| Rigorous mathematical simulation verification
| Unpredictable driving decision-making algorithm
| Protecting your vehicle from real-world road traffic conditions…
| SFF computations performed on vehicle sensor data

NVIDIA announced today that it has further enhanced its NVIDIA DRIVE AV autonomous vehicle software suite with a control layer designed to deliver a safe and comfortable driving experience.

A key component of this software is the Safety Force Field (SFF), a robust driving strategy that protects the vehicle, its occupants, and other road users.
SFF collects sensor data and analyzes and predicts the movements of the surrounding environment, determining a series of actions to protect vehicles and other road users. The SFF framework ensures that unsafe situations do not occur or lead to unsafe conditions during operation and includes measures necessary to mitigate potential risks.

With powerful computational capabilities, SFF enables vehicles to maintain safety based on mathematical verification of zero collisions, rather than attempting to model the high complexity of real-world scenarios with limited statistics. Running on the NVIDIA DRIVE platform, frame-by-frame, physics-based SFF computations are performed on vehicle sensor data.

SFF has also undergone real-world data and bit-accurate simulation validation, including scenarios related to highway and urban driving that are difficult to replicate in the real world.


Mitigating risk situations and preventing complete collisions
With the U.S. National Highway Traffic Safety Administration (NHTSA) estimating that approximately 94% of crashes are caused by human error, the industry is striving to achieve a "zero accident rate" through the development of autonomous vehicles. Enhanced Drive AV platforms, with safety features such as path planning, prediction, and SFF, aim to achieve this goal.

A unique feature of SFF is its ability to consider both braking and vehicle steering constraints. This dual functionality helps prevent vehicle behavioral anomalies that could occur if operated separately. This policy follows one core principle: conflict prevention, rather than large-scale rules and predictions.

“By removing the human error factor from driving, we can prevent the vast majority of crashes and minimize their impact,” said David Nister, vice president of autonomous driving software at NVIDIA. “Autonomous vehicles using SFF are mathematically designed to act like magnets, protecting themselves from dangerous situations and not causing them.”


Transparent open platform
SFF is an open platform that can be integrated with other driving software. Serving as a safe decision-making policy within the motion planning stack, SFF monitors and prevents unsafe behavior. It also clearly distinguishes obstacle avoidance from complex and extensive road rules.

When running on high-performance computing platforms like NVIDIA Drive, it provides a high level of safety by adding another layer of diversity and redundancy features.