Using Explainable Scheduling for the Mars 2020 Rover Mission

Explainability is critical for trust in automated systems: The paper demonstrates that for the Mars 2020 Perseverance Rover mission, understanding why a scheduling system makes decisions—rather than just accepting its outputs—is essential for operators to trust and fully utilize the system in high-stakes applications.

Transparency helps identify missing constraints: When a scheduler produces invalid schedules, the paper shows that visual inspection with explanatory tools can help users quickly identify what constraints were overlooked, making the system more useful for iterative problem-solving rather than a "black box."

Explainability reduces debugging time for complex failures: The paper addresses a key challenge: when a scheduler fails to schedule activities due to conflicting constraints, Crosscheck's explanatory capabilities help operators understand which specific constraints caused the failure—information that would otherwise be difficult and time-consuming to determine manually.

Practical guidance improves system usability: Beyond explaining failures, the tool provides actionable recommendations for how to modify constraints to achieve desired schedules, transforming the system from a passive decision-maker into an interactive problem-solving partner for mission planners.

Explainable AI enables operational deployment: The paper notes that both the scheduling system and Crosscheck became "the baseline for operational use" for Mars 2020, demonstrating that explainability features are not optional add-ons but essential components for deploying AI systems in real-world critical missions.