MIT’s VLASH System Lets Robots Plan While They Move

MIT researchers built a planning system called VLASH that lets a robot start figuring out its next move before it finishes the current one. Most robot control systems don't work that way.

They finish a motion, stop, replan, then move again. Those pauses are why robots look jerky, and they've been holding back both speed and fluidity for a long time.

What Happened

VLASH eliminates most of that dead time by predicting where the robot will be a moment from now and planning the next action based on that prediction, running in parallel with the current one. It doesn't add computing overhead, and it doesn't.

• If you are seeking financial advice, find a professional who is right for you.

• Disclaimer: This article is written purely for entertainment and educational purposes and should not be taken as financial advice in any way.

• Consider leaving a tip to support my writing efforts.

Key Details

The results: pick-and-place tasks ran about twice as fast in testing. Table tennis and Whack-a-Mole, tasks that need constant motion, showed even bigger gains.

• That's probably why VLASH-style planning is likely to show up in warehouses, factories, and service robots that need both speed and motion that doesn't feel abrupt around people.

• The motion starts to look less mechanical and more like how a living thing moves.

• It closes the gap between how fast a robot can think and how smoothly it can actually move.

Why It Matters

There's also a companion technique called action quantization, which trades a bit of positional accuracy for speed, roughly 1.5 to 2 times faster task completion. VLASH works with existing robot platforms too, so none of this requires new hardware.

• Cutting out the forced pause between acting and replanning fixes a problem that's been slowing robots down for years.

• The more aggressive quantization settings still cost some accuracy, and messier real-world environments, ones with unpredictable obstacles or a lot of contact, still need more testing.

What Reports Say

Coverage of the story so far points to:

• Continued reporting by HackerNoon as more details emerge