Mbp-728x90
Location: Am Kielortplatz 2, 22850 Norderstedt, Hamburg, Germany Germany
Founded in: 2016
Stage: Pre-launch
Number of employees: 1-5
Short URL: vator.co/intellimo
Mbp-300x250

intellimo

Hyper-realistic simulation environment for autonomous vehicle validation and driving-algorithm creating and benchmarking.
Startup/business
Hamburg, Germany Germany
http://intellimo.com
  • About
  • >
  • Activities
Company description

Executive Summary:

intellimo aims to develop a highly realistic 3D-simulation environment for the economically reasonable development and validation of autonomous cars in order to achieve provably safe autonomous cars. We want to help OEMs and their suppliers to rapidly accelerate their development and testing/validation processes for autonomous cars and (therefore) reduce resource input and time to market for safe autonomous driving Level 5 by 2021.

 

We’re initially targeting the car OEM and supplier industry. We plan to make our money by collecting a SaaS fee and we will acquire customers by our automotive and suppliers network.


We have an experienced team and core network with deep domain expertise, unique technology and a significant first mover advantage by capturing stereo images. A proof of concept for this has been developed by us and a functioning prototype is currently built around this.

The Problem:

Autonomous driving must be unprecedentedly and provably safe, therefore OEMs and suppliers will have to test their hard- and especially software during billions of miles involving billions of development expenses.

Safety - Autonomous driving must be unprecedentedly and provably safe, therefore OEMs and suppliers will have to test their hard- and software over and over again - in the range of 6.6 billion road miles with real vehicles.

Expenses – For the testing of an autonomous car prototype for only 1 billion kilometers alone, OEMs would have to operate 4000 real test-vehicles in parallel over the regular lifespan of a car. This is related to billions of dollars in development and validation expenses for vehicle-prototypes, test-drivers, test-execution and iteration as well as documentation.

Time to market – Given the fact of the above roadblocks, time to market will exceed the mark of 2021 or 2025 for autonomous driving by far and can only be accelerated in a reasonable, economic manner by applying simulation solutions.

The Solution:

intellimo develops a highly realistic 3D simulation environment for developing and validating/testing autonomous cars as well as its sensors x-times faster than in real-time and highly cost efficient, thus allowing to prove their high safety.

Complex Environment Simulation – intellimo aims to train and benchmark new algorithms for an end2end learning Advanced Driver Assistant System (ADAS) and to screen existing algorithms for rare malfunctions faster than real-time and in parallel within a simulation environment of realistic complexity and graphical fidelity.

Stereo-Camera Simulation – intellimo is able to perform stereo-capture in order to simulate stereo camera sensors.

Depth-Buffer reconstruction with Convolutional Neural Networks – the stereo capturing ability is used to actively research the reconstruction of depth-information from stereo-images.

Sensor Input Simulation – intellimo aims to simulate the specific sensor-data from Camera, LIDAR, RADAR and ultrasound to provide a Framework capable of training and validating new and existing algorithms within a realistic simulation.

Competitive Advantages:

Resource/DEVEX reduction – intellimo can enable OEMs as well as TIER1 suppliers to reduce their development and especially testing expenses necessary to provide provably safe autonomous driving by 2021/25

Proving and Improving safety – using intellimo’s highly realistic simulation environment it will be possible to prove the failure resistance and thus safety of autonomous cars. Furthermore by simulating a broader range of scenarios including those that cannot be replicated in reality due to technical or ethical concerns it will be possible to further improve safety. 

Accelerating time to market safely – intellimo’s approach to testing and validating ADAS for billions of test-kilometers in super-realtime and in parallel through simulation will not only reduce costs but accelerate time to market for the safest autonomous driving possible as well as mitigate the risk of accidents during testing.

Stereo-Camera Simulation – in comparison to potential competition, intellimo is able to perform stereo-capture in order to simulate stereo cameras for depth reconstruction.

Depth-Buffer reconstruction with Convolutional Neural Networks – intellimo will also utilize the ability to capture highly realistic stereo images from simulation in order to research and develop promising depth buffer reconstruction with convolutional neural networks

Sensor Input Simulation – intellimo aims to simulate the specific sensor-data from Stereo Camera, LIDAR, RADAR and ultrasound to provide a comprehensive development and validation framework for autonomous driving algorithms.

 

 

 

 

 

 

 

 

 

 

 

 

 


Team