Autonomous Tip & Cue on-board Scheduler for Earth Observation

The rapid growth in satellite deployments, driven by increasingly accessible launch opportunities, has led to dense constellations that must autonomously coordinate complex operations. This work focuses on the scheduling problem for heterogeneous constellations composed of two satellite classes: tip and cue satellites. Tip satellites are equipped with wide-field-of-view (FoV) or lower resolution sensors for broad-area detection of events or ground targets of interest. Upon detection, a tip satellite selects and tasks a cue satellite, which carries a higher resolution payload for targeted acquisition and detailed characterization of the alerted area. All satellites in the constellation feature Inter-Satellite Link (ISL), enabling direct data exchange without immediate reliance on ground stations, whose availability is limited by communication windows and operational costs. Autonomous scheduling must therefore account for a diverse set of constraints, which include both physical factors, such as orbital geometry, targets visibility, and downlink opportunities, and device-level limitations related to on-board resources. These constraints span memory occupation, battery usage, CPU load distribution between the on-board computer and payload processors, and ISL availability. The objective is to enable each satellite to autonomously generate feasible schedules not only for itself but also for other nodes in the constellation when a critical event arises. A detailed satellite model is constructed, capturing the system dynamics, platform architecture, operational tasks, and their associated constraints, and a set of scheduling algorithms are designed and validated using a high-fidelity satellite simulator. This multi-stage process supports the development of robust autonomous scheduling strategies for heterogeneous satellite constellations operating in real-world scenarios.