Quadruped robots can traverse low obstacles, but many 2D planning pipelines still model obstacles as binary occupied regions and rely on sampling-based search that can be inefficient under a limited budget. We propose a perception-assisted height-adaptive planning framework based on CMP-IRRT*, a Channel Mamba PointNet-guided Informed RRT* planner. Given a calibrated top-view RGB observation, the perception module estimates obstacle regions and converts depth predictions into a ground-relative height map. The planner then performs height-conditioned collision checking, treating high obstacles as blocked while allowing low obstacles to be traversed, and uses the CMP guide to bias sampling toward promising regions while retaining standard free-space and informed sampling fallbacks. Experiments on 2D planning benchmarks show that CMP-IRRT* reduces explored nodes and iterations compared with classical and neural-guided baselines, and a controlled ablation supports the contribution of the Mamba-based guide. In constructed traversability-aware scenarios, the proposed planner reduces path length by up to 16.3% when low obstacles are traversable, and a Unitree Go2 demonstration further shows executable bypassing and traversal behaviors. Our code is publicly available at https://github.com/MingfanZhao/height-adaptive-planner.
Figures & tables
Figure 1: Overview of the proposed perception-assisted traversability-aware planning pipeline. The perception module estimates obstacle geometry from a calibrated top-view RGB observation. The planning module uses height-conditioned collision checking and CMP-guided informed sampling to find a low-cost path under a finite planning budget. The execution module decodes the path and traversal mode into robot commands.
Figure 2: Structure of the Channel Mamba PointNet guide used to predict promising search regions for CMP-IRRT*.
Figure 3: Comparison of planning performance on Center Block and 2D Random World problems. The compared methods include RRT* ( Karaman and Frazzoli, 2011 ) , IRRT* ( Gammell et al., 2014 ) , and NIRRT* ( Huang et al., 2024 ) . The plots report finite-budget planning behavior, including iterations to reach a cost threshold and path-cost improvement after an initial solution is found.
Method
RRT*
IRRT*
NIRRT*
CMP-IRRT*(Ours)
Initial Time(s)
0.17
0.19
1.81
1.37
Nodes
241.86
251.61
102.92
99.49
Iterations
299.04
312.75
151.69
146.55
Final Cost
233.01
232.63
206.51
203.41
Table 1: Average comparison results on the 2D Random World dataset. Lower values are better for all metrics.
Guide
Mamba
Channel Attn.
Nodes Ratio
Iterations Ratio
PointNet++
No
No
1.000
1.000
PointNet++ and Mamba
Yes
No
0.984
0.980
CMP
Yes
Yes
0.919
0.920
Table 2: Normalized controlled ablation of the guide architecture on 2D Random World. Nodes Ratio and Iterations Ratio denote the average number of nodes and iterations required to find the initial feasible path, normalized by the PointNet++ guide baseline. Lower values indicate better initial-solution search efficiency.
Evaluation setting and metric
NIRRT*
CMP-IRRT*
Table 1 protocol: success rate
100.0%
100.0%
Table 1 protocol: initial-solution time (s)
1.813±0.770
1.380±0.650
Table 1 protocol: final path cost after 18 s of post-initial optimization
207.27±26.79
204.11±25.73
2000-expansion budget: success rate
99.8%
100.0%
2000-expansion budget: iterations to first solution
151.71±106.24
146.61±99.87
2000-expansion budget: terminal path cost
207.06±26.62
203.87±25.29
Table 3: Paired repeated results on the 150-map Random World test set. Continuous metrics are reported as map-level mean and standard deviation, while success rates are computed over individual runs. Path cost is measured in map units. Lower is better except for success rate.
Figure 4: Traversability-aware planning results. (a) Visualization on four maps with low-height obstacles that can be traversed. Green obstacles denote obstacles below the threshold Hα . (b) Path shortening after applying height-conditioned traversal decisions.
Figure 5: Real-robot demonstration on a Unitree Go2 robot. The robot executes planned motions that include both obstacle bypassing and traversal of low obstacles.
Configuration
Task completion
Executed path, matched n=10
Intervention/safety
Binary-CMP
10/10
8.56±0.11 m
0 interventions
Oracle-height CMP
10/10
7.25±0.09 m
0 interventions
Full predicted- height CMP
18/20
7.71±0.13 m
2/20 trials required intervention; 0 falls; 0 body collisions
Table 4: Controlled Unitree Go2 trials. Task completion uses all trials for each configuration, whereas executed-path length is reported as mean and standard deviation on the common matched subset of ten trials.
Quadruped robots are increasingly expected to navigate through narrow passages, cluttered indoor scenes, and large-scale 3D unstructured environments. Existing local planners commonly approximate the robot using isotropic geometric inflation or rely on planar and elevation-map representations, leading to conservative motion in tight spaces and limited reasoning about overhanging structures. This letter presents SCAN-Planner, a spatial collision-aware local planning framework for long-range quadruped navigation. A yaw-aware twin-cylinder footprint is used to model the elongated robot body, enabling whole-body collision evaluation through sparse queries in an inflated 3D occupancy map. We further introduce a projected A* search that generates collision-free guidance on an interpolated ground-following surface, with z-gradient suppression to avoid obstacles horizontally while maintaining vertical stability. For large-scale deployment, a robot-centric sliding map with boundary fallback provides high-resolution local collision checking and recovery from local dead ends. Simulation and real-world experiments demonstrate that SCAN-Planner generates safe, smooth, and efficient trajectories in dense clutter, 3D unstructured scenes, stair traversal, and long-range navigation tasks.
Navigating quadruped robots in unstructured 3D environments poses significant challenges, requiring goal-directed motion, effective exploration to escape from local minima, and posture adaptation to traverse narrow, height-constrained spaces. Conventional approaches employ a sequential mapping-planning pipeline but suffer from accumulated perception errors and high computational overhead, restricting their applicability on resource-constrained platforms. To address these challenges, we propose Hierarchical Posture-Adaptive Navigation (HiPAN), a framework that operates directly on onboard depth images at deployment. HiPAN adopts a hierarchical design: a high-level policy generates strategic navigation commands (planar velocity and body posture), which are executed by a low-level, posture-adaptive locomotion controller. To mitigate myopic behaviors and facilitate long-horizon navigation, we introduce Path-Guided Curriculum Learning, which progressively extends the navigation horizon from reactive obstacle avoidance to strategic navigation. In simulation, HiPAN achieves higher navigation success rates and greater path efficiency than classical reactive planners and end-to-end baselines, while real-world experiments further validate its applicability across diverse, unstructured 3D environments.
We present Riemannian Informed Trees (RIT*), a planning framework that replaces Euclidean primitives in batch-informed search with their Riemannian counterparts. RIT* constructs a tighter, cost-consistent informed set, performs a nearest-neighbour search under an anisotropic distance metric, and evaluates edge costs efficiently via a cascading scheme. We further introduce a Collision-Adaptive Metric Refinement (CARM), which learns an obstacle-proximity cost field online from collision feedback, reducing the reliance on prior metric design in practical settings. Experiments across environments from 2-D to 14-D show that RIT* is competitive in low-dimensional and spatially constant-metric settings and produces substantially lower-cost solutions when the metric varies spatially in high-dimensional configuration spaces. Performance gains scale with anisotropy and dimension, reaching up to 13.0% improvement in median initial cost over BIT* in the 3-D anisotropic benchmark, up to 9.0% in median final cost over BIT* in 6-DOF manipulation, and 24.8-63.5% in a 14-DOF bimanual planning problem, where Euclidean-informed baselines degrade. Videos and code can be found here: https://muhayyuddin.github.io/ritstar/
Muhayy Ud Din, Ahmed Nadar, Jan Rosell +1
Center for Autonomous Robotic Systems, Khalifa University (KU-CARS) · Institute of Industrial and Control Engineering, Universitat Politècnica de Catalunya, Barcelona, Spain