y_nnn : change learn_to_drive and add some bash print

deepQlearn_0/src/deepQlearning/learn_to_drive.c

@@ -130,6 +130,7 @@ struct print_params * create_print_params(float scale_x, float scale_y,  struct
   pprint->delay = delay;
   pprint->string_space = malloc(LOG_LENTH+1);
   pthread_mutex_init(&(pprint->mut_printed), NULL);
+  pprint->go_on = 1;
 
   int i;
   for( i=0; i<LOG_LENTH; ++i)  
@@ -204,6 +205,15 @@ void free_delay_params (struct delay_params *dly_p){
   free(dly_p);
 }
 
+int check_go_on_print_params(struct print_params *pprint){
+  int ret=0;
+  pthread_mutex_lock(&(pprint->mut_printed));
+  ret=pprint->go_on;
+  pthread_mutex_unlock(&(pprint->mut_printed));
+
+  return ret;
+}
+
 void free_print_params (struct print_params *pprint){
   free(pprint->string_space);
   pthread_mutex_destroy(&(pprint->mut_printed));
@@ -388,15 +398,15 @@ void learn_to_drive(struct RL_agent * rlAgent){
   //struct print_params * pprint = rlAgent->pprint;
   char msg[100];
   
-  pthread_t threadPrint;
-  pthread_create(&threadPrint, NULL, runPrint, (void*)rlAgent);
+  ////pthread_t threadPrint;
+  ////pthread_create(&threadPrint, NULL, runPrint, (void*)rlAgent);
   
  // while(true){
     for(size_t index_episode = 0; index_episode < qlParams->number_episodes; ++index_episode){
       reset(car);
       qlStatus->nb_training_after_updated_weight_in_target = 0;
       qlStatus->index_episode = index_episode;
-      while(true){
+      while(!is_ending(qlStatus) /*true*/){
         ++(qlStatus->nb_episodes);
         ++(qlStatus->nb_training_after_updated_weight_in_target);
         action = select_action(rlAgent);
@@ -435,6 +445,6 @@ void learn_to_drive(struct RL_agent * rlAgent){
 			pthread_mutex_unlock(qlStatus->mut_ending);
 //  }
 
-  pthread_join(threadPrint, NULL);
+  ////pthread_join(threadPrint, NULL);
 }
 

deepQlearn_0/src/deepQlearning/learn_to_drive.h

@@ -25,6 +25,7 @@
 
 //float d_reLU(float x);
 
+extern char *action_name[8];
 
 struct qlearning_params {
   float gamma;
@@ -67,8 +68,11 @@ struct print_params {
   float scale_y;
   struct delay_params *delay;
   char *string_space;//[LOG_LENTH];
+  int go_on;
 };
 
+int check_go_on_print_params(struct print_params *pprint);
+
 struct networks_qlearning {
   config_layers *config;
   neurons_TYPE_FLOAT *main_net;
@@ -136,5 +140,9 @@ void train_qlearning(struct RL_agent * rlAgent,
 //  long reward;
 
 void learn_to_drive(struct RL_agent * rlAgent);
+int is_ending(struct status_qlearning *qlStatus);
+
+///char *fileNameDateScore(char * pre, char* post,size_t score);
+///int select_action(struct RL_agent * rlAgent);
 
 #endif /* __LEARNING_VEHICLE__C_H____ */

deepQlearn_0/src/deepQlearning/vehicle.c

@@ -370,7 +370,8 @@ void goto_xy(int x, int y)
 }
 
 
-static struct winsize w;
+//static 
+struct winsize w;
 
 void init_win(){
   ioctl(1, TIOCGWINSZ, &w);

deepQlearn_0/src/deepQlearning/vehicle.h

@@ -130,4 +130,8 @@ float distance2_coordinate(coordinate *c0, coordinate *c1);
 void print2D_blocks(struct blocks *blk, float scale_x, float scale_y, char pad);
 void print2D_blocks_withPoint(struct blocks *blk, float scale_x, float scale_y, char pad, coordinate *coordPoint);
 
+extern struct winsize w;
+void init_win();
+struct blocks * block_neighbord_Point(coordinate *coord, float *radius );
+
 #endif /* __VEHICLE__C_H__  */