[Question]The state I get not match with my action

[cainiaodidi]

example code as below

def func()
  action = agent.act(state)
  next_state, reward, done, truncated, info = env.step(action)

I found that the observation of actuator in next_state is not equal to which in action but equal it two steps later, for example: t : action = 1 observation = 0 t+1: action = 2 observation = 0 t+3: action = 3 observation = 1 t+4 :action = 4 observation = 2

[AlejandroCN7]

Hi @cainiaodidi, Duplicate, the answer can be found in the issue #404