Steering Holding Determination Using Kalman Filter Variance

The patent replaces complex additional sensors with a computational approach using a Kalman filter algorithm. The system substitutes physical measurement devices with a mathematical model that processes existing sensor data (steering angle, steering torque, wheel acceleration) to estimate driver holding state, thereby achieving high measurement precision without increasing hardware complexity

The patent introduces a Kalman filter as an intermediary computational layer between the physical sensors and the steering control system. This intermediary processes raw sensor data through prediction and update steps, extracting meaningful information about driver holding state from noisy measurements without requiring direct physical contact or additional sensors

The patent enables the existing steering system to self-diagnose driver holding state using its own operational data. The Kalman filter utilizes steering angle, steering torque, and wheel acceleration data already generated during normal operation to determine whether the driver is holding the steering wheel, eliminating the need for external detection sensors

The patent makes existing sensors serve multiple functions: steering angle sensors provide both steering control data and driver holding detection data; steering torque sensors provide both assist control information and driver input detection; wheel acceleration sensors provide both stability control data and holding state verification. This multi-functionality achieves high detection accuracy without adding dedicated sensors