MemoGuard: An Adaptive Runtime for Guarding Against Memory Traps in Communication-Limited Robot Navigation

2026-07-16

AI Summary

This study addresses the “memory trap” problem in communication-constrained robotic systems, where reusing highly similar past memories can lead to task failure or safety hazards—particularly under environmental topology changes or insufficient resources. The work formally characterizes this issue and introduces MemoGuard, a lightweight runtime mechanism that validates memory reuse by modeling experiences as graphs and verifying multi-dimensional contracts encompassing topology, resource availability, and expected outcomes. Only upon validation failure does the system fall back to local inference using a large language model (Llama3.2-3B). Experimental results in corridor inspection simulations demonstrate that, compared to similarity-based memory reuse alone, MemoGuard reduces battery safety violations by 76.6%; relative to always performing local inference, it decreases fallback invocations by 21.4%, saving 3.67 seconds and 36.97 joules per task.

Abstract

Communication-limited robots in mission-critical scenarios such as disaster inspection and search-and-rescue must make reliable onboard decisions without access to remote operators or high-capacity reasoning services. Episodic memory reuse is an attractive low-cost fallback, but retrieval similarity does not guarantee execution validity, i.e., a retrieved action may match the current context yet be unsafe due to changed topology, insufficient battery margin, or unreliable prior outcomes. We call such high-similarity but execution-invalid episodes memory traps. This creates a safety-efficiency design space where similarity only reuse minimizes fallback cost but can be unsafe, while always invoking local reasoning improves safety at high computational and energy cost. This paper presents MemoGuard, a lightweight adaptive runtime that validates episodic memories against topology, resource, and outcome contracts before reuse, invoking fallback only when validation fails. In a graph-based corridor-inspection simulator, MemoGuard reduces battery safety violations by 76.6% over similarity-only top-1 reuse while reducing fallback calls by 21.4% over always reasoning. On an NVIDIA Jetson AGX Xavier with local llama3.2:3b fallback reasoning, this corresponds to 3.67 s and 36.97 J of avoided fallback-reasoning overhead per trial. We open-source MemoGuard at https://github.com/hetheiin/memoguard.

Problem

• memory traps
• communication-limited robot navigation
• episodic memory reuse
• safety-efficiency tradeoff
• execution validity

Innovation

• memory traps
• episodic memory reuse
• adaptive runtime
• safety validation
• resource-aware reasoning