cs.ROOct 30, 2025

ART-TEB: Adaptive Trajectory Planning for Mobile Robots in Cluttered Environments

Authors: Hahjin Lee, Young J. Kim

Organizations: Department of Computer Science and Engineering at Ewha Womans University in Korea

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Trajectory planning for mobile robots remains a major challenge, particularly in cluttered environments, where existing planning algorithms often fail or produce suboptimal paths. To address this issue, we propose an adaptive trajectory refinement algorithm, ART-TEB, comprising two main stages. First, to ensure safety at the path-segment level, a segment-wise conservative collision test is applied, recursively subdividing risky path segments until the collision risk is eliminated. Second, to guarantee pose-level safety, pose correction based on separation direction and line search is applied, ensuring that each pose in the trajectory is collision-free and maximally clear from obstacles. Simulation results demonstrate that the ART-TEB achieves up to 3.12x higher success rates and up to 23.7x faster average planning times than state-of-the-art approaches. Furthermore, real-world experiments confirm that the robot can safely pass through highly constrained environments while maintaining rapid planning performance.

Figures & tables

Fig. 1 : Limitations of TEB (a) o3\mathbf{o}_{3} is the closest to x3\mathbf{x}_{3} . The trajectory can penetrate into obstacles since (b) ∥x3−o3∥=∥x3′−o3∥\lVert\mathbf{x}_{3}-\mathbf{o}_{3}\rVert=\lVert\mathbf{x}_{3}^{\prime}-\mathbf{o}_{3}\rVert and (c) x3′\mathbf{x}_{3}^{\prime} yields a shorter trajectory than x3\mathbf{x}_{3} . (d) Waypoints are sparsely placed near obstacles, leading to collisions. (e) The original temporal resolution in the trajectory. (f) Reducing the resolution can cause collisions.
Fig. 2 : Algorithm Overview. (a) An initial plan from TEB hyper-graph optimization. (b–d) Iterative collision detection, pose correction, and orientation updates ensure all poses are collision-free. (e-f) Segment-wise CCD subdivides risky segments, with new poses refined through (b-d). (g) A collision-free trajectory is obtained. This trajectory is then fed back to the hyper-graph optimization, forming an iterative planning pipeline.
Fig. 3 : Separation direction v\mathbf{v} determined when (a) the robot’s position is located at x\mathbf{x} inside an obstacle (gray grids), (b) on the obstacle boundary (red grids), (c) outside the obstacle but still close to the obstacle.
Fig. 4 : BARN dataset with benchmark number. The Jackal robot navigates through 300 different environments with varying obstacle densities, represented by red cylinders.
Fig. 5 : Real-world experimental scenarios designed to evaluate performance. Scenario 1 requires the robot to pass through a narrow doorway. Scenario 2 involves navigating past a cart and a tight passage in a confined corridor. Scenario 3 features a tight corner in an office environment that leads to the goal. Scenario 4 presents a challenging test environment with sharp turns formed by barriers. The red circle labeled “G” denotes the goal position, while the arrows illustrate the robot’s navigation path.
MethodSuccess Rate (%)Avg. Planning Time (ms)Max. Planning Time (ms)Traversal Time (s)
DWA54.6711.7614.6930.76
MPC31.7853.65153.9136.95
TEB58.75 / 83.785.27 / 8.9912.78 / 19.6017.68 / 16.79
egoTEB86.83 / 60.673.25 / 4.737.33 / 17.5115.75 / 15.18
ART-TEB99.25 / 95.112.26 / 2.375.15 / 5.8815.93 / 15.69
ART-TEB w/o CCD82.56 / 52.331.97 / 1.985.18 / 4.7916.17 / 15.1
TABLE I : Quantitative performance comparison in simulation. Avg. Planning Time is averaged over all planning steps, whereas Max. Planning Time is obtained by averaging the maximum planning time per trial. Traversal Time denotes the time required for the robot to travel from start to goal. For TEB variants, -/- in each cell indicates results corresponding to planning horizons of 1.0s1.0s and 3.0s3.0s , respectively.
ScenarioTask SuccessProgress Rate (%)Avg. Planning Time (ms)Max Planning Time (ms)Avg. Traversal Time (s)
TEBegoTEBART-TEBTEBegoTEBART-TEBTEBegoTEBART-TEBTEBegoTEBART-TEBTEBegoTEBART-TEB
Scen. 10/51/55/541.7747.76100.0-3.813.06-13.618.96-10.0310.03
Scen. 20/50/55/549.5849.5100.0--3.87--10.09--20.15
Scen. 30/50/54/536.2737.8595.49--6.71--10.58--18.94
Scen. 40/50/55/514.9751.81100.0--5.78--12.94--22.88
TABLE II : Performance in real-world environments. ART-TEB achieved the highest task success across all scenarios, with lower planning times compared to the baselines. “ −- ” denotes cases in which the method failed to succeed.
Fig. 6 : Real-world BARN Challenge environment. The benchmark features narrow, cluttered passages constructed with cardboard boxes.

Explore similar work

CardsList

  1. Reactive Planning based Control for Mobile Robots in Obstacle-Cluttered Environments

    May 14, 2026Li Tan, Junlin Xiong, Yan Wang +1Obstacle AvoidanceMobile Robots

    Department of Automation, University of Science and Technology of China, Hefei, 230031, China. · School of Intelligence Science and Engineering, Harbin Institute of Technology Shenzhen, Shenzhen 515100, China · School of Control Science and Engineering, Dalian University of Technology, Dalian 116024, China.

  2. STITCHER: Constrained Trajectory Planning in Complex Environments with Real-Time Motion Primitive Search

    Oct 16, 2025Helene J. Levy, Brett T. LopezClassical PlanningQuadrotor

    VECTR Laboratory, University of California, Los Angeles, Los Angeles, CA, USA

  3. RoboNav-Arm: Agentic AI-Driven Navigation and Obstacle Avoidance for Robotic Manipulator in Cluttered Environments

    Jun 25, 2026Aachal Sharma, Narendra Kumar DharObstacle AvoidanceRobotic Manipulation

    Centre for Artificial Intelligence and Robotics (CAIR), Indian Institute of Technology Mandi, Himachal Pradesh 175005, India.