Something current Large Vision Language Models can’t do: trace cables, the first step for a broad range of robot tasks in homes, factories, construction sites, transportation, and healthcare. Introducing our #CoRL2023 (oral) paper, HANDLOOM (https://t.co/RFKxM91WmE). (1/7)
We applied HANDLOOM to trace a single cable among identical cables, trace cables unseen during training, detect knots, tie knots with unseen distractor cables, and untangle knots unseen during training. (6/7)
When building robotic systems, we create policies to make progress on our task. But what if we encounter states where attempting progress is challenging or risky? In our #ICRA2023 paper, we shape behavior in difficult parts of the state space for the task of untangling cables.
Tangled cables are unwieldy, unsightly, and a tripping hazard in homes and workplaces – including retail, factories, boats, and rock concerts ;). Can robots help? Thread 🧵👇 (1/7)
In our experiments, we observe a 67% success rate when untangling cables with 1 knot and 50% untangling success rate when untangling cables with 2 knots. (6/7)
Could robots help us organize cables in homes, offices, and factories? IRON-MAN is an efficient algorithm for disentangling dense multi-cable knots. @AUTOLab_Cal@ToyotaResearch#IROS2021
(1/7)
Experiments with real cables and a robot in our lab suggest that IRON-MAN can achieve 80.5% success on disentangling dense knots of up to three cables.
(6/7)