diff --git a/dl-with-carla/README.md b/dl-with-carla/README.md
index 868eeea..f2a4111 100644
--- a/dl-with-carla/README.md
+++ b/dl-with-carla/README.md
@@ -1,17 +1,22 @@
 # Key concepts
 
-- To obtain driving data, we used Carla, an open-source simulator.
-- We employed two main methods to acquire driving data from Carla:
-    1. Automatic Control
-        
-        Using the autonomous driving code implemented by Carla, we save five types of data while the vehicle is in motion: speed, steering, throttle, brake, and lane invasion. These data types are stored according to the VSS standard. Depending on the need, this data can be formatted into CSV or input into a virtual CAN.
-        
-    2. Manual Control
-        
-        You can manually drive according to Carla's physics engine, although it is quite challenging. Like in automatic control, the same five data types are saved.
-        
-- LSTM Learning Model
+## 1. Make data for model
+
+To obtain driving data, we used Carla, an open-source simulator. We employed two main methods to acquire driving data from Carla:
+
+- Automatic Control
+    
+    Using the autonomous driving code implemented by Carla, we save five types of data while the vehicle is in motion: speed, steering, throttle, brake, and lane invasion. These data types are stored according to the VSS standard. Depending on the need, this data can be formatted into CSV or input into a virtual CAN.
+    
+- Manual Control
     
-    We utilized an LSTM model to calculate driving scores. We chose this model because of its strength in handling time-series data, which is crucial since our CAN data is time-dependent. For more details on the model, please refer to lstm-training.py.
+    You can manually drive according to Carla's physics engine, although it is quite challenging. Like in automatic control, the same five data types are saved.
     
-    When the five types of data form a sequence of 100, a corresponding score is assigned. With 300,000 CAN data points, a total of 300 data sequences have been trained. Although it's not a vast amount of data, it should be sufficient to represent your driving score!
+
+## 2. LSTM Learning Model
+
+![lstmmodel.png](./images/lstmmodel.png)
+
+We utilized an LSTM model to calculate driving scores. We chose this model because of its strength in handling time-series data, which is crucial since our CAN data is time-dependent. For more details on the model, please refer to lstm-training.py.
+
+When the five types of data form a sequence of 100, a corresponding score is assigned. With 300,000 CAN data points, a total of 300 data sequences have been trained. Although it's not a vast amount of data, it should be sufficient to represent your driving score!"
\ No newline at end of file
diff --git a/images/lstmmodel.png b/images/lstmmodel.png
new file mode 100644
index 0000000..32839a6
Binary files /dev/null and b/images/lstmmodel.png differ