cs.ROOct 7, 2026

Lifelong small-object navigation in changing object layouts: a benchmark and method

Authors: Jiagan Huang, Zikun Zhou, Zijian Ni, Hongpeng Wang, Guangming Lu, Jun Yu, Wenjie Pei

Organizations: Harbin Institute of Technology, Shenzhen

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Household robots need to continually navigate to different objects in the same environment, many of which are small and portable, such as tools and toys. Their small visual footprint and frequent occlusion make reliable observation difficult, and they may be moved by people without the robot observing the changes. We formulate this challenging task as Lifelong Small-object Navigation in Changing Object Layouts (LiSoNav-COL). Agents must seek suitable viewpoints for reliable observation, accumulate and reuse scene knowledge to efficiently locate subsequent targets, and update outdated memory after object relocation. To eliminate the need for prior scene scanning, we also require agents to start navigation with empty scene memory. Although practical, this task still lacks benchmarks designed around its defining assumptions. To bridge this gap, we introduce LiSoNav-Eval, a dedicated benchmark spanning 28 indoor scenes with 45 small-object categories. Its lifelong navigation sequences include both unchanged and relocated targets to evaluate memory reuse and adaptation to object relocation. To address this challenging task, we propose a navigation method based on multi-view Inspection with Viewpoint-Anchored Memory, dubbed IVAM-Nav. IVAM-Nav actively observes supporting surfaces from complementary viewpoints for reliable small-object perception and anchors the resulting memory to their observation viewpoints, supporting relational memory reuse and revalidation under similar viewing conditions. Extensive experiments on LiSoNav-Eval demonstrate favorable performance of IVAM-Nav against representative methods. Benchmark analyses also show that smaller objects, larger environments, and longer relocation distances pose greater challenges. The dataset and code are available here.

Figures & tables

Figure 1: An example navigation sequence in LiSoNav-Eval, in which the agent is instructed to find the CAN on the table, the SPOON on the counter, the NEWSPAPER on the table, and other targets in sequence. Starting with no prior scene memory, the agent searches for the CAN while constructing memory from observed objects, which is later reused to find the SPOON . Between episodes, randomly selected objects are relocated without the agent observing the changes. When searching for the NEWSPAPER , the agent first revisits the remembered location, finds the newspaper missing, updates the scene memory, and then locates the newspaper elsewhere. This example illustrates how LiSoNav-Eval evaluates small-object navigation as agents accumulate scene knowledge, reuse it across tasks, and respond to changes in object locations.
 ScenesSubtasksSemantic target categoriesObject relocationRelocated subtasksRelocated ratioPrior scene knowledgeSubtasks per sequence
MultiON 3-ON val ( Wani et al., 2020 )113,1501No––No3.00
GOAT val-unseen ( Khanna et al., 2024b )362,66936No––No7.41
OpenIN †\dagger ( Tang et al., 2025 )4––Yes–100%Yes4–5
LMEE-Bench test ( Wang et al., 2026 )3682838No––No4.99
LiSoNav-Eval (Ours)281,19945Yes54545.45%No8.27
Table 1: Comparison of LiSoNav-Eval with existing lifelong navigation benchmarks. †\dagger indicates that the corresponding evaluation data are not publicly released.
Figure 2: Overview of IVAM-Nav. IVAM-Nav maintains a viewpoint-anchored graph memory through three processes: (a) Memory construction with multi-view inspection actively observes supporting surfaces from complementary viewpoints to acquire reliable small-object evidence. (b) Viewpoint-anchored memory reuse aggregates relational evidence around historical viewpoints to retrieve those relevant to the current query. (c) Viewpoint-anchored memory revalidation and update revisits retrieved historical viewpoints to verify memory under similar viewing conditions and locally update inconsistent information.
MethodCross-task MemoryStale-memory RevisionOverall SR ↑\uparrowOverall SPL ↑\uparrowHRRelocated-target SR ↑\uparrow
VLFM ( Yokoyama et al., 2024 )No–12.3%8.8%27.6%11.7%
AstraNav ( Hu et al., 2026 )YesNo12.2%9.5%23.2%9.2%
DualMap ( Jiang et al., 2025 )YesYes13.9%10.2%35.2%10.3%
DynaMem ( Liu et al., 2025 )YesYes18.0%12.7%38.1%14.9%
DynaMem Oracle ( Liu et al., 2025 )YesYes20.1%14.4%39.6%15.8%
3D-Mem ( Yang et al., 2025 )YesNo21.9%15.5%34.3%17.2%
Table 2: Overall performance on LiSoNav-Eval. Oracle variants use ground-truth detection assistance as described in Sec. 5.1 .
Figure 5
MethodCross-task MemoryStale-memory RevisionSRSPL
StaticChangingDrop (%)StaticChangingDrop (%)
AstraNav ( Hu et al., 2026 )YesNo14.7%11.0%25.2%12.9%8.5%34.1%
DynaMem ( Liu et al., 2025 )YesYes20.8%17.6%15.7%15.7%12.6%19.7%
IVAM-NavYesYes40.8%36.7%10.0%25.6%21.4%16.4%
Table 4: Performance under static and changing-object-layout settings on the LiSoNav-Eval analysis subset. Relative drops are computed from the static to the changing setting.
Table 7
Table 8

Explore similar work

CardsList

  1. OmniNav: Robust Long-Horizon Target Navigation in Dynamic Environments

    Sep 8, 2026Yujie Tang, Meiling Wang, Jinhao Jiang +4Object Goal NavigationEnvironment

    School of Automation, Beijing Institute of Technology, Beijing, China.

  2. EvoNav-Bench: Benchmarking Lifelong Navigation in Evolving Environments

    Sep 8, 2026Xilin Wang, Guoxi Zhang, Hongming Xu +3Object Goal NavigationNavigation

    State Key Laboratory of General Artificial Intelligence, BIGAI · 2Tsinghua University

  3. CoReLIN: Constraint-based Reasoning for Zero-shot Lifelong Interactive Navigation

    Feb 23, 2026Apoorva Vashisth, Manav Kulshrestha, Pranav Bakshi +3Robot NavigationMobile Robots

    Purdue University · Purdue University, USA · IIT Kharagpur +5