Trace Repair for Temporal Behavior Trees

2025-09-10

AI Summary

Repairing execution traces of robots and cyber-physical systems under Temporal Behavior Tree (TBT) specifications remains challenging due to the prohibitive computational cost and poor scalability of existing Mixed-Integer Linear Programming (MILP)-based approaches. Method: This paper proposes two efficient trace repair strategies—incremental repair and landmark-guided repair—leveraging the robust semantics of TBTs. Both methods approximate MILP-based repair using linear programming and piecewise iterative optimization, thereby avoiding combinatorial explosion. Contribution/Results: The proposed methods significantly improve scalability and efficiency: on trace instances with over 100,000 time steps, repair completes within ≤10 minutes, whereas MILP fails due to memory overflow. Furthermore, the framework enables fault attribution analysis and high-quality training sample generation, establishing a novel paradigm for trustworthy operation and maintenance of TBT-driven systems.

Technology Category

• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

We present methods for repairing traces against specifications given as temporal behavior trees (TBT). TBT are a specification formalism for action sequences in robotics and cyber-physical systems, where specifications of sub-behaviors, given in signal temporal logic, are composed using operators for sequential and parallel composition, fallbacks, and repetition. Trace repairs are useful to explain failures and as training examples that avoid the observed problems. In principle, repairs can be obtained via mixed-integer linear programming (MILP), but this is far too expensive for practical applications. We present two practical repair strategies: (1) incremental repair, which reduces the MILP by splitting the trace into segments, and (2) landmark-based repair, which solves the repair problem iteratively using TBT's robust semantics as a heuristic that approximates MILP with more efficient linear programming. In our experiments, we were able to repair traces with more than 25,000 entries in under ten minutes, while MILP runs out of memory.

Problem

• Repairing traces against temporal behavior tree specifications
• Reducing computational cost of mixed-integer linear programming
• Providing practical repair strategies for large-scale systems

Innovation

• Incremental repair splits trace segments
• Landmark-based repair uses TBT robust semantics
• Efficient linear programming replaces MILP