autowarefoundation · kminoda · Jun 1, 2022 · May 30, 2022 · May 30, 2022 · May 31, 2022
diff --git a/docs/how-to-guides/.pages b/docs/how-to-guides/.pages
@@ -1,4 +1,5 @@
 nav:
   - index.md
+  - running-autoware-without-cuda.md
   - advanced-usage-of-colcon.md
   - how-to-integrate-autoware-with-your-vehicle.md
diff --git a/docs/how-to-guides/index.md b/docs/how-to-guides/index.md
@@ -1,5 +1,6 @@
 # How-to guides
 
+- [Running Autoware without CUDA](running-autoware-without-cuda.md)
 - [Advanced usage of colcon](advanced-usage-of-colcon.md)
 - [Integrate Autoware with your vehicle](how-to-integrate-autoware-with-your-vehicle.md)
 

diff --git a/docs/how-to-guides/running-autoware-without-cuda.md b/docs/how-to-guides/running-autoware-without-cuda.md
@@ -0,0 +1,34 @@
+# Running Autoware without CUDA
+
+Two following algorithms in Autoware Universe depend on a CUDA environment.
+
+- 2D/3D object detection
+- traffic light detection
+
+However, it is still possible to run Autoware in the environment without CUDA.
+The following subsections briefly explain how to run these algorithms in such an environment.
+
+!!! warning
+
+    It is highly recommended to prepare a CUDA environment to achieve better performance.
+
+## Running 2D/3D object detection without CUDA
+
+As for the object detection system in Autoware, the following three packages require the CUDA environment:
+
+- `lidar_centerpoint`
+- `lidar_apollo_instance_segmentation`
+- `tensorrt_yolo`
+
+Instead of these modules, you can use the `euclidean_cluster` module, which is a 3D object detection algorithm that does not require a CUDA environment (refer to [the readme](https://github.com/autowarefoundation/autoware.universe/tree/main/perception/euclidean_cluster) for detail).
-Instead of these modules, you can use the `euclidean_cluster` module, which is a 3D object detection algorithm that does not require a CUDA environment (refer to [the readme](https://github.com/autowarefoundation/autoware.universe/tree/main/perception/euclidean_cluster) for detail).
+Instead of these modules, you can use the `euclidean_cluster` module which contains a 3D object detection algorithm that does not require CUDA. For more details, refer to the [`euclidean_cluster` module's README file](https://github.com/autowarefoundation/autoware.universe/tree/main/perception/euclidean_cluster).
-Instead of these modules, you can use the `euclidean_cluster` module, which is a 3D object detection algorithm that does not require a CUDA environment (refer to [the readme](https://github.com/autowarefoundation/autoware.universe/tree/main/perception/euclidean_cluster) for detail).
+Instead of these modules, you can use the `euclidean_cluster` module which contains a 3D object detection algorithm that does not require CUDA. For more details, refer to the [`euclidean_cluster` module's README file](https://github.com/autowarefoundation/autoware.universe/tree/main/perception/euclidean_cluster).
+
+## Running traffic light detection without CUDA
+
+As for the traffic light recognition system in Autoware, there are two package that require the CUDA environment:
+
+- `traffic_light_ssd_fine_detector`
+- `traffic_light_classifier`
+
+To run traffic light detection without CUDA, set [`enable_fine_detection` in this file](https://github.com/autowarefoundation/autoware.universe/blob/9445f3a7acd645d12a64507c3d3bfa57e74a3634/launch/tier4_perception_launch/launch/traffic_light_recognition/traffic_light.launch.xml#L3) to `false` so that `traffic_light_ssd_fine_detector` be disabled and the traffic light detection solely be executed by `map_based_traffic_light_detector`.
+
+To run traffic light classification without CUDA, set [`use_gpu` in this file](https://github.com/autowarefoundation/autoware.universe/blob/9445f3a7acd645d12a64507c3d3bfa57e74a3634/perception/traffic_light_classifier/launch/traffic_light_classifier.launch.xml#L7) to `false` so that `traffic_light_classifier` use classification algorithm that does not use CUDA environment or GPU.