AAS 01-181 ORBIT DETERMINATION ACCURACY REQUIREMENTS FOR COLLISION AVOIDANCE Robert G. Gottlieb, Steven J. Sponaugle, David E. Gaylor A simulation has been developed to examine how orbit determination accuracy, size of avoidance maneuvers, encounter geometry and warning time affect the probability of collision. The simulation shows that reasonably sized collision avoidance maneuvers are effective in reducing the probability of collision only if accurate orbit determination data is available for both the active satellite and the threat. Also, a method to determine the maneuvers required to reduce the probability of collision to a desired value is presented. INTRODUCTION A collision between satellites in a constellation and other objects in space could potentially destroy one, many or possibly even all satellites in the constellation. In order to build an effective collision avoidance system for such a constellation, trade studies must be performed to understand the sensitivity of the probability of collision (P c ) to various characteristics of the system. The purpose of this paper is to present the orbit determination (OD) accuracy required for an effective collision avoidance system for a commercial satellite constellation. In addition, this paper will present methods for predicting potential collisions between satellites, computing the probability of collision and determining orbital maneuvers to avoid those collisions GENERIC COLLISION AVOIDANCE SIMULATOR The Generic Collision Avoidance Simulator (GCAS) was developed to support analysis of collision avoidance requirements. Some mathematical simplifications were made in the formulation of GCAS to allow very fast run times to enable exploration of the entire design trade space. The following assumptions are made in GCAS: 1. GCAS simulates two satellites in near-circular orbits at the same inclination and semi-major axis, but with different values of RAAN ( 1 deg 359 deg ≤ ∆Ω ≤ ). All motion is described by the restricted two body equations. 2. Maneuvers performed by active satellites produce a change in the close approach distance in only the intrack direction. No maneuvers to change inclination or RAAN are considered and only small changes in mean semi-major axis (< 20 m) are considered. The change in radial and crosstrack distance due to any collision avoidance maneuver is small compared to the intrack change. GCAS simulates close approaches by computing the orbit elements necessary for a given miss distance to occur at the point of closest approach. These orbit elements are converted to J2000 Earth Centered Inertial (ECI) state vectors and fed into the probability of collision calculation, which is described in the Appendix. The user may vary the initial covariance matrices, the difference in RAAN between the two objects, the miss distance, and the prediction time span to determine the effects on the probability of collision. Though orbit determination performance is usually quoted in terms of position error and velocity error, a more meaningful error criteria is the semi-major axis error. Semi-major axis error propagated in time drives the intrack position uncertainty. This can be readily seen by analyzing the two body orbit equation for mean anomaly. In the two body problem, all orbit elements are constant except for mean anomaly, which changes as: ( ) 0 0 3 M M t t a μ = + - (1)