While some companies are already focusing their efforts on (semi-)autonomous driving, in which it does not matter what the driver’s condition is, the American car manufacturer General Motors has filed a patent application for an advanced system that uses cameras, lidars and advanced sensors to assess whether the driver is capable of driving the vehicle.
AI stagger detection
The system, as stated in the recently published patent, is designed to follow a person as they approach the car. In doing so, it monitors and analyzes specific walking parameters with its sensors and cameras, such as movement speed, stride length or path stability, i.e. it detects if the driver stumbles when approaching the car.
The technology relies on machine learning, specifically a recursive neural network, which processes the collected data and generates a “gait score”. If this rating exceeds a predefined safety limit, the system activates countermeasures to prevent a potentially dangerous situation on the road. In cases where the system estimates that the driver is under the influence of alcohol or drugs, several levels of action are foreseen. These include visual and audible warnings, activation of secondary systems such as breathalyzers, or, in the last resort, complete denial of access to the vehicle.
GM emphasizes the importance of transparency of the system so that the user is clearly informed of the blockage while still outside the car, which reduces the possibility of attempts to bypass the protection. In addition to the detection of alcohol intoxication, the patent also mentions a potential application in the recognition of emergency medical conditions. A set of built-in sensors could identify health problems that the driver might not even be aware of yet, so this system takes on a wider role in protecting the user’s health, not just preserving traffic safety and saving the driver’s license.
Although the patent was filed in September 2024 and recently published, there is still no official confirmation as to whether, when and in which production models this technology will appear.