Reference:
https://github.com/ericlavigne/CarND-Path-Planning


Overview:
 1. astar.h  Applied to Trajectory State for 4-dimensional path planning: s, d, v, t
 2. discrete_prediction.cpp
     Predicts future positions of other cars by assuming that they will move at constant speed within their current lanes.
     Cars that are changing lanes are conservatively predicted to be in both lanes two seconds later.
 3. trajectory_state.cpp
     Represents 4-dimensional car state (s,d,v,t) for A* algorithm in context of discrete trajectory planner.
     Includes basic car physics: acceleration, deceleration, and lane changing
     Includes goal of advancing along the highway
     Applies penalties for crashing or tailgating
 4. discrete_trajectory_planner.cpp
     Discretization with 1 meter precision along the road and 2 meter precision across the road to reduce A* search space
 5. trajectory_planner.cpp
     Converts continuous data about self and other cars into discrete representation for Discrete Trajectory Planner.
     Converts discrete representation back to continuous representation for use in controlling car
 6.  track.cpp
      Converts between cartesian (x,y) and frenet (s,d) coordinates.
      Needed because trajectory planning uses frenet while sensors and controller use cartesian.
      Uses splines for much better accuracy compared with linear interpolation
 7. controller.cpp
     Directly controls speed and path of car to follow waypoints from Trajectory Planner.
     Determines acceleration based on assumption of constant acceleration between current state and one second later in the Trajectory Planner's waypoints.
     Determines precise path in 0.02 second increments by applying spline from current state through all of the Trajectory Plannner's waypoints.

视频展示:
   GoogleDrive:
    https://drive.google.com/open?id=1hFwpXNSZrBgeeeX0J-EVrTKNLbZ_PONm
   云盘:
    https://www.jianguoyun.com/p/DdkOloMQvsmaCBjDqvUC