Making Autonomous Vehicles Safe

by bringing the human element into traffic simulation.

Highly realistic traffic simulation

Based on our cognitive AI, cogniBIT develops algorithms simulating complex human road user behavior. Our patent-pending cogniBOT™ AI architecture realistically simulates how humans perceive, think, and act.

edward-ma-1bBRjHcMZNY-unsplash

Our technology supports all stages of the ADAS and AD development process

Research

Train your AI algorithms with automatically generated life-like data, e.g. for the development of AD motion planning.

Criticality analysis

Identify critical edge cases scenarios such as pre-crash or driver hand-over.

Development

Integrate our traffic models into your CI for continuous function testing in your development process.

Verification & Validation

Our interactive traffic models support virtual endurance tests as well as scenario-based testing.

Certification

Challenge AD systems in simulation with highly realistic human road users.

Highly Realistic Traffic Simulation


Instead of using pregenerated trajectories, our simulated road users realistically interact with the vehicle under test at runtime. Just like humans, our simulated road users have typical human limitations such as a limited field of view, reaction times or lack of attention. Factors such as age and emotions can be adapted for each simulated individual driver to represent the continuum of human behavior in road traffic.

city traffic , Cars and people on street aerial

Validated

Our models produce fully explainable behavior and are validated by comparing data resulting from our simulation with real-world data, such as traffic statistics or physiological data.

Bar graph and line graph templates, business infographics, vector illustration. Graphs and charts set. Statistic and data, information infographic

Automated Data Generation

Our models are deterministic in the sense of generating fully reproducible results. At the same time, they contain stochastic elements representing uncertainties in human information processing. This automatically generates valid and realistic scene variations without the need for manual trajectory adjustment or real-world data acquisition.

traffic-top-view

Easy Integration

Our road user models are platform-independent and support common industry standards. This allows easy integration into any simulation platform.

IT Technician with a Laptop Computer and Black Male Engineer Colleague are Talking in Data Center while Walking Next to Server Racks. Running Diagnostics or Doing Maintenance Work.

Our cognitive models display highly realistic modes of behavior

edward-ma-1bBRjHcMZNY-unsplash

Highly realistic traffic simulation

Instead of using pregenerated trajectories, our simulated road users realistically interact with the vehicle under test at runtime. Just like humans, our simulated road users have typical human limitations such as a limited field of view, reaction times or lack of attention. Factors such as age and emotions can be adapted for each simulated individual driver to represent the continuum of human behavior in road traffic.

Validated

Our models produce fully explainable behavior and are validated by comparing data resulting from our simulation with real-world data, such as traffic statistics or physiological data.

edward-ma-1bBRjHcMZNY-unsplash
edward-ma-1bBRjHcMZNY-unsplash

Automated data generation

Our models are deterministic in the sense of generating fully reproducible results. At the same time, they contain stochastic elements representing uncertainties in human information processing. This automatically generates valid and realistic scene variations without the need for manual trajectory adjustment or real-world data acquisition.

Easy integration

Our road user models are platform-independent and support common industry standards. This allows easy integration into any simulation platform.

edward-ma-1bBRjHcMZNY-unsplash