cs.ROAug 24, 2026

Resilience Matters for Embodied Agents System: New Metrics, Systematic Evaluation, and Optimization

Authors: Yapeng Liu, Yuanzhao Zhai, Xudong Gong, Dawei Feng, Bo Ding, Lin Wang, Huaimin Wang

Organizations: PDL Lab, College of Computer Science and Technology, National University of Defense Technology · State Key Laboratory of Complex & Critical Software Environment, Changsha 410073, Hunan, China · EmPACT Lab, Nanyang Technological University, Singapore

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Embodied Agents System (EAS) are increasingly deployed in open-world physical domains, where reliability directly dictates deployment quality and human-agent trust. However, existing evaluations rely on outcome-centric metrics as success rate or safety scores that collapse diverse execution trajectories into coarse scores, obscuring the dynamic processes underlying agent behavior. Therefore, they ignore a critical property of EAS -- which we define as the Resilience -- that reflects how EASs recover, stabilize, and extend under perturbations and across iterative updates. The lack of resilience is particularly critical in open-world environments due to continuous unexpected disruptions, thus directly affecting the quality of EAS deployment. To address this problem, we gain insight from the resilience-engineering concepts to EAS groundings and propose a novel resilience evaluation framework that can be flexibly applied to any EAS. Specifically, we define the first comprehensive resilience metrics suite for EASs system that exposes Rebound, Stability, and Graceful Extensibility across embodied tasks execution, providing a practical grounding for EAS resilience analysis. We further implement the resilience evaluation layer that transforms execution process into assessments for diagnosis and optimization. Across 400 household tasks with 10 EAS, we reveal the process-level distinction hidden by outcome metrics, including recovery cost differences among successful episodes (ΔCrec=25.2ΔC_{rec}=25.2), increased instability and task-family degradation. Metrics-guided optimizations reduce recovery cost and increase stability, graceful extensibility completion, showing the diagnostic effect of resilience evaluation. Our results reveal a trade-off among resilience characteristics, suggesting that a resilient EAS construction should be configured according to deployment-specific requirements.

Figures & tables

Figure 1: Outcome hide process resilience difference in EAS. (A) A resilient success trajectory handles inner and outer stress through resilient mechanisms, (B) A costly success relies on repeated backtrack, replan and unsafe behaviors.
Sym.Metric NameCalculationAspectTypeData Source
CrecC_{\mathrm{rec}}Recovery CostCrec=1Wrem∗(atd)∑tdtrΔτtτ∗(at)⋅gtrecC_{\mathrm{rec}}=\frac{1}{W_{\text{rem}}^{*}\left(a_{t_{d}}\right)}\sum_{t_{d}}^{t_{r}}\frac{\Delta\tau_{t}}{\tau^{*}\left(a_{t}\right)}\cdot g_{t}^{\text{rec}}ReboundD, SSystem Logs
β\betaPolicy Stabilityβ=sup⁡z,ρ∣ℓ(π,ρ(z))−ℓ(π,z)∣∥ρ∥\beta=\sup_{z,\rho}\frac{\left|\ell(\pi,\rho(z))-\ell(\pi,z)\right|}{\|\rho\|}StabilitySSensitivity Test
Mf(λ)M_{f}(\lambda)Stress Curve MappingMf(λ)=1K∑j=1Kr~(j),f(λ)−1M_{f}(\lambda)=\frac{1}{K}\sum_{j=1}^{K}\tilde{r}_{(j),f}(\lambda)-1ExtensibilityD, SStress Test
λf∗\lambda_{f}^{*}Stress Capacityλf∗=sup⁡{λ∈[0,1]∣Mf(λ)≥0}\lambda_{f}^{*}=\sup\{\lambda\in[0,1]\mid M_{f}(\lambda)\geq 0\}ExtensibilityS, AStress Grid
Table 1: Overview of resilience metrics with aspect, type ( D : deterministic outcome, S : statistical, A : analysis, T : time-series), and data source
Figure 2: Different CrecC_{\mathrm{rec}} under outcomes.
Figure 3: Resilience metrics construct validity results ( zoom in for better view ).
Figure 4: Resilience benchmark landscape.
MethodResilience MetricsOutcome Ref.Sim.
CrecC_{\mathrm{rec}} ( ↓\downarrow )β\beta ( ↓\downarrow )Stress Cap. λ∗\lambda^{*} ( ↑\uparrow )SRComp.Steps
ReAct6.5 ±\pm 0.460.239 ±\pm 0.0110.50 ±\pm 0.04238.5% ±\pm 3.1%57.8% ±\pm 4.6%3374 ±\pm 285
Reflexion57.3 ±\pm 5.070.299 ±\pm 0.0210.66 ±\pm 0.03556.3% ±\pm 4.2%75.0% ±\pm 3.8%4390 ±\pm 312
CycleVLA35.5 ±\pm 2.910.400 ±\pm 0.0260.72 ±\pm 0.04722.2% ±\pm 2.8%47.5% ±\pm 4.1%4688 ±\pm 420
CLARE21.6 ±\pm 1.680.344 ±\pm 0.0160.72 ±\pm 0.05228.4% ±\pm 3.5%46.0% ±\pm 4.5%4249 ±\pm 360
SayCan38.5 ±\pm 2.490.283 ±\pm 0.0130.78 ±\pm 0.04657.0% ±\pm 4.1%72.6% ±\pm 3.7%4682 ±\pm 380
Table 2: Comparative EAS resilience benchmark results in resilience aspects.
Figure 5: The Resilience Diagnosis. Visualizing the distinct behavioral archetypes of EASs.
TargetMethodsRebound CrecC_{\mathrm{rec}} ↓\downarrowRec. Win. ↓\downarrowβstep\beta_{\mathrm{step}} ↓\downarrowβ\beta ↓\downarrowGE. Comp. ↑\uparrowFormal GE
 Baseline59.113.130.28810.68340.887–
ReboundOptimized33.73 ↓\downarrow 42.94%2.38 ↓\downarrow 23.96%0.2079 ↓\downarrow 27.84%0.6870 ↑\uparrow 0.53%0.819 ↓\downarrow 7.67%–
 Baseline51.012.630.31850.67880.903–
StabilityOptimized43.01 ↓\downarrow 15.68%3.00 ↑\uparrow 14.07%0.2552 ↓\downarrow 19.87%0.6772 ↓\downarrow 0.24%0.750 ↓\downarrow 16.94%–
 Baseline36.701.331.45940.69170.833Incomplete
GEOptimized99.59 ↑\uparrow 171.36%2.25 ↑\uparrow 69.17%0.7215 ↓\downarrow 50.56%0.6371 ↓\downarrow 7.89%0.917 ↑\uparrow 10.08%Complete †
Table 3: Comparative results of metric-guided targeted improvements.
More in this topic

AI Agent Evaluation

AI agent evaluation measures whether tool-using systems complete tasks reliably, safely, and efficiently across interactive settings.

See all 519 papers

Explore similar work

CardsList
Feb 18, 2026

Towards a Science of AI Agent Reliability

AI agents are increasingly deployed to execute important tasks. While rising accuracy scores on standard benchmarks suggest rapid progress, many agents still continue to fail in practice. This discrepancy highlights a fundamental limitation of current evaluations: compressing agent behavior into a single success metric obscures critical operational flaws. Notably, it ignores whether agents behave consistently across runs, withstand perturbations, fail predictably, or have bounded error severity. Grounded in safety-critical engineering, we provide a holistic performance profile by proposing twelve concrete metrics that decompose agent reliability along four key dimensions: consistency, robustness, predictability, and safety. Evaluating 15 models across two complementary benchmarks, we find that recent capability gains have only yielded small improvements in reliability. By exposing these persistent limitations, our metrics complement traditional evaluations while offering tools for reasoning about how agents perform, degrade, and fail.

Stephan Rabanser, Sayash Kapoor, Peter Kirgis +3

Princeton University.

AI Agent ReliabilityAI Agent EvaluationAI Agent SafetyAI Agent Benchmarks
AskPDF
Apr 13, 2026

EmbodiedGovBench: A Benchmark for Governance, Recovery, and Upgrade Safety in Embodied Agent Systems

Recent progress in embodied AI has produced a growing ecosystem of robot policies, foundation models, and modular runtimes. However, current evaluation remains dominated by task success metrics such as completion rate or manipulation accuracy. These metrics leave a critical gap: they do not measure whether embodied systems are governable -- whether they respect capability boundaries, enforce policies, recover safely, maintain audit trails, and respond to human oversight. We present EmbodiedGovBench, a benchmark for governance-oriented evaluation of embodied agent systems. Rather than asking only whether a robot can complete a task, EmbodiedGovBench evaluates whether the system remains controllable, policy-bounded, recoverable, auditable, and evolution-safe under realistic perturbations. The benchmark covers seven governance dimensions: unauthorized capability invocation, runtime drift robustness, recovery success, policy portability, version upgrade safety, human override responsiveness, and audit completeness. We define a benchmark structure spanning single-robot and fleet settings, with scenario templates, perturbation operators, governance metrics, and baseline evaluation protocols. We describe how the benchmark can be instantiated over embodied capability runtimes with modular interfaces and contract-aware upgrade workflows. Our analysis suggests that embodied governance should become a first-class evaluation target. EmbodiedGovBench provides the initial measurement framework for that shift.

Xue Qin, Simin Luan, John See +2

School of Software, Harbin Institute of Technology, Harbin, China · School of Computer Science and Technology, Harbin Institute of Technology, Harbin, China · School of Mathematical and Computer Sciences, Heriot-Watt University, Malaysia Campus, Putrajaya, Malaysia +1

AI Agent GovernanceAI Agent Safety BenchmarksAI Agent BenchmarksEmbodied AI
AskPDF
May 11, 2026

Consistency as a Testable Property: Statistical Methods to Evaluate AI Agent Reliability

This paper establishes a rigorous measurement science for AI agent reliability, providing a foundational framework for quantifying consistency under semantically preserving perturbations. By leveraging UU-statistics for output-level reliability and kernel-based metrics for trajectory-level stability, we offer a principled approach to evaluating agents across diverse operating conditions. Our proposal highlights the important distinction between the core capability and execution robustness of an agent, showing that minor task-level variations can induce complete strategy breakdowns despite the agent possessing the requisite knowledge for the task. We validate our framework through extensive experiments on three agentic benchmarks, demonstrating that trajectory-level consistency metrics provide far greater diagnostic sensitivity than traditional pass@1 rates. By providing the mathematical tools to isolate where and why agents deviate, we enable the identification and rectification of architectural concerns that hinder the deployment of agents in high-stakes, real-world environments.

Harsh Raj, Niranjan Orkat, Suvrorup Mukherjee +3

Northeastern University · University of P´ecs · University of Michigan +2

AI Agent ReliabilityAgent ReliabilityAI Agent Evaluation
AskPDF