Apollo-ADS_路径规划&决策 模块 _study [ADS]

REF:

source code :apollo/tree/master/modules/planning

https://github.com/ApolloAuto/apollo/blob/master/docs/specs/Class_Architecture_Planning_cn.md

https://github.com/ApolloAuto/apollo/tree/master/modules/planning README.md

---

改变：

• 3.0以前，Apollo 用一套规划算法处理所有case.

 

• 3.5之后，改成场景base 的规划算法，scenario-based planning

针对不同场景（车道中心保持，停车（车库停车，靠边停车 ...），掉头，十字路口(有红绿灯，无红绿灯 ...），选用不同的规划算法，生成profile，再执行。

其中decider --是基于规则的规划--也不是最理想的，会有bug，有遗漏...

• Apollo FSM: A finite state machine that determines the vehicle state given its location and routing together with HD Map.
  • 有限状态机，还是有缺陷的，bug易发。
• Planning Dispatcher: Call an appropriate planner given the vehicle's state and some other relevant information
• Planner: Obtain the required context data and other information, decide a corresponding vehicle intention, execute the required planning tasks for this intention and generate the planning trajectory. It will also update the context for future jobs.
• Deciders & Optimizers: A set of stateless libraries that implement decision tasks and various optimizations. Optimizers specifically optimize the vehicle's trajectory and speed. Deciders are rule-based decision makers that