Breaking news: Autonomy in surgical robots

Some people think this is the stuff of fantasy. But this paper has grabbed my attention today, and helps me see a potential roadmap for the Da Vinci 5 and other surgical robotics. The pieces are starting to fall into place.

Autonomous robotic surgery

Many people think there will be sudden leap to fully autonomus robotic surgery. Erm... no.

Instead this is going to happen much like we have seen "assistive" technology creep into our cars.

First we get automatic gear changing...

Then we get brake assistance with ABS...

Then we get lane assistance...

Then we get advanced cruise control...

Then we get corner assistance...

Then we get self parking...

Then we get auto overtaking...

etc etc

We get gradual assistive features that take some of the routine steps out of driving and well... just do it better than most drivers. And that has been on the path to full self driving.

I think we are on that path to this with surgical robots... where even today certain functions are automatically and autonomously helping surgeons. Instruments that straighten themselves to help removal. Self re-orientating at the push of a button. Clash avoidance etc etc. Step by step (very slowly) to autonomy.

A glimpse of the future in autonomous surgery

In this paper : Link here please do go here >> If the insert doesn't work.

We start to get a glimpse into how current AI technology and ML models are being used to "train" robots to mimic surgical manoeuvres (self parking, lane assist) such as picking up needles, throwing a suture tie, lifting tissue.

You should not be looking at what you see today - but the vision of what you see for tomorrow when this is implemented by a giant like Intuitive.

Download the full PDF of their work here >>

This is a fantastic view into what is potentially coming... and way sooner than you think to surgical robots.

Autonomy: Dentistry a first frontier?

I am working closely with an amazing technology of Lupin Dental - that is bringing a transformational technology in dental aesthetics. But to cut a long story short they have coded into software the hand movements of two masters in dental prep. And under the supervision of the dentist, the Lupin robot automatically performs the prep (the hard part) but to the specifications of the experts inside the software. It's started!

Lupin Dental

A little easier than soft tissue because you are dealing with pre-scanned hard structures. But it starts to show that automated steps - especially repeatable steps is really on the horizon.

Da Vinci 5 upgrades to 10,000X compute power.

It does not take a genius to suddenly see why you might want so much compute power in a DV5 ,and why such a massive upgrade vs the Xi. Well for vision based scene assessment and running heavy AI software on the system... you might just need it.

I've said it before, but let me repeat. 80% of the power of the DV5 has not yet been unleashed - and this chassis is just a hardware frame to be upgraded with predominantly software add ons over the next decade until DV6 comes out.

If these research groups out in the wild with older Da Vinci development kits are able to get rudimentary demonstrators working well. Then imagine what the R&D team at Intuitive are knee deep into, and how far along they are. They have access to all the DV video steams, kinematics and hardware upgrades. They also know the vision upgrades that are coming (soon) that will also make this kind of real time vision based autonomy real.

What to expect

The distant future will be that clinicians will be much like pilots. They will plan the route - do some of the more "tricky" parts of the procedure. Be sat in the cockpit supervising, and basically overseeing a lot of autopilot stuff... but being ready to take over in an emergency.

That may even come from a single surgeon supervising 4 or 5 tele-surgical autonomous robots out in different centres via telesurgery. Intervening when needed.

Okay - back to earth.

What I predict is that within five years we will see automatic steps be regulatory approved in the Da Vinci 5. Rudementary steps - but the start down that path of automation towards a degree of autonomy. As we get to the end of the decade and we see a launch of a da Vinci 6 - we will have a significant parts of procedures be carried out with automatic steps: such as needle pick up, suturing and knot tying. Watch the video again - it's not (you see what I did there) too far away. I think we will get assistive dissection (to get better margins), no fly zones for sure... and a few other interesting things I've not fully got my brain around yet.

Look for that 10,000X compute power to start to be unleashed through software upgrades (over the air by the way). Vision upgrades (better information coming to the chip sets) and regulatory clearances.

Automatic steps and autonomy is coming to an operating theatre near you... sooner than you think.

These are simply opinions of the author for educational purposes only.

P.s. congrats to the team in publishing this research.