diff --git a/README.md b/README.md
index ae74979f539f97762b87caac28c98a31559d857b..298d63be73c09ab6023a1ecbfcd372cc5004a188 100644
--- a/README.md
+++ b/README.md
@@ -38,8 +38,9 @@ This project is the [ROS](https://www.ros.org/) implementation of the Robot Beha
     </li>    
     <li><a href="#usage">Usage</a>
       <ul>
-        <li><a href="#gestures-action-server">Gestures Action Server</a></li>
-        <li><a href="#gestures-action-client">Gestures Action Client</a></li>
+        <li><a href="#look_at-actions">look_at actions</a></li>
+        <li><a href="#go_to-actions">go_to actions</a></li>
+        <li><a href="#local-path-planner-selection">Local Path Planner Selection</a></li>
       </ul>
     </li>
     <li><a href="#roadmap">Roadmap</a></li>
@@ -134,6 +135,9 @@ The Robot Behavior project is an [actionlib](http://wiki.ros.org/actionlib) base
 
 .
 
+
+### 
+
 Now that all the ROS modules required for the navigation are running, you will be able to launch all the Robot behavior actions.
 
 ### look_at actions
@@ -374,8 +378,19 @@ Four types of `go_to` actions are possible:
 
 __note__: One `look_at` action and one `go_to` action can be run at the time. If a `go_to` actions is launched while one is running, the running action will be canceled and the new one will be run.
 
-###
+### Local Path Planner Selection
+In order to select the  Deep Reinforcement Learning (DRL) local path planning algorithm or he Model Predictive Control (MPC) local path planning algorithm, you can set a parameter using [rosparam](http://wiki.ros.org/rosparam):
 
+```sh
+rosparam set /planning_manager/local_path_planner_name social_mpc
+```
+use `social_mpc` you want to choose the MPC algorithm or `social_rl` you want to choose the DRl algorithm. By default, the DRL is used.
+
+In order to check the local path planner algorithm, type:
+```sh
+rosparam get /planning_manager/local_path_planner_name
+```
+the output should be `social_mpc` or `social_rl`.
 
 Below, we can see two resulting examples of the Behavior Generator running in real time in ARI robot through ROS: