A Study about Curve Extraction and Lane Departure Determination of Linear Curved Road

Recently, intelligent vehicle are required accurate information about road condition for safety driving. However, vehicle safety systems are irregularly caused an error in judgment by road patterns. In this study, which has extracted curve road segments from line information of GIS with alignment of road. This algorithm are based modified Bezier curve method to detect curve lines from straight lines. Extracted line information are formally analyzed with lane departure warning system (LDWS) that was robust curves. For performance evaluation, the system has installed DGPS with 20cm RMS and WAVE communication with 5.8GHz, and implemented vehicle test in road environment with groups of curves and straight lines. Curvature LDWS improve correctness criterion by applying to predicted trajectory condition to previous algorithm. Test results are shown robust determination and confirmed false alarm did not appear. Furthermore, this algorithm are not only reduced error of functioning properly, but also detected lane departure at all-road environments in real-time.