Robot Path Planning That Smoothly Dodges Moving Objects

A team of researchers has blended two proven techniques—an enhanced Artificial Potential Field (APF) and a refined Bidirectional Rapidly-exploring Random Tree (Bi‑RRT)—to create a cutting‑edge navigation system for collaborative robots.

Key Innovations

• Cost‑Sensitive Waypoint Selection
  The Bi‑RRT search now incorporates a cost metric that prefers small adjustments to the robot’s joint angles. This strategy yields smoother paths and eliminates unnecessary stops.

• Joint‑Space Potential Field
  The APF is reimagined to operate in joint space:
• Attractive force pulls the robot toward its goal.
• Redesigned repulsive force pushes it away from any moving obstacle that could collide with its links.

Results

• Smooth, Continuous Motion
  Simulations demonstrate that robots can navigate around both stationary and moving barriers while keeping all joints turning smoothly.

• Real‑Time Obstacle Avoidance
  The combined method shows reliable performance in real‑time scenarios, making it suitable for dynamic factory floors and collaborative work environments.

The hybrid approach promises safer, more efficient collaboration between robots and humans in complex settings.