Installation and hands-on records and adoption notes from the field. The numbers come from what we installed and measured ourselves.
We used reinforcement learning to teach only one step, approaching the block, to an educational AMR robot arm. Of 11 trainings, four collapsed on reward loopholes such as "standing still," and one that looked good by training score never actually arrived. The final policy arrived in 199 of 200 simulated runs but fell short of the target time.
Read →
To have a robot learn to walk on its own, you first need to install the simulators. This one article covers what Omniverse, Isaac Sim, Isaac Lab and MuJoCo each are, the official download addresses and the versions we used, the official specs and our PC's specs, the install steps and how to check them, and where we actually got stuck.
Read →
We attached Booster K1's reinforcement-learned gait to a museum guide simulation. Reading markers, walking up to each exhibit and giving a talk, three gaits trained from different seeds each ran the course and all reached the exit, with no falls. All of it is simulation, and whether the gait is good enough is judged separately, by the average and worst value over several trainings with different seeds.
Read →
Through V4, the arms only had a penalty for "stay still," and they drifted sideways while walking. V5 adds a reward for swinging the arms back and forth in time with the leg gait, and abrupt motion and torso wobble dropped slightly. We also record the result at 0.7 m/s and what we have not verified.
Read →
There is a method where you do not teach a robot "walk like this" step by step. It tries things and gets scored, and that alone lets it find how to walk on its own. We explain reinforcement learning with an easy analogy, leading into the result of applying it to our K1 robot.
Read →
We moved a Booster K1 walking policy trained in Isaac Lab to MuJoCo. Walking forward and turning carried over, but sideways walking fell. Randomized training (v2) removed the fall but the body turned 195°, and v4, with a gait shape score, cut the heading drift to -11.6°. This gives the method and the results as they came out.
Read →
Installing Isaac Sim 5.1.0 and Isaac Lab 2.3.2 on Windows with the official commands stopped us in five places. This article gives the pkg_resources error, an access violation right after training starts, and the missing URDF extension 2.4.31, with the error text and fix commands as they appeared.
Read →
Yes. On one PC with a single GPU, Booster K1 walking trained in 32 minutes into a gait that does not fall, and v4, with the gait shape corrected, trained in 1 hour 40 minutes into a human-like gait. This covers the PC specs, the tools, the cross-check in another simulator, and the limits.
Read →
One GPU ran 4,096 K1 robots at once and produced a gait that does not fall in 32 minutes. But it was a shuffle of 15.2 foot steps per second. Retrained with a gait shape score, v4 walked at 3.6 steps per second. We also record where the install got stuck and the limits we have not yet checked.
Read →