Satellite Collision
Avoidance

In orbital operations, a conjunction is a predicted close approach between two space objects. One object is usually treated as the protected or primary spacecraft. The other is the secondary object, which might be another active satellite, a spent rocket body, or a fragment of debris. The prediction reaches its smallest relative distance at the time of closest approach, commonly abbreviated TCA.

A conjunction does not mean that a collision will happen. It means a screening system has found an encounter close enough to deserve more analysis. The first prediction may be based on limited observations, an uncertain orbit, or a maneuver plan that has not yet been included. As new measurements arrive, the estimated trajectories and their uncertainty can change. An event that initially looks alarming may become harmless, while another may become more concerning.

This distinction matters because satellites move quickly and the orbital environment is crowded, especially in low Earth orbit. Operators cannot respond to every geometric close approach with an immediate burn. Propellant is limited, maneuvers interrupt missions, and an unnecessary change can create a new conjunction with a different object. Collision avoidance is therefore a repeated decision process, not a last-second steering reflex.

Conjunction assessment starts with orbit knowledge. Ground radars, optical sensors, and spacecraft operators produce observations or ephemerides that estimate where objects are and how they are moving. Orbit determination converts those measurements into a state estimate. Propagation models then predict each state forward over a screening interval.

The protected spacecraft trajectory is compared with a catalog of tracked objects. Screening first uses a volume around the predicted trajectory to identify candidates efficiently. A candidate that enters this screening volume receives more detailed analysis. NASA describes operational screenings for protected assets as the first step in a three-stage process: conjunction prediction, close-approach risk assessment, and risk mitigation.

Prediction is only as useful as its data. An operator ephemeris may include planned maneuvers and high-quality navigation data unavailable in a public orbital element set. A catalog orbit for non-cooperative debris may depend entirely on remote observations. Each estimate therefore carries uncertainty, and that uncertainty must be propagated to TCA along with the nominal trajectory.

A Conjunction Data Message, or CDM, is a standardized way to exchange information about a predicted close approach. The CCSDS format allows different organizations and software systems to communicate using the same structure. A CDM can include object identifiers, TCA, relative position and velocity, miss distance, reference frames, covariance information, and metadata about how the solution was produced.

The CDM is not a command to maneuver. It is an input to analysis. Operators may receive a sequence of messages for the same event as new observations refine the orbits. The history is often as important as one message: analysts watch whether miss distance, uncertainty, and probability of collision are stabilizing, increasing, or decreasing.

Data quality also matters. A covariance matrix is useful only when it realistically represents prediction error. If uncertainty is too optimistic, calculated risk can be misleadingly low. If it is too broad or poorly modeled, the result can be overly conservative. Operational teams therefore examine both the numbers and the assumptions behind them.

Miss distance is the predicted separation between the object centers at TCA. It is intuitive, but it cannot describe collision risk by itself. Two events with the same miss distance can have very different uncertainty. In one case, both trajectories may be tightly known and clearly separated. In another, the predicted positions may have broad overlapping uncertainty regions.

Analysts often study the relative geometry in an encounter plane perpendicular to the relative velocity. The primary object can be represented by a combined hard-body radius that accounts for the effective sizes of both objects. After both covariance matrices are expressed in a common frame, they are combined and projected into that plane. The resulting relative-position probability distribution is centered on the predicted secondary position. Collision probability, usually written Pc, is the probability mass from that distribution that lies inside the hard-body region.

This explains a counterintuitive result: a very small nominal miss distance does not automatically produce the highest Pc. If uncertainty is enormous, the probability density can be spread across a large region, leaving only a small fraction inside the collision area. Conversely, a somewhat larger miss distance with a compact uncertainty region positioned near the hard-body boundary can produce a more significant Pc. Neither number should be interpreted without the other.

Probability of collision is an important metric, but there is no single threshold that every spacecraft must use in every situation. ESA has described 1 in 10,000 as a typical operational level for initiating maneuver preparation in some contexts. NASA missions use documented operational plans, mission-specific criteria, and multiple risk indicators. A threshold is therefore a policy within a wider decision framework, not a law of orbital mechanics.

Teams also consider the expected consequence. A close approach involving a large intact rocket body may have different environmental implications from one involving a small fragment. Human spaceflight applies additional safety considerations. A spacecraft with large solar arrays, limited propulsion, sensitive pointing requirements, or a critical observation schedule may face different maneuver constraints.

Analysts also ask whether the probability estimate is stable. If TCA is several days away and new tracking is expected, waiting for another update may reduce uncertainty without consuming propellant. If the decision deadline is approaching, delaying can remove safe maneuver options. Operations is a balance between better information and remaining response time.

Orbital predictions change because observations contain error and because the physical environment is not perfectly known. Atmospheric density affects drag in low Earth orbit and varies with solar activity, local time, altitude, spacecraft attitude, and weather in the upper atmosphere. Solar radiation pressure, maneuvers, attitude changes, and modeling choices can also influence propagation.

Covariance normally grows as prediction moves away from the latest well-observed state. New measurements can shrink or rotate that uncertainty. They may also move the nominal trajectory. That is why a sequence of CDMs can show Pc rising, falling, or fluctuating even though neither object has maneuvered.

A public TLE and SGP4 propagation are valuable for visualization and general tracking, but they are not a replacement for an operational conjunction solution. Public element sets do not normally provide the complete covariance and operator intent needed for a defensible avoidance decision. High-quality assessment uses the best available ephemerides, uncertainties, object sizes, planned maneuvers, and operational constraints.

A collision-avoidance maneuver changes the spacecraft's future state so that the predicted encounter geometry becomes safer. The burn can be small because its effect accumulates over time. A slight change in velocity performed hours or days before the original TCA changes where the spacecraft will be when the secondary object reaches the encounter region. Recomputing both trajectories after the maneuver may shift the updated TCA or remove the conjunction from further concern.

There is no universal burn direction. Depending on geometry and mission constraints, a maneuver may change along-track timing, radial separation, cross-track separation, or a combination. The selected option must preserve spacecraft safety, power, thermal limits, communications, payload operations, and future orbit requirements. Teams also check whether the candidate maneuver creates new conjunctions.

After execution, orbit determination resumes. The updated ephemeris is screened again against the catalog, and analysts verify that the original risk has fallen without introducing unacceptable new events. Collision avoidance therefore ends with confirmation, not with engine cutoff.

Most alerts do not become maneuvers. Many are filtered out as tracking improves, the objects pass with adequate separation, or uncertainty no longer overlaps the collision region. Maneuvering for every early warning would consume propellant, interrupt missions, and make the traffic picture less predictable.

Coordination is especially important when both objects are active. Operators exchange ephemerides and maneuver intentions so they do not both move in incompatible ways. When the secondary object is debris, only the active spacecraft can respond. A satellite without propulsion may use attitude changes or differential drag in limited situations, but those methods require time and suitable geometry.

The correct response can therefore be continued monitoring, better tracking, operator coordination, maneuver preparation, maneuver execution, or a combination in sequence. Responsible operations means preserving options while avoiding unnecessary disturbance to the shared orbital environment.

Jewawud Orbital Tracker (JOT) helps visitors visualize public satellite positions, orbital paths, altitude classes, and the changing geometry around Earth. It can build intuition about why many objects occupy similar regions and why their ground tracks and orbital planes intersect when viewed on a map or globe.

JOT is not an operational collision-warning system. A visual overlap on a 2D map does not mean two satellites are close in three-dimensional space or present at the crossing point at the same time. The display also does not provide the operator-grade covariance, hard-body radii, maneuver intent, high-accuracy ephemerides, and validated risk algorithms required for real conjunction assessment.

The safest way to use a public tracker is educationally: inspect the orbit, identify altitude and inclination, observe how quickly geometry changes, and then use this article to understand what professional systems add. Public visualization is the beginning of orbital awareness, not the final safety decision.

"The orbit lines cross, so the satellites will collide." Crossing paths are not enough. The objects must reach the same three-dimensional location at the same time, within their combined physical size.

"The smallest miss distance always has the highest risk." No. Position uncertainty, encounter geometry, and object size change Pc.

"A high initial Pc means impact is certain." No. It is a probability estimate based on the current data and model. It can change as new observations arrive.

"Avoidance is a last-second dodge." Operational maneuvers are normally planned ahead so a small velocity change can accumulate into useful separation by TCA.

"Any satellite tracker can issue collision warnings." No. Reliable assessment requires validated ephemerides, covariance, object-size assumptions, consistent reference frames, and operational analysis.

What is TCA? The currently predicted time of closest approach between the two trajectories.

What is a CDM? A standardized Conjunction Data Message carrying close-approach geometry, covariance, object, and solution metadata.

What is Pc? The estimated probability that the true relative position at TCA falls within the combined collision region.

Can debris maneuver? Normally no. When the secondary object is uncontrolled debris, the active spacecraft must provide any avoidance action.

Does every close approach require a burn? No. Most alerts are monitored and refined without a maneuver.

NASA Conjunction Assessment - operational definitions and the three-step assessment process.

NASA CARA Services - screening, risk analysis, ephemeris, covariance, and mitigation support.

NASA Close Approach Risk Assessment - probability of collision and uncertainty-based risk interpretation.

ESA Reentry and Collision Avoidance - CDM processing, risk factors, and operational conjunction services.

ESA and the CCSDS Conjunction Data Message - standardized exchange of collision-warning information.