Mission Extension Vehicle 2

Overview

A geostationary life-extension vehicle has to dock with a satellite that was never built to be docked with, using whatever interface the client happens to carry, in a belt crowded with active traffic and with no crew able to inspect a failed approach. Industry’s economic argument for building such a vehicle rests on the observation that GEO comsats fail well short of design life often enough to make a fleet of clients worth visiting [1]. MEV-2 is the second Mission Extension Vehicle built to that argument. It shares its docking mechanism, rendezvous and proximity operations (RPO) sensor suite and client interface with MEV-1, and differs from it in which satellite it was sent to and the orbital regime that satellite occupied.

Design space

Servicing designs differ along four axes: how the servicer approaches (guidance and navigation), how it senses an uncooperative client, how it captures it, and what it does once attached. MEV chose the narrowest answer on the last axis, station-keeping and pointing only, and spent its design effort on the first three.

Approach guidance. Every rendezvous guidance law in this lineage derives from the linearized relative-motion equations of a chaser near a circular-orbit target, first written down for terminal rendezvous guidance in 1960 [16]. Closed-form multipulse glideslope and flyaround schemes built on them were developed for autonomous approach and departure in circular orbit [17]. Shuttle crews flew proximity operations manually for thirty years, and the flight dynamics history of that program attributes the lack of full automation to the absence of a programmatic requirement and the cost of certifying it, not to sensor limits alone [23]. Optimization-based planners for tumbling clients, using variable planning horizons and linear programs, have been demonstrated only in simulation against a defunct Envisat model [26]. MEV’s clients were station-kept comsats, so tumbling-target guidance was not required of it. Ground rehearsal of the final approach used hardware-in-the-loop facilities such as DLR’s EPOS, which validated a delay-compensated closed-loop rendezvous filter down to 1.8 m of standoff before any commercial GEO servicer flew [7].

Sensing an uncooperative client. The flown precedents split by whether the client carried aids. ETS-VII (flown 1998 to 1999) docked autonomously with a cooperative target carrying LED markers and a corner cube reflector, reaching 7 mm lateral and 0.6 mm/s residual rate at capture against requirements of 25 mm and 5 mm/s, with GPS relative navigation good to about 10 m [18]. MEV’s client carries no retroreflectors, so its rendezvous suite combines narrow- and wide-field visible imagers, long-wave infrared imagers and a scanning lidar, three dissimilar phenomenologies feeding one relative-navigation solution [3]. The lidar was modified from its ISS-approach heritage with an added amplification stage and reduced beam divergence for this reason [3].

Capture. The interface MEV uses, a tool driven into the client’s liquid apogee engine nozzle inside its launch adapter ring, was proposed for the same market by SMART-OLEV more than a decade earlier, on the argument that roughly 80 percent of GEO communications satellites carry that nozzle-and-ring geometry as a leftover of how they reached orbit [2]. Other flown or ground-tested capture schemes differ in what they assume of the client. Orbital Express used a three-arm soft capture mechanism with a 15 cm axial, plus or minus 5 cm lateral and 3 cm/s contact envelope, qualified in hardware-in-the-loop contact tests against a client designed to be captured [20], [22]. DLR’s ROKVISS operated a two-joint torque-controlled manipulator on the ISS from 2005, and the same overview outlines the TECSAS free-flyer capture mission and Orbital Recovery’s CX-OLEV GEO life-extension tug [21]. A launch-adapter-ring capture tool with two fingers has been built as a 5.85 kg prototype and bench-tested at 1 g, with no microgravity or thermal-vacuum testing [27]. A standardized docking port with a lance and capture cone, tested on parabolic flights to derive a sub-1-centimeter-per-second approach velocity, was worked out at small scale on SPHERES and never flown as a full docking on orbit [5]; the same testbed was used to study control reconfiguration after docking, where a docked payload mattered only when its mass was comparable to the satellite’s own [25]. The 2020 survey of space manipulators for debris mitigation places MEV beside ETS-VII, Orbital Express and SMART-OLEV as the short list of flown or near-flown precedents [6].

Servicing beyond life extension. NASA’s parallel effort, RESTORE-L turned OSAM-1, took the opposite path: a government-run refueling and dexterous-servicing demonstrator, canceled in 2024 after eight years and roughly 1.4 billion dollars against an original 750 million dollar estimate, never having flown [4]. Its propellant transfer subsystem was nonetheless qualified on the ground: a hose management assembly, a seal-less hypergolic pump and a turbine flow meter in the 2018 status [28], and hose cycling plus leak detection to 5x10^-3 sccs of helium in the FY2024 account [29]. Refueling hardware for the Robotic Refueling Mission 3 cryogen demonstration was loaded with about 50 liters of liquid methane under manual ground procedures [30]. A broader review places these within on-orbit servicing robotics as of 2014, covering manipulator kinematics, target motion estimation and capture stabilization [24].

What changed, and what it enabled

MEV-1 docked to Intelsat 901, a satellite already retired to the graveyard orbit roughly 300 km above the geostationary belt [3], so the operation had no active neighbors to avoid. MEV-2 docked to Intelsat IS-10-02, a satellite still carrying revenue traffic in its assigned slot, which removed the option of clearing the surrounding airspace before the approach and made MEV-2 the first vehicle to dock with an operating satellite in an active GEO slot [8]. Once docked, MEV-2 assumed the client’s station-keeping and pointing, letting Intelsat disable the satellite’s own propulsion and attitude control for the service period [9].

The flight imagery gave the first measured detection ranges for an uncooperative GEO client: visible cameras detected and tracked it beyond 30 km at sub-pixel resolution, long-wave infrared beyond 10 km, and the lidar held 3-DOF position beyond 2 km and a 6-DOF solution with model overlay from 60 m down to 2 m [3]. The same camera suite is idle for most of the vehicle’s life between roughly four-week installation windows that recur four or five times a year. A later proposal argued for reusing servicer cameras as a part-time space domain awareness sensor viewing the belt from an in-track direction ground radars cannot reach; its detection ranges of about 600 km for 1 square meter and 1800 km for 3 square meters are model outputs, not measurements [9].

The taxonomy that later NASA surveys use to place MEV-2, robotic manipulation; rendezvous, capture, docking and mating; relocation; repair and maintenance; refueling; and structural assembly, was set by the first of those surveys and has been carried forward largely unchanged [10]. MEV’s published thruster performance has become a stand-in figure for fleet-level logistics studies: a 2021 optimization framework for trading high- and low-thrust servicer architectures took its XR-5 Hall thruster figures from Northrop Grumman’s public MEV data [11], and a depot-placement study argued that a shared servicing depot could cut the combined logistics mass of servicing two large constellations by 15 to 21 percent relative to separate fleets [12].

Mission profile

Parameter	Value	Source
Launch	15 August 2020, rideshare with Galaxy 30	[13]
Client	Intelsat IS-10-02	[8], [9]
Docked	12 April 2021, in an active GEO slot	[8], [9]
Function assumed	station-keeping and pointing; client propulsion and attitude control disabled	[9]
Service	five years of extended life	[13]

The published record has the vehicle operating with its original client as of 2022 [14] and as of 2025 [15], and gives its launch as 2020 [9]; the vendor fact sheet gives the 15 August 2020 rideshare date [13].

What is not established

MEV-2’s own performance is documented mainly by the vendor and by two conference papers from the vehicle’s own sensor engineers; no independent flight-data review of the docking or station-keeping handover has been published [8], [9], [13]. The sensor paper reports detection ranges as thresholds exceeded, with no navigation accuracy, covariance or detection probability [3]. The three NASA ISAM/OSAM state-of-play editions share large blocks of verbatim text and are not three independent confirmations [10], [14], [15].

SMART-OLEV’s docking figures were pre-flight projections published by the company selling the vehicle, and SMART-OLEV never flew [2]. ETS-VII and Orbital Express flew against cooperative or purpose-built clients; the ETS-VII results are single values per flight from three docking events, so they support no reliability or capture-envelope claim [18], [20]. The SPHERES docking-port work was validated on a flat floor and a reduced-gravity aircraft, never on orbit [5]. The ring capture tool [27] and the tumbling-target guidance [26] are ground or simulation results and do not bound MEV’s hardware. The OSAM-1 propellant transfer results are ground qualification of a canceled mission [4], [29]. The market argument rests on secondary insurance and failure-rate statistics assembled by policy researchers, not operator data [1], and the thruster-reuse and depot-sharing figures are optimization outputs built on assumed cost and performance parameters [11], [12].