How does automated docking work and why isn't it used more often?

Question

It seems on every flight to the ISS for either people or just cargo, there is always a big deal over manual docking, lots of back and forth over comms, pinpointing the right spot, slowing down in a long procedure, all controlled by hand.

But I keep hearing about automated docking. What are the technologies used for such a thing currently? What are technologies that could be used soon or down the road? Is there some reason automated docking is not used more often?

I've seen info on Kurs, what was used on Mir, but it seems it is not used anymore.

And further but not necessary, are there any proposed machine learning methods proposed working on visual or other electromagnetic-based sensors?

Answer 1

As noted in my answer to What is the difference between docking and berthing on the ISS?, the primary difference between docking and berthing is whether the chaser vehicle (the vehicle docking with the target vehicle, in this case, the ISS) performs all of the translational and rotational maneuvers right up to contact, or relinquishes control to some other system (in the case of the ISS, the robotic arm) at some point, with that other system performing the final maneuvers that bring the chaser into contact with the target.

The question asks about automated docking. The Japanese HTV, SpaceX's Dragon, and Orbital's Cygnus perform automated berthing, up to the point where those vehicles enter the berthing box and relinquish control to the ISS. The final act of berthing is performed by the ISS robotic arm, which requires a human operator.

Rendezvous and proximity operations, including automated rendezvous, can be divided into three phases: far field, near field, and approach. In the far field, the chaser roughly knows where both it and the target vehicle are, but has yet to detect the target using far field sensors such as radar.

Once the chaser is in reliable communication range with a cooperative target, the target can work with the chaser to improve the odds of a successful operation. Rendezvous with an non-cooperative target is a difficult problem. The ISS is a cooperative target. It can and does provide active aids to rendezvous.

One of the key near field sensors formerly used by Shuttle and the ATV, and currently used by the HTV, Dragon, and Cygnus is relative GPS. This entails the target signaling via radio information on each of the GPS satellites seen by the target. The chaser selects GPS satellites common to both to obtain provides highly accurate relative position data. Another use of the radio is having the target embed a frequency signal generated by the chaser into the signal returned to the chaser. This provides highly accurate range rate measurements, much better than radar with a non-cooperative target.

Relative GPS becomes problematic once the chaser gets close to the target. The ISS is large and either blocks the view to GPS satellites or creates a lot of reflections, aka multi-path (multi-path is the key reason GPS doesn't work so well in the downtown areas of large cities). This is the time to switch to a different set of approach sensors. These include passive sensors such as cameras and active sensors such as LIDAR. Cooperative targets such as the ISS can aid here as well. The ISS has retroreflectors and visible markers placed at key locations that enable an automated chaser to fully determine relative state (position, velocity, orientation, and rotation rate).

Answer 2

First a general image of what the port situation at the ISS kind of looks like. This is a bit dated, but one of the best images, showing some of the possible vehicles docked.

To note the ATV at the left end, the Soyuz/Progress up/down are all using Russian segment ports.

The HTV on the right, next to the Shuttle is where a PMA now resides with an IDA. The Shuttle port has had an IDA added to it.

On to the question, right now, only the Russian Soyuz and Progress vehicles dock automatically, and they continue to use the updated Kurs. There is a manual backup mode that is seemingly being used more often than would be liked of late. Progress, no crew on board can be remotely piloted/docked from the ISS itself.

Thus the goal is pretty much always dock as automated as possible.

Dragon, Cygnus, and HTV (JAXA) all berth, where the fly to a holding point and the CanadArm2 grabs and completes the berthing. This holding point is an automated process, but a human flies the arm to grab the vehicle and then berth it.

A core difference between berthing and docking is that the actual connection of vehicle to station for docking can be done automated, such that a crew abandoning an empty station, can close the door behind them and leave. A berthed vehicle needs someone on the other side to fully close the door and thus someone needs to stay on the station. So even if you automated the berthing, you still need someone to open or close the door.

The future Dragon Crew (Dragon V2, whatever it officially ends up being called) and Boeing CST-100 will be using a US automated docking system when they dock to the PMA ports (with IDA's attached). (As I write this, Nov 2018 flight of Dragon V2 first expected. We shall see).

The Shuttle (retired) used the PMA's as well, but without the IDA adapters that were recently added (Brought up on Dragon Cargo missions).

The ESA ATV (retired) used Kurs to dock to the Russian segment ports.