Elastic band

• Also called: elastic bands, elastic band method
• Category: Planning & manipulation › Motion planning
• Tags: motion-planning, local-planning, obstacle-avoidance, path-deformation, mobile-robots

A local path deformation method that treats a planned path as a stretchable band, pulled tight by internal forces and pushed away from obstacles by external ones, so it adapts in real time.

Definition

An elastic band is a method for deforming a path in real time, introduced by Quinlan and Khatib in 1993 to connect global path planning with reactive control. A collision-free path from a global planner is represented as a chain of points, or of bubbles, regions of free space around each point sized by the distance to the nearest obstacle. The band is then treated as a physical object. Internal contraction forces, like springs between neighboring points, pull it tight so that it becomes shorter and smoother. External forces, derived from a repulsive potential around obstacles, push it away from them. Each control cycle the band relaxes toward a balance of these forces, so when an obstacle moves or a new one appears, the path bends around it while keeping its start and goal.

The bubbles keep the band connected: as long as neighboring bubbles overlap, the straight segments between them are collision-free, and points are added or removed as the band stretches or shrinks. Because only the path is deformed, not recomputed, the method is fast enough to run inside a control loop. Later work extended the idea to elastic strips for mobile manipulators, which deform a sequence of whole-robot configurations in the workspace, and to the Timed Elastic Band, which adds time intervals between poses and solves a sparse optimization with velocity, acceleration and kinematic constraints. The Timed Elastic Band is available as a local planner in ROS.

The elastic band differs from a potential field method, which moves the robot directly along the gradient of an attractive and repulsive potential: here the forces act on an entire path that already connects the start to the goal, so the robot does not get trapped in the local minima that can stop a pure potential-field controller. It remains a local method. A band can only deform continuously, so it cannot jump to a different way around an obstacle, such as passing on the left instead of the right; if an obstacle cuts the band or the corridor closes, a global planner must supply a new path. Some variants keep several candidate bands in different homotopy classes for this reason.

Why it matters

Robots in changing environments need paths that react to moving obstacles within a control cycle without discarding the global plan. Elastic bands and their descendants are a standard way of providing that reactive layer for mobile robots and manipulators, and the force-based path model is also used to generate smooth, obstacle-aware trajectories for other purposes.

Key applications

• Local planning and obstacle avoidance for mobile robots
• Real-time path deformation around moving people and objects
• Reactive motion for mobile manipulators through elastic strips
• Smoothing and shortening paths from global planners
• Generating obstacle-aware trajectories for training data and demonstrations

In practice

A warehouse mobile robot receives a global path along an aisle from a grid-based planner. Its local planner represents the next few meters of the path as a Timed Elastic Band. When a person steps into the aisle, the lidar cost map registers a new obstacle, the band's points are pushed away from it, and the optimization bends the path around the person while respecting the robot's velocity and acceleration limits and its minimum turning radius. The robot follows the deformed path, and once the person has passed, the contraction terms pull the band back toward the original line. If the aisle is blocked completely, the band cannot be deformed into a feasible path, and the navigation stack asks the global planner for a new route.

Common misconceptions

• An elastic band is a global planner — it deforms an existing path and needs a global planner to supply the initial route and to replan when the band breaks
• Elastic bands are just potential fields — potential fields move the robot directly, while an elastic band applies forces to a whole path that already reaches the goal, which avoids the local minima that trap a pure potential-field controller
• An elastic band finds the shortest path — it converges to a locally short, smooth path that passes obstacles on the same side as the initial path, which may not be the globally shortest

Who should know this

• Mobile robot navigation engineers
• Motion planning researchers working on reactive planning
• Engineers configuring ROS navigation local planners
• Robotics students studying planning and control

Related technologies

• Timed Elastic Band
• Elastic Strips
• ROS Navigation Stack
• Artificial Potential Fields
• Dynamic Window Approach

Related terms

• Potential Field
• Obstacle Avoidance
• Trajectory Planning
• Motion Planning
• Probabilistic Roadmap

Related projects

• OMPL: Sampling-based motion planning library powering MoveIt and countless robotics applications
• MoveIt 2: Motion planning framework for ROS 2 robot arms, with kinematics and collision checking
• PyRoki: Modular Python toolkit for kinematic optimization across CPU, GPU, and TPU via JAX
• mink: Python inverse kinematics library built on MuJoCo supporting 20+ robot platforms
• Ruckig: Real-time jerk-constrained time-optimal trajectory generation for robots and machines
• NeuPAN: Real-time end-to-end planner that maps raw lidar points directly to robot actions