An approach to understanding driving environment using network-type sensor fusion method

A method for recognizing a driving environment by fusion processing using the concept of an immune network, which is a distributed diagnostic model, has been studied. The method classifies a recognition system into multiple modules from the point of view of sensor processing software, connects them into a network, and relates the network states obtained by interactive tests with a driving environment. As an experimental example, the lane of a car in front is determined. The method is advantageous in that constructing fusion processing is easy without requiring changes in lower-level processing, because it applies interactive diagnostics only to upper-level processing.