You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The nuPlan Planning Challenge is now live! The warm-up stage has begun. Please check our competition landing [page](https://nuplan-devkit.readthedocs.io/en/latest/).
33
+
## Planning challenges
34
+
#### The nuPlan Planning Competition has been postponed to January 2023.
35
+
#### The nuPlan Dataset v1.1 has been released. Please download the latest dataset from the nuPlan [page](https://www.nuscenes.org/nuplan#download).
Please refer to the [installation page](https://github.com/motional/nuplan-devkit/blob/master/docs/installation.md) for detailed instructions on how to setup the devkit.
63
+
Please refer to the [installation page](https://nuplan-devkit.readthedocs.io/en/latest/installation.html) for detailed instructions on how to setup the devkit.
58
64
59
-
Please refer to the [dataset page](https://github.com/motional/nuplan-devkit/blob/master/docs/dataset_setup.md) for detailed instructions on how to download and setup the dataset.
65
+
Please refer to the [dataset page](https://nuplan-devkit.readthedocs.io/en/latest/dataset_setup.html) for detailed instructions on how to download and setup the dataset.
Training configurations are important to ensure your expected system performance, for example preprocessing cost, training speed, and numerical stability. If you encounter problems related to aforementioned aspects, please refer to [performance tuning guide](https://github.com/motional/nuplan-devkit/blob/master/docs/performance_tuning_guide.md) to find potential solutions.
We provide several baselines within the devkit. These baselines are standard comparison points in which to compare new
4
+
planners. Moreover, the baselines serve as a starting point for users to prototype their planner or simply tinker with it.
5
+
6
+
## SimplePlanner
7
+
The SimplePlanner, as the name suggests, has little planning capability. The planner plans a straight line at a constant
8
+
speed. The only logic of this planner is to decelerate if the current velocity exceeds the `max_velocity`.
9
+
10
+
Link to the [code](https://github.com/motional/nuplan-devkit/blob/master/nuplan/planning/simulation/planner/simple_planner.py)
11
+
12
+
## IDMPlanner
13
+
The Intelligent Driver Model Planner (IDMPlanner) consists of two parts:
14
+
15
+
1. Path planning
16
+
2. Longitudinal control (IDM policy)
17
+
18
+
### Path planning
19
+
The path planning is a breadth-first search algorithm. It finds a path towards the mission goal.
20
+
The path consists of a serie of lane and lane connectors that leads to the roadblock containing the mission goal.
21
+
The baseline is then extracted from the found path and is used as the reference path for the planner.
22
+
23
+
### IDM Policy
24
+
Now that the planner has a reference path, it must then decide how fast to go along this path. For this, it follows
25
+
the [IDM policy](https://en.wikipedia.org/wiki/Intelligent_driver_model). The policy describes how fast the planner should
26
+
go based on the distance between itself and a given agent. Of course, it is wise to choose the closest agent in the path of the planner.
27
+
28
+
Hence, the IDMPlanner uses breadth-first search to find the path towards the mission goal, and the IDM policy describes how far along that path the planner should be.
29
+
30
+
Link to the [code](https://github.com/motional/nuplan-devkit/blob/master/nuplan/planning/simulation/planner/idm_planner.py)
0 commit comments