Sensor Validation and Fusion for Automated Vehicle Control Using Fuzzy Techniques

This brief introduces a fuzzy sensor validation and fusion methodology and applies it to automated vehicle control in Intelligent Vehicle Highway Systems (IVHS). Sensor measurements are assigned confidence values through sensor-specific dynamic validation curves. The validation curves attain minima of zero at the boundaries of the validation gate. These in turn are determined by the largest physically possible change a system—in our example vehicles of the IVHS—can undergo in one time step. A fuzzy exponential weighted moving average time series predictor determines the location of the maximum value of the validation curves. Sensor fusion is then performed using a weighted average of sensor readings and confidence values, and—if available—the functionally redundant values calculated from other sensors.