@@jonathanr4242 Hello, I attempted to implement this on the Franka robot, but after training, when I try to simulate, my target actions are always the same, like this: target [0, 0, 0, 0, 0, 1, 0, 0]. Could you please help me?
@@SakethBajjuri hi. Yes. I did a wheeled robot but got caught up in another project. I’m making a cheaper version of mobile aloha at the moment. But the focus is on getting the actual robot working rather than any AI. I should release a video in a couple of weeks.
I need your suggestion on onething. I have a industrial cobot like UR robot. I would like to jog the robot manually since i dont have the follower arm like the one which u have . I can get the robot current state by topic joint state just like the current and i can move the robot just by using jogging mode. But my question here is wont be the observation and action will has same joint state..? Do i need to take action as t+ 1 and t as observation or is it model itself does that here is the reference what do i meant by jogging means ua-cam.com/video/GqNUUmlj_sk/v-deo.htmlsi=BNLygCeyTYFsz1dy&t=31
@@SakethBajjuri yes, as far as I remember the observations are 1,…,t and the action is t+1. Ideally, I would expect the observations and actions to be able to be read from the same joint state.
Thanks for sharing your work.. really helpful for beginners to start
This is amazing, thank you for the share!
@@mikementele thank you, for your kind words.
Amazing work
Thank you. That is very kind of you.
@@jonathanr4242 Hello, I attempted to implement this on the Franka robot, but after training, when I try to simulate, my target actions are always the same, like this: target [0, 0, 0, 0, 0, 1, 0, 0]. Could you please help me?
Amazing robot!❤ Awesome work friend!🤩💥💯😃👍
Thank you, for your kind words.
Liked it !
thank you.
did u tried with the wheeled robot
@@SakethBajjuri hi. Yes. I did a wheeled robot but got caught up in another project. I’m making a cheaper version of mobile aloha at the moment. But the focus is on getting the actual robot working rather than any AI. I should release a video in a couple of weeks.
I need your suggestion on onething. I have a industrial cobot like UR robot. I would like to jog the robot manually since i dont have the follower arm like the one which u have . I can get the robot current state by topic joint state just like the current and i can move the robot just by using jogging mode. But my question here is wont be the observation and action will has same joint state..? Do i need to take action as t+ 1 and t as observation or is it model itself does that
here is the reference what do i meant by jogging means
ua-cam.com/video/GqNUUmlj_sk/v-deo.htmlsi=BNLygCeyTYFsz1dy&t=31
@@SakethBajjuri yes, as far as I remember the observations are 1,…,t and the action is t+1. Ideally, I would expect the observations and actions to be able to be read from the same joint state.