Abstract Spacecraft modularity is a powerful design methodology for shortening satellite development cycles and reducing costs. The exploitation of standardized and reusable modules, universal interfaces and protocols, generic design, and a set of defined conventions, allows for rapidly integrating a flexible and cost-effective platform with ease of component modifications. Anyhow, the implementation of modularity poses several issues at any level of mission and spacecraft design, including subsystem design. This paper deals with the Attitude Determination and Control Subsystem (ADCS), with a particular accent on making the estimation software independent of any specific requirements dictated by the application in question. In this respect, a generic and reusable Murrell's Multiplicative Extended Kalman Filter (MMEKF) that can operate on different satellites, with diverse estimation requirements, working with changeable integrated ADCS boxes and sensor suites, but without changes either in the equations or in the interface with the rest of the system, has been proposed. The algorithm has been designed to maximize flexibility and ease of applicability to different missions without a priori-knowledge of the actual configuration. The execution of the algorithm is based on a plug-and-play logic in which middleware software enables the automatic discovery of the available sensors and the transmission of required data to the estimation algorithms. An adaptable and modular 3D simulator has been built to test and validate the performance of the generic MMEKF over multiple test cases with assigned configurations. The results of the simulations confirm the capability of the algorithm to meet estimation requirements for each simulated case, without ad-hoc, case-by-case, optimization, or customization.

Highlights • A generic and reusable estimation algorithm is proposed to estimate the attitude in a modular plug-and-play satellite. • The algorithm is designed by adapting Murrell's Kalman Filter to be fully parametric and mission abstracted. • The method involves a plug-and-play execution based on centralized middleware software. • Relevant requirements and constraints to reuse the same estimation algorithm in heterogenous missions are pointed out. • Performance in the estimation accuracy is verified over multiple cases with assigned configurations.