Watch how Bimo learns in 30 minutes, equivalent to 50 million steps, how to handle a forward push using reinforcement learning inside Isaac Sim.
Even though the end result seems simple and intuitive, it actually took months of work to refine and create a simulation pipeline that works as intended.
Te most exciting part? This pipeline is very flexible, and with small parameter changes it can be used to teach new behaviors, and spoiler alert, Bimo can learn how to walk now too! Stay tuned for the next video.
#TheBimoProject #Robotics #Innovation #Startup
Investment
Collaboration
Partnership