Autonomous navigation in real-world environments often requires long-horizon operation under partial observability and evolving goals. Existing mapless navigation approaches predominantly assume fixed goals throughout an episode, limiting applicability to practical scenarios where goals change online during execution. We study online chained-goal navigation, where goals evolve through existing-goal redirection or new-goal emergence without explicit transition signals. Under this setting, we find navigation may require either continued exploration or strategic revisitation, creating conflicting memory requirements. We propose Global-Local Spatio-Temporal Belief Learning (GLST), a memory-decomposition framework that separates latent beliefs into global-local spatial and long-short temporal representations for long-horizon reasoning. We further introduce goal-conditioned rewards to balance exploration and exploitation under evolving goals. Extensive experiments in simulation across diverse terrains demonstrate improved chained-goal navigation performance over baselines. Real-robot experiments using the Spot quadruped further validate our memory decomposition framework.
Figures & tables
Figure 1: Overall framework and training pipeline of the proposed GLST method. (a) GLST framework showing the global-local spatial and temporal belief branches and actor-critic policy. Real-robot depth images or simulated height scans serve as shared inputs to both spatial branches. (b) Training pipeline including locomotion policy training, global spatial supervised pretraining, environment generation with chained-goal sampling, and navigation policy training.
Figure 2: Performance on chained-goal navigation. (a) Chained-goal success rate (SR), where empirical compounding is compared with independent-goal assumptions. (b) Single-goal SR and failure decomposition into timeout, collision, and excessive roll. (c) Navigation efficiency measured by SPL and optimal path length distributions. Increasing SPL across goals suggests improved efficiency under comparable task difficulty.
Figure 3: Visualization of learned spatial beliefs in two simulation environments: (1) maze and (2) pillars. (a) Global spatial beliefs reconstructed from predicted height, mask, and confidence maps, with robot trajectories overlaid. (b) Local observations obtained by fusing depth images from five cameras (front-left, front-right, left, right, and rear) into local height maps. (c) Simulation environments with the quadruped robot and executed trajectories overlaid.
Method
# Params (M)
Memory
Stats
Path Length 5 × [4,15] m
Path Length 5 × [4,20] m
SR ↑
SPL ↑
TO ↓
Coll. ↓
Roll ↓
SR ↑
SPL ↑
TO ↓
Coll. ↓
Roll ↓
(%)
(%)
(%)
(%)
(%)
(%)
(%)
(%)
Ours
5.69
GLST
Mean
79.2
0.70
10.1
9.1
1.6
64.6
0.56
20.4
12.9
2.1
Std
0.8
0.01
0.6
0.9
0.5
1.8
0.02
1.1
0.7
0.3
SRU [ 2 ]
11.8
[ 2 ]
Mean
64.2
0.54
15.8
19.1
0.9
50.3
0.42
21.2
27.9
0.6
Std
2.8
0.03
2.1
1.5
0.4
3.2
0.03
3.0
1.8
0.4
Table 1: Navigation results over 3 independent runs of 1000-episode simulations. Failure modes include timeout (TO), collision (Coll.), and excessive roll (Roll). Best SR and SPL are bolded and underlined, while lowest failure rates are underlined.
Figure 4: Real-robot navigation experiments on the Spot quadruped across diverse indoor and outdoor environments. (1) Indoor experiments. (2) Outdoor forest experiments. (a) Experimental scenes with ghost robot overlays illustrating executed trajectories. (b) Inferred global spatial beliefs reconstructed from latent representations. (3) Long-horizon online chained-goal experiments with dead ends and dynamic obstacles. Additional chained-goal indoor and outdoor experimental results are provided in Appendix G .
Appendix figures & tables5 assets
Supplementary material from the paper’s appendix.
Appendix
Figure 5: Chained-goal success rate (%) as a function of goal index ng , where ng=1,…,5 denotes the ng -th goal in the sequence. The lower figures show the compounded success rate from 100% to the final success rate under the assumption of independent goal completion. The upper figures show the single-goal success rates reported in prior work, which are used to construct the chained-goal success rates. Data are adapted from (a) [ 2 ] , (b) [ 40 ] , (c) [ 41 ] , (d) [ 42 ] , and (e) [ 9 ] .
Figure 6: Detailed architecture of the proposed GLST framework. Observations are decomposed into global-local spatial and temporal belief representations. Global beliefs are recursively accumulated and refined through goal-conditioned attention, while fused belief representations are processed by a shared actor-critic policy. Auxiliary prediction heads supervise spatial belief learning through height, mask, and confidence reconstruction.
Terrain
Goal
Path length
# Episodes
Speed
SR (%) ↑
SPL ↑
Failure modes (%)
(m)
(m/s)
Timeout ↓
Collision ↓
Roll ↓
Path Length: 5×[4,15] m
Maze
Overall
46.4±0.7
1000
1.1
79.2±0.9
0.70±0.01
10.1±0.7
9.1±1.1
1.6±0.5
G1
9.3±0.3
1000
1.0
90.6±0.3
0.84±0.01
5.0±0.9
4.0±0.8
0.4±0.2
G2
9.2±0.3
906±3
1.1
94.4±0.5
0.85±0.01
3.2±0.2
2.0±0.2
0.4±0.2
G3
9.4±0.0
856±5
1.1
96.5±0.7
0.86±0.01
1.9±0.4
1.2±0.4
0.2±0.1
Appendix
Table 2: Chained-goal navigation statistics across terrains and path-length settings. Results are aggregated over 3 runs of 1000-episode simulations and reported as mean ± std across runs.
Method
# Params (M)
Memory
Stats
Path Length 5 × [4,15] m
Path Length 5 × [4,20] m
SR ↑
SPL ↑
TO ↓
Coll. ↓
Roll ↓
SR ↑
SPL ↑
TO ↓
Coll. ↓
Roll ↓
(%)
(%)
(%)
(%)
(%)
(%)
(%)
(%)
Ours
5.7
GLST
Mean
79.8
0.77
7.9
10.7
1.6
64.2
0.61
12.1
19.7
4.0
Std
4.8
0.05
2.4
2.5
0.9
5.4
0.05
2.6
2.7
1.0
SRU [ 2 ]
11.8
[ 2 ]
Mean
64.7
0.54
15.0
18.9
1.4
50.7
0.41
22.0
26.7
0.6
Std
2.9
0.03
2.2
1.4
0.4
3.3
0.03
2.7
1.9
0.3
Appendix
Table 3: Navigation results over 1,000-rollout simulations ( ×10 seeds). Failure modes include timeout (TO), collision (Coll.), and excessive roll (Roll). Best SR and SPL are bolded and underlined, while lowest failure rates are underlined.
Figure 7: Additional real-robot experiments: (a) single-goal navigation on stairs, (b) indoor navigation with three chained goals, and (c) forest navigation with three chained goals. Images are contrast-enhanced for visualization.
Active exploration and semantic navigation require an embodied agent to build memory from partial observations, predict how the evolution of observed spatial memory may support future motion, and convert that prediction into actionable plans. We present GLAM, a goal-conditioned latent world model trained over global spatiotemporal memory, and GLAM NAV, the complete navigation system built around it. Given historical map tokens, a navigation goal, and the current robot pose, GLAM jointly predicts future map representations and robot-centric waypoint latents, allowing future spatial context and navigation intent to be inferred in a shared representation space. The model follows a JEPA-like latent prediction paradigm, operates directly on map-level latent tokens rather than RGB reconstruction, and uses a pretrained waypoint encoder-decoder to supervise and decode navigation plans within GLAM NAV. Training data are collected by replaying ObjectNav expert trajectories in Habitat over HM3D v0.2 scene assets and slicing them into multi-timescale prediction samples. On a controlled HM3D-ObjectNav subset reproduction setting, GLAM NAV improves over a reproduced BSC-Nav baseline in both success rate and success weighted by path length.
I-Tak Ieong, Ruizhi Feng, Zhaoyang Lu +5
Tsinghua University · Lab for Brain-Inspired Embodied Intelligence, EBKernel Technologies Co., Ltd · Northeastern University +2
Object Goal Navigation (ObjectNav) refers to an agent navigating to an object in an unseen environment, which is an ability often required in the accomplishment of complex tasks. While existing methods demonstrate proficiency in isolated single object navigation, their limitations emerge in the restricted applicability of lifelong memory representations, which ultimately hinders effective navigation toward continual targets over extended periods. To address this problem, we propose OVAL, a novel lifelong open-vocabulary memory framework, which enables efficient and precise execution of long-term navigation in semantically open tasks. Within this framework, we introduce memory descriptors to facilitate structured management of the memory model. Additionally, we propose a novel probability-based exploration strategy, utilizing a multi-value frontier scoring to enhance lifelong exploration efficiency. Extensive experiments demonstrate the efficiency and robustness of the proposed system.
Jiahua Pei, Yi Liu, Guoping Pan +3
Shenzhen International Graduate School, Tsinghua University, Shenzhen 518055, China
Multimodal lifelong navigation requires an agent to autonomously explore unseen environments while sequentially completing navigation tasks specified by object categories, language descriptions, or reference images. Existing methods primarily accomplish these tasks by constructing state-centric semantic scene graphs. By treating scene graphs as persistent repositories of semantic observations, these methods struggle to distinguish similar instances, jointly represent semantic targets and exploration frontiers, and effectively exploit navigation memory and trajectory experience. To address these limitations, we propose Spatio-Temporal Event Graph Navigation (STEGNav), a training-free framework that extends conventional scene graphs into spatio-temporal event graphs along complementary spatial and temporal axes. The spatial axis performs query-conditioned instance grounding and jointly represents semantic targets and occupancy-aware exploration frontiers characterized by reachability, path cost, and exploration utility. The temporal axis employs trajectory-aware dual-window memory to retain recent decision--trajectory events and verified cross-subtask navigation outcomes. A VLM-based navigation agent reasons over the resulting spatio-temporal event graph and selects either a target instance or an exploration frontier as its next navigation goal. STEGNav achieves 66.3% SR and 39.7 SPL on GOAT-Bench, as well as SR scores of 64.0% and 69.4% on HM3Dv1 and HM3Dv2, respectively. Ablation studies and error analyses validate the complementary effects of the two axes, demonstrating that event-driven spatio-temporal representations improve navigation reliability and cross-subtask experience reuse.
Yang Chen, Zhenyu Huang, Wenbo Fu +4
State Key Laboratory of Novel Software Technology, Nanjing University · School of Intelligence Science and Technology, Nanjing University · School of Artificial Intelligence, Nanjing University