Automatic Scheduling for Earth Observation Satellite With Temporal Specifications

Jianqing Li,Chaoyong Li,Feng Wang
DOI: https://doi.org/10.1109/taes.2020.2966902
IF: 3.491
2020-08-01
IEEE Transactions on Aerospace and Electronic Systems
Abstract:This article investigates a mission scheduling problem for earth observation satellite (EOS) with specific and temporal requirements. As is well known, the challenge is how to automatically generate an optimal actionable sequence for EOS given various physical constraints and temporal/spatial specifications that are often too intuitive to reason. To this end, we propose an automatic scheduling algorithm for EOS with the help of a linear temporal logic (LTL). In particular, LTL semantics are introduced to automate the constraints as well as temporal specifications, such that the argument of the mission specifications can be properly formulated. Toward this, we proceed to construct the state transition system by abstracting the mission objectives and specifications into a directed graph with weighted edges, and the automata theory is applied to obtain the observation sequence that satisfies the said specifications. Numerical examples verified the performance of the proposed strategy.
telecommunications,engineering, electrical & electronic, aerospace
What problem does this paper attempt to address?
The paper is primarily dedicated to addressing the task planning and scheduling problem of Agile Earth Observation Satellites (AEOS), particularly how to automatically generate the optimal task execution sequence while satisfying various physical constraints and time/space-specific requirements. Specifically, the main contributions of the research include: 1. **Problem Background**: With the development of observation technology and the increase in task demands, traditional Earth Observation Satellite (EOS) task planning methods have become inadequate for handling more complex task requirements. AEOS, due to its flexible maneuverability, can perform more complex observation tasks, such as extending observation windows and flexibly selecting observation targets. Therefore, its task planning becomes more complex and challenging. 2. **Proposed Method**: The authors propose an automatic task scheduling algorithm based on Linear Temporal Logic (LTL) to solve the AEOS task planning problem. This method uses the semantics of LTL to automatically describe constraint conditions and time-specific requirements, allowing task specifications to be appropriately formulated. The specific steps are as follows: - Use LTL semantics to formalize the desired observation requirements. - Abstract tasks and constraint conditions to construct a state transition system. - Apply automata theory to generate observation sequences that meet the specified specifications. 3. **Key Technologies and Assumptions**: The paper details the foundational knowledge of LTL semantics, the construction method of the state transition system, and the application of automata theory. It also provides some basic assumptions to simplify the problem, such as ignoring the impact of data transmission conditions, cloud cover, and sunlight. 4. **Case Analysis**: The paper demonstrates how to use the proposed LTL specifications to describe complex observation requirements through a specific example and provides the corresponding mathematical model and solution strategy. In summary, this paper aims to solve the automatic scheduling problem in AEOS task planning, particularly providing an effective solution for handling complex temporal requirements.