基于分段多项式曲线的自动泊车系统速度规划OA

To address the comprehensive requirements of automatic parking systems for comfort,high-precision control,and scenario adaptability in complex low-speed environments,this paper proposes a longitudinal speed planning algorithm based on segmented quintic polynomial curves.A three-phase planning framework,comprising"maximum speed cruising,target speed cruising,and braking-to-stop"is built,with candidate quintic polynomial curve sets generated for each phase.An adaptive rapid optimization strategy prioritizing jerk limitation as the primary constraints is designed.Real-world vehicle experiments are conducted across three typical scenarios:flat roads,speed bumps,and slopes.Results demonstrate the algorithm consistently constrains jerk magnitude within±0.5 m/s3 and achieves a distance control accuracy of less than 5 cm under comfortable braking conditions.It markedly outperforms conventional single-segment planning algorithms,providing a highly comfortable,precise,and easily deployable solution for automatic parking systems.