Abstract
To evaluate the possibility of collision in the driving environment of handle type electric wheelchairs, we suggested the collision predictive index based on the analysis of the video image data of driving handle type electric wheelchairs and examined the validity of it in comparison with the risk perception data obtained from expert sales persons. By using this index, we can quantitatively evaluate the possibility of collisions with pedestrians, bicycles and the rushing out from blind spots. As a result of examination of validity, we found that the collision predictive index could correctly identify 63% of objects within the range of the scenes we analyzed.
