With　the　rapid　development　of　intelligent　driving　technology,　people　have　higher　requirements　for　the　safety　of　vehicle　driving.　The　framework　of　this　paper　is　divided　into　two　layers.　The　upper　layer　is　responsible　for　the　path　planning　of　the　driving　trajectory,　and　the　lower　layer　is　the　trajectory　tracking　control.　In　the　path　planning　layer,　an　obstacle　model　based　on　the　potential　energy　function　is　proposed,　while　considering　the　sensing　range　of　the　vehicle,　so　that　the　vehicle　can　make　an　obstacle　avoidance　response　in　advance.　The　search　efficiency　is　improved　by　combining　PSO　to　segment　the　trajectory.　The　differential　flatness　of　the　intelligent　vehicles　is　used　to　smooth　the　whole　trajectory　and　segment　boundaries,　making　the　driving　more　comfortable.　The　trajectory　tracking　layer　utilizes　the　kinematic　model　and　ensures　that　the　inputs　are　within　the　control　boundaries.　Simulation　and　real　experiment　based　on　the　Quanser　self-driving　car　(QCar)　have　been　accomplished　to　prove　the　effectiveness　of　the　proposed　algorithm.　©　2023　IEEE.