--- a
+++ b/Real_Time_Prediction_From_Microcontroller.py
@@ -0,0 +1,97 @@
+import socket
+import msvcrt
+import csv
+import pandas as pd 
+import tensorflow as tf 
+import matplotlib as plt
+import glob
+import numpy as np 
+from tensorflow import keras
+from keras import Sequential
+from sklearn.utils import shuffle
+import sklearn.model_selection
+
+STEP_SIZE = 20
+SENSOR_NUM = 6
+
+
+Label = { 'STD':0, 'WAL':1, 'JOG':2 , 'JUM':3, 'FALL':4 , 'LYI':5,'RA':6} #, 'JUM':3, 'LYI':4, 'FOL':5, 'FKL':5, 'BSC':5, 'SDL':5, 'STU':6, 'STN':7, 'SCH':8, 'SIT':9, 'CHU':10, 'CSI':11, 'CSO':12}
+class_names = { 0:'STD', 1:'WAL', 2:'JOG' , 3:'JUM', 4:'FALL', 5:'LYI',6:'RA'}#, 3:'JUM', 4:'LYI', 5:'Falling', 6:'STU', 7:'STN', 8:'SCH', 9:'SIT', 10:'CHU', 11:'CSI', 12:'CSO'}
+
+inputSensor = []
+model = tf.keras.models.load_model('./model_4/')
+
+model.summary()
+
+
+l = []
+packet = []
+p = ""
+tp = []
+s = socket.socket()         
+ 
+s.bind(('0.0.0.0', 80 ))
+s.listen(0)    
+i = 0      
+
+record = input("Press R to start recording...")
+if record == 'r' :
+        
+
+    # while True:
+    
+    client, addr = s.accept()
+    #     x = kbfunc()
+
+    carry = ''   
+    try:
+        while True:
+            content = client.recv(1)
+            
+
+            if len(content) ==0:
+                # print("heree?")
+                break
+
+            else:
+                #print(len(content))
+              
+                temp = ''
+                content = content.decode("utf-8")
+
+                if content == '!':
+                    
+                    tp =[]
+                    p = ''
+
+                
+                elif content == '@':
+                    tp.append(float(p))
+                    # print(type(tp[0]))
+                    p = ''
+                    inputSensor.append(tp)
+                    if len(inputSensor) > STEP_SIZE:
+                        inputSensor.pop(0)
+                    if len(inputSensor) == STEP_SIZE:
+                        temp = np.array(inputSensor).reshape(-1, STEP_SIZE, SENSOR_NUM)
+                        #print(temp)
+                        pred = model.predict(temp)
+                        results = np.argmax(pred, axis=1)
+                        print("prediction: ", class_names[results[i]])
+                        # inputSensor = []
+                    
+
+
+                elif content ==',':
+                    p = float(p)
+                    p = p
+                    tp.append(p)
+                    p = ''
+
+                else:
+                    p += content
+
+    except KeyboardInterrupt:
+        pass   
+    
+