GM's new safety system uses gait analysis to prevent impaired driving

General Motors has filed a patent for a new safety system that can analyze a driver's condition before they even get behind the wheel. The technology uses cameras and sensors to track a person's gait as they approach the vehicle.

The system evaluates several parameters simultaneously: walking speed, stride length, stability, and deviations from a straight path. Based on this data, it generates a so-called "gait score" that reflects the likelihood of impaired coordination—due to alcohol or fatigue, for instance.

Information is processed using machine learning algorithms, including LSTM neural networks. If the system detects deviations from the norm, the car can respond in various ways: it might warn the driver, suggest an additional check, or completely block the ability to operate the vehicle.

Additionally, the technology can determine whether a person actually intends to drive—for example, by detecting the opening of the driver's door. This reduces the risk of false alarms and makes the system more accurate. In practice, this represents a new level of active safety, where the car begins monitoring the driver's state even before the engine starts. Such solutions could become an important step in reducing accident rates and tightening safety standards.

If this technology reaches production models, cars could become a full-fledged filter against dangerous drivers. However, this raises privacy concerns—and that issue might become the main obstacle to widespread adoption.