I) using only off-the-shelf security IP cameras (less than USD 40).

All this is done (current version) in 15 GFlops. It does 2 tasks:
- Segments road (Segmentation mask)
- Detects obstacles (Bounding Box)

Previous demo of DGN showed it working at day, night and in cluttered shadow conditions. This demo shows the capability of DGN-I with only 15 GFlops computation. Our upcoming demo will show both the ADAS version and demo over several KMs (>100) of road.

DGN-I shares the same property as DGN (due to same backbone): A key feature is that it allows us to increase the number of computations while reducing the memory super-linearly (until a threshold) w/o affecting accuracy. This makes it suitable for embedded platforms.

Our team is currently also researching on making DGN-I more robust with addition of an RNN layer in it, it will be called DGN-RI.

, , ,

A sample demo of our Deep Geometric Network with Integrated Bounding Box Proposal Pipeline (DGI-I). It is rated at 15 GFlops, making it the fastest network t...

09/08/2019

We now at PoC level have the technology to perceive environments for autonomous driving and ADAS tasks using both the NIR and RGB cameras.

Here's the demo of our lane detection and generation algorithm working at night. It detects lane markers if they are present on road, and generates them automatically if they are not present.

The cameras are regular off-the-shelf security IP cameras mounted on the front bumber of our autonomous vehicle Swaayatt-1.

For us, we consider an algorithm real-time if it does less than 15 Billion Flops per image.