Final Proximity Space Systems

Let’s Talk Space

02Chronology of rendezvous and docking

Two spacecraft,
one Meeting.

Sixty years of finding each other in orbit, from a crew flying by eye to autonomous approach against a target that never agreed to be approached.

T minus 00:10:00   THE LINE

How the problem kept changing

Rendezvous has been solved several times over, because the problem keeps being replaced. Each generation inherited the previous answer and then removed one of its assumptions: the crew, then the cooperation of the target, then the ground link, then daylight, and finally the docking port itself.

What follows is the full sequence rather than the highlights, because the pattern only shows up at length. Roughly forty missions, grouped into six eras.

37operations listed
31success
4partial
2failed
84%fully successful

T minus 00:09:00   THE PIONEERING ERA

The pioneering era

Everything in this period was a first, and most of it was flown by hand. The question was not how to do it well but whether it could be done at all.

Dec 1965

first orbital rendezvous

Gemini 6A and 7success

Schirra and Stafford flew Gemini 6A to find Gemini 7, already in orbit with Borman and Lovell aboard. They closed to about one foot and held station for more than five hours. Neither had a docking port, so the two never touched. Rendezvous was the hard part and it was solved here.

Closing on a target in orbit is counter-intuitive: thrusting towards it raises your orbit and makes the gap grow. Gemini 4 had tried exactly that and watched its target drift away.

NSSDC, Gemini 6A

Gemini 7 seen from Gemini 6A during the first rendezvous. NASA, S65-63171.

Mar 1966

first docking

Gemini 8partial

Armstrong and Scott joined two spacecraft in orbit for the first time. Twenty-seven minutes later a stuck thruster on their own vehicle set the docked pair rolling, and undocking made it worse because the Agena mass had been damping it. The full account is here.

A docked pair is one vehicle, with mass properties neither half had alone. Separating from an unlocalised fault was the intuitive move and it was very nearly fatal.

NASA, Gemini VIII

The Agena Target Docking Vehicle from Gemini VIII, minutes before the first docking. NASA, S66-25782.

Jun 1966

the target that could not be docked

Gemini 9Apartial

Stafford and Cernan reached their target to find its launch shroud still attached, jaws half open. Stafford called it an angry alligator. They flew three separate rendezvous profiles instead, which turned a failed mission into the best rendezvous training of the programme.

The target with its shroud still attached, jaws half open. The angry alligator. NASA, s66-37923.

Jul 1966

the first dead target

Gemini 10success

Young and Collins docked with their own Agena, used its engine to climb, then rendezvoused with the derelict Agena left behind by Gemini 8. No transponder, no lights, no attitude control. The first approach to something that could not help.

Sep 1966

docking on the first orbit

Gemini 11success

Conrad and Gordon docked 94 minutes after launch, a direct ascent rendezvous with no phasing orbits at all. They then used the Agena engine to reach 1,369 km, still the altitude record for a crewed spacecraft in Earth orbit.

The Agena at the far end of a 100-foot tether during Gemini 11, an early experiment in keeping two vehicles station without burning propellant. NASA, S66-54571.

Nov 1966

rendezvous by hand

Gemini 12success

The rendezvous radar failed. Aldrin computed the closing manoeuvres from a sextant and a chart, and they docked anyway. His five hours of EVA, rehearsed underwater beforehand, finally made spacewalking a controllable task.

Aldrin outside Gemini 12, the EVA rehearsed underwater beforehand. NASA, s66-63537.

Oct 1967

first automatic docking

Kosmos 186 and 188success

The Soviet Union docked two uncrewed Soyuz-derived vehicles with nobody aboard either, using the Igla radio system. This is the branch point between two philosophies. The Soviet programme treated automatic docking as the default with manual as backup, and the American programme did the reverse for decades.

The branch point between two philosophies: the Soviet programme made automatic docking the default with manual backup, the American programme the reverse for decades. Igla became Kurs, which is still flying.

NSSDC, Kosmos 186

Jan 1969

first crew transfer

Soyuz 4 and 5success

Two crewed spacecraft docked, and two cosmonauts moved between them by spacewalking across the outside. There was no internal transfer tunnel. It established that a docked pair could exchange people, which is what a space station needs.

Mar 1969

the lunar module rehearsed

Apollo 9success

Rendezvous and docking of the command and lunar modules in Earth orbit, including a separation to 180 km and a return. The first flight of a spacecraft that could not re-enter, so the rendezvous was not optional.

Spider in lunar landing configuration, photographed from Gumdrop on the fifth day of Apollo 9. NASA, AS09-21-3212.

May 1969

the full profile short of landing

Apollo 10success

The lunar module descended to 15 km above the Moon and then rendezvoused with the command module in lunar orbit. Everything except the landing, at the real distance and with the real light delay.

Snoopy’s ascent stage climbing back from 15 km above the Moon, photographed from the command module before docking. NASA, AS10-34-5112.

Jul 1969

the rendezvous nobody talks about

Apollo 11success

The landing is the famous part. The engineering achievement was the ascent stage finding the command module in lunar orbit afterwards, with one engine, no backup, and no possibility of rescue if it missed.

The whole architecture rested on lunar orbit rendezvous working: it is why the lander could be left behind and the vehicle that went did not have to be the vehicle that came home.

NSSDC, Apollo 11

The Apollo 11 ascent stage climbing back towards the command module, lunar horizon and a half Earth behind it. NASA, as11-44-6642.

T minus 00:07:30   STATIONS, AND THE FIRST ROUTINE DOCKING

Stations, and the first routine docking

Once there was somewhere to go, docking stopped being an achievement and became a scheduled operation. Resupply is what forced it to become reliable.

Apr 1971

Salyut 1

Soyuz 10 and 11partial

The first space station. Soyuz 10 docked but could not open the hatch and returned. Soyuz 11 got aboard and stayed 23 days, then the crew died during re-entry from a valve that opened early. Docking had become routine faster than everything around it.

May 1973

docking to a damaged station

Skylabsuccess

Skylab reached orbit with a solar array torn off and a heat shield missing. The first crew flew a fly-around inspection before docking, then repaired it from outside. Approach and inspection as a diagnostic tool, not just a step towards docking.

Skylab from the approaching crew vehicle, one solar array wing plainly missing, the gold parasol sunshade deployed. NASA, sl3-114-1683.

Jul 1975

first international docking

Apollo and Soyuzsuccess

Two spacecraft with incompatible hatches, incompatible atmospheres and incompatible everything else, joined by an androgynous adapter built for the purpose. The hard problem was agreeing an interface, not flying the approach.

Androgynous means neither side is probe nor drogue, so either vehicle can be the active one. The principle runs through APAS into the standard the ISS uses now.

NASA History, ASTP

Soyuz photographed from Apollo, July 1975. The two vehicles are not shown joined. NASA, ast-01-056.

Jan 1978

first uncrewed resupply

Progress 1success

A stripped-down Soyuz with no heat shield and no seats, carrying propellant and cargo to Salyut 6, docking automatically and later burning up on purpose. It made long-duration occupation affordable and is still flying today.

Progress 77 on approach to the ISS. NASA, iss064e033797.

1986 onward

assembled by docking

Mirsuccess

A station built module by module, each one arriving under its own power and docking itself, some using a manipulator arm to relocate to a side port afterwards. Assembly in orbit became an engineering discipline rather than a demonstration.

Mir a decade into assembly, seen from Atlantis during the STS-76 departure in March 1996. Every module in view arrived and docked under its own power. NASA, STS076-713-083.

T minus 00:06:00   THE SHUTTLE ERA

The Shuttle era

A vehicle with an arm and a payload bay changed the question from whether two spacecraft could meet to what could usefully be done once they had.

Apr 1984

first satellite repair

Solar Maxsuccess

A crew captured a failed satellite, replaced its attitude control module in the payload bay and released it working. The first capture attempt failed and left it tumbling, which cost a day of stabilising. A failed capture can make a target harder to capture.

Nelson approaching Solar Max in a manned manoeuvring unit. This first attempt failed and left it tumbling. NASA, 41c-34-1380.

Nov 1984

first satellite recovery

Palapa B2 and Westar 6success

Two communications satellites stranded in useless orbits were captured by astronauts flying manned manoeuvring units, wrestled into the payload bay and brought home. Still the only time satellites have been retrieved to Earth and later reflown.

Gardner rides the manned manoeuvring unit to a hard dock with the slowly spinning Westar 6, using the stinger probe built for the capture. NASA, 51A-104-029.

Dec 1993

Hubble servicing mission 1success

The most demanding proximity operation attempted to that point: capture a 11-tonne telescope, hold it in the bay through five spacewalks, correct its optics and release it. Four more servicing missions followed to 2009.

Solar Max proved a satellite could be caught and repaired; Hubble proved it could be planned for, with handholds, standard fasteners and modular boxes designed in.

NSSDC, STS-61

Hubble berthed in the payload bay on STS-61, 1993, the first servicing mission. NASA, STS061-79-087.

Jun 1995

unequal masses

Shuttle and Mirsuccess

A 100-tonne orbiter docking with a 120-tonne station, nine times over three years. Contact dynamics between two large flexible structures became a real design driver, and the approach corridor was constrained as much by where the thrusters could safely point as by where the port was.

Contact dynamics stop being a detail at this mass: the docking system had to absorb the meeting of two large flexible structures, and the corridor was set by where the plumes could point.

NASA SP-4225, Shuttle-Mir

Atlantis seen from a window on Mir while the two were docked, STS-71. NASA, sts071-741-004.

Dec 1998 onward

ISS assemblysuccess

More than thirty flights across a decade, from five agencies, docking or berthing hardware they had not built to hardware they had not built. Interoperability stopped being a courtesy and became a requirement.

The ISS under construction, February 2001, photographed from Atlantis after separation. NASA, sts098-713a-004.

T minus 00:04:30   AUTONOMY, AND WHAT IT COSTS TO GET WRONG

Autonomy, and what it costs to get wrong

Removing the crew from the loop is the step that makes servicing and debris removal economic. It is also the step where the failures start.

Nov 1997

first fully autonomous rendezvous and docking

ETS-VIIsuccess

Japan flew two uncrewed spacecraft that rendezvoused, docked, separated and docked again with nobody aboard and nobody flying it. It went further and deliberately released its target, let it drift, then captured it again with a robotic arm. It also proved the loop had to close on board, because the round trip to the ground was seconds long.

Apr 2005

an estimator that was confidently wrong

DARTfailed

An autonomous rendezvous demonstrator collided with the satellite it came to inspect. Its navigation had drifted and its own estimate of closing velocity was badly wrong, with nothing independent to check it against. It retired early on depleted propellant. An estimator without a cross-check is a single point of failure that reports success.

2007

servicing demonstrated end to end

Orbital Expresssuccess

Two purpose-built spacecraft autonomously docked, transferred hydrazine, swapped a battery and a flight computer by robotic arm, and separated. It demonstrated the whole servicing chain and then, notably, was not followed up for over a decade.

T minus 00:03:00   CARGO, AND THE SPLIT BETWEEN DOCKING AND BERTHING

Cargo, and the split between docking and berthing

Once resupply became competitive, two different approaches appeared, and the difference between them is not cosmetic.

1978 onward

and Kurs

Progresssuccess

The Igla system gave way to Kurs in the 1980s, which is still the workhorse. Automatic approach with a manual TORU backup flown from inside the station, a fallback that has been needed more than once.

2008 to 2014

automatic docking to the ISS

ATVsuccess

Europe flew five Automated Transfer Vehicles, navigating the final approach with videometers tracking retro-reflectors on the Russian segment. Before it was allowed near the station it had to demonstrate a collision avoidance manoeuvre and a full escape sequence, which made passive safety a licence condition rather than a design preference.

The demonstration requirement is the interesting part: ATV had to prove it could abort safely before it was permitted to approach, which makes passive safety a licence condition rather than a design preference.

ESA, ATV

ATV Jules Verne during the Demo Day 2 rehearsal, 31 March 2008, closing to about 36 feet of Zvezda. The docking followed on 3 April and was not photographed from the station. NASA, iss016e034176.

2009 onward

berthing instead of docking

HTV Kounotorisuccess

Japan chose a different answer: fly to a hold point about ten metres away, stop, and let the station arm capture and berth it. Slower, and it needs the arm and a crew, but the vehicle never has to fly itself into contact.

Kounotori 2 in the grapple of Canadarm2 at the Harmony nadir port. The vehicle flew to a hold point and the arm did the rest. NASA, iss026e024076.

2013 onward

Cygnussuccess

A commercial freighter that also berths. Its later flights have doubled as a reboost stage for the station, which turns a cargo vehicle into part of the station propulsion system.

Cygnus held by Canadarm2: berthing, not docking. NASA, iss064e036984.

May 2012

first commercial vehicle to the ISS

Dragon C2+success

A privately built spacecraft reached the station, held station while its systems were checked, and was captured by the arm and berthed. Berthed, not docked: the arm did the final positioning.

Berthing and docking are different operations with different failure modes: berthing puts an arm and a human in the loop for the last metres and tolerates a slower approach.

NSSDC, Dragon C2+

T minus 00:01:30   CREW, COMMERCE, AND UNCOOPERATIVE TARGETS

Crew, commerce, and uncooperative targets

The current era. Docking is now something several companies do on contract, and the interesting targets are the ones that were never meant to be approached at all.

Mar 2019

first autonomous commercial docking

Crew Dragon Demo-1success

A commercial crew vehicle flew itself to the ISS and docked with no crew aboard and no arm involved, using the International Docking Adapter. The first use of the IDSS standard interface in flight.

Crew Dragon closing on the station during Demo-1, 3 March 2019, with nobody aboard and no arm waiting. NASA, iss058e027349.

Dec 2019

a rendezvous that never happened

Starliner OFTfailed

A mission elapsed time error put the spacecraft into the wrong burn at the wrong moment, burning the propellant reserved for the rendezvous. It never reached the station. A timing fault, not a guidance fault, which is a useful reminder of where the failure modes actually live.

May 2020

Crew Dragon Demo-2success

The same docking with two astronauts aboard, and a manual mode demonstrated on the way in. Autonomous by default with a human fallback, which is the Soviet 1967 answer arrived at from the opposite direction.

Autonomous by default with a human fallback is the answer the Soviet programme reached in 1967, arrived at from the opposite direction fifty years later.

NSSDC, Crew Dragon Demo-2

Crew Dragon on final approach during Demo-2, nose cone open, docking ring exposed. NASA, iss063e021563.

Feb 2020

the uncooperative era begins

MEV-1 and Intelsat 901success

A servicing vehicle docked with a commercial satellite in graveyard orbit that had no docking port, no markers, no transponder and no way to assist. MEV-1 used the launch adapter ring and the apogee engine nozzle as capture features, took over station-keeping, and returned it to service for five years.

The population of satellites worth approaching is overwhelmingly things designed with no thought of ever being approached, drifting, in bad light, on a long ground loop.

Northrop Grumman, SpaceLogistics

Apr 2021

docking to a satellite still working

MEV-2success

The second vehicle docked with Intelsat 10-02 while it remained in service, without interrupting it. Servicing moved from rescuing dead satellites to extending live ones.

May 2022

Starliner OFT-2success

The uncrewed flight test repeated, successfully docking to the ISS after the first attempt failed to reach it.

Starliner approaching the Harmony forward port on OFT-2, 20 May 2022, docking ring exposed. NASA, iss067e066717.

2021

capture by magnet

ELSA-dpartial

A demonstration of repeated capture and release of a client fitted with a ferromagnetic plate, including a capture of a deliberately tumbling target. Built for the debris removal case, where the target has to be caught rather than docked with.

2021 onward

Shenzhou and Tianzhou at Tiangongsuccess

China now flies crewed and cargo docking to its own station routinely, with fast automated rendezvous profiles reaching the station in a few hours. A third independent line of development, arrived at separately.

A Soyuz spacecraft approaching the International Space Station
Soyuz MS-29 approaching, arrays deployed and docking probe extended, July 2026. Nearly sixty years after Kosmos 186, the same basic approach is still flying crews. NASA, iss074e0814091.

T minus 00:00:50   ANNOUNCED

What is coming: servicing, refuelling, removal

The entries below are announced or in flight but not yet concluded, so they carry no outcome mark. What they share is the shape of the market that MEV-1 opened: docking as a service, sold by the operation rather than flown by a state.

2024 onward

approaching real debris

ADRAS-J and ADRAS-J2

Astroscale’s inspector flew within metres of a derelict Japanese upper stage in 2024, the first commercial close approach to a genuine piece of large debris, uncontrolled and uncooperative. The follow-on mission is contracted to capture and deorbit it with a robotic arm.

2024, cancelled

the cautionary entry

OSAM-1

NASA’s mission to robotically refuel Landsat 7, a satellite never designed for it, was cancelled in 2024 after years of cost growth. The lesson cuts the other way from the successes: servicing an unprepared client is hard enough to sink a flagship project.

2025 onward

Mission Robotic Vehicle and Mission Extension Pods

The successor to MEV changes the economics: rather than one large vehicle staying docked for years, a robotic servicer installs small propulsion pods on client satellites and moves on. One vehicle, many clients, and a robotic arm doing the attachment.

Announced

multi-client debris removal

ELSA-M

The production version of the ELSA-d magnetic capture demonstration: one servicer removing several prepared satellites per mission, aimed at constellation operators whose spacecraft carry a docking plate from the factory.

Announced

capturing with arms

ClearSpace-1

ESA’s commissioned debris removal flight, capturing an uncooperative object by enclosing it with four robotic arms rather than docking with any feature. Embracing a target that offers nothing to grab is the hardest version of the capture problem.

Announced

Orbit Fab and refuelling interfaces

Fuel depots and tankers built around a standardised refuelling port that satellites fit on the ground. The bet is the opposite of MEV’s: rather than docking with anything, make the next generation dockable by design. Refuelling turns delta-v from a fixed budget into a consumable.

In flight

Starship orbital propellant transfer

Lunar landings under Artemis depend on transferring hundreds of tonnes of cryogenic propellant between Starships in orbit, which requires docking between two vehicles larger than anything ever joined in space. The rendezvous is the enabling step for the whole architecture.

Late 2020s

assembly around the Moon

Gateway

The lunar station repeats the ISS assembly story at a distance where abort to Earth takes days: modules and vehicles docking autonomously in an orbit no crew can quickly reach. Every capability in this chronology, exercised far from help.

T minus 00:00:40   WHAT IT ADDS UP TO

What was removed, and what is left

Removed, one era at a time

  • The crew, from Kosmos 186 in 1967
  • The pilot on the vehicle, with Progress and later ATV
  • The ground loop, with ETS-VII and ATV
  • The cooperative target, with Gemini 10 by hand and MEV-1 autonomously
  • The docking port, using launch adapter rings
  • Daylight, with lidar and thermal sensing

Still open

  • A target that tumbles rather than holding attitude
  • A target that manoeuvres, deliberately or not
  • Pose estimation from a single camera at close range
  • Plume impingement on a structure you are about to touch
  • Proving an approach is safe before it is flown, rather than observing that it was

Sixty years in, the vehicles have changed completely and the hard part has not moved. It is still the last hundred metres, and it is still about whether the constraints hold when the world does not match the model.

That last item is what we work on. Our capabilities are here.

T minus 00:00:20   SOURCES

Where this comes from

Fehse, Wigbert. Automated Rendezvous and Docking of Spacecraft. Cambridge: Cambridge University Press, 2003.

Goodman, John L. “History of Space Shuttle Rendezvous and Proximity Operations.” Journal of Spacecraft and Rockets 43, no. 5 (2006): 944–959.

JAXA , Engineering Test Satellite VII, KIKU-7 (ETS-VII)

ESA , Automated Transfer Vehicle

NASA , 40 Years Ago: STS-41C, the Solar Max Repair Mission

International Docking System Standard (IDSS), Interface Definition Document

Northrop Grumman SpaceLogistics, Mission Extension Vehicle, on the MEV-1 docking with Intelsat 901 in February 2020.

How the outcomes are marked

success
The mission achieved its primary objective. The spacecraft arrived, the instruments worked, and the data or the crew came home. Judged against the mission’s own stated objective, not against later ambitions: Voyager 2 at Neptune, Chang’e 5 returning its sample, Apollo 11 despite landing long of its target.
partial
The mission reached its target and returned something real, but not everything it was built for: an instrument that failed after arrival, a lander that tipped, an antenna that never opened. Galileo delivered eight years of science through a jammed antenna. SLIM landed on its nose with its precision objective met. Philae bounced twice and still worked from shadow.
failed
The mission did not achieve its primary objective: it never arrived, it crashed, or it arrived unable to work. DART colliding with the satellite it came to inspect, Luna 25 into the surface, IM-2 on its side in shadow with no power and none of its science.

What is counted: every rendezvous, docking or capture operation listed, judged on whether the operation itself achieved its objective. Announced and in-flight missions in the final section carry no mark, because a mission that has not flown has no outcome. Gemini 8 is partial: the docking succeeded and the mission was then aborted. Gemini 9A is partial: the rendezvous worked perfectly and the target could not be docked with.

All nine topics
PreviousHistorical overview of space flight NextEvolution of interplanetary missions