IKLink: End-Effector Trajectory Tracking with Minimal Reconfigurations

TL;DR

The paper addresses end_effector trajectory tracking when continuous tracking is infeasible due to kinematic constraints by minimizing joint_space reconfigurations. It introduces IKLink, a graph_based approach that samples diverse IK solutions per waypoint and uses dynamic programming to link these solutions into a trajectory with the fewest reconfigurations. The method demonstrates favorable reconfiguration counts and competitive motion quality in simulations and a real_robot demonstration, and provides an open_source implementation. This work advances practical trajectory tracking for redundant robots in cluttered or constrained environments, reducing time and energy overheads by avoiding unnecessary reconfigurations.

Abstract

Many applications require a robot to accurately track reference end-effector trajectories. Certain trajectories may not be tracked as single, continuous paths due to the robot's kinematic constraints or obstacles elsewhere in the environment. In this situation, it becomes necessary to divide the trajectory into shorter segments. Each such division introduces a reconfiguration, in which the robot deviates from the reference trajectory, repositions itself in configuration space, and then resumes task execution. The occurrence of reconfigurations should be minimized because they increase the time and energy usage. In this paper, we present IKLink, a method for finding joint motions to track reference end-effector trajectories while executing minimal reconfigurations. Our graph-based method generates a diverse set of Inverse Kinematics (IK) solutions for every waypoint on the reference trajectory and utilizes a dynamic programming algorithm to find the globally optimal motion by linking the IK solutions. We demonstrate the effectiveness of IKLink through a simulation experiment and an illustrative demonstration using a physical robot.

Figures (4)

• Figure 1: An illustration of the three approaches described in this paper and evaluated in our experiment.
• Figure 2: Left: A UMAP visualization of 20 robot motions that track an identical end-effector trajectory. Each motion starts with a random configuration and greedily propagates to the subsequent end-effector pose using an optimization-based IK solver. Convergence of motions occurs in regions 1 and 2. Right: We visualize motions A and E using a solid color robot and a translucent robot, respectively. The two motions start from different configurations but converge together over the course of the trajectory. This figure shows that greedy propagation does not consistently yield diverse solutions.
• Figure 3: Our experiment involves four benchmark applications. In Random Trajectory Tracking, the reference end-effector trajectories are visualized in green curves, with coordinate frames attached to show orientations. In the other three benchmarks, the translucent robot and the solid-color robot show the configurations before and after a reconfiguration. These visualizations were generated using Motion Comparator.
• Figure 4: In our experiment, IKLinkconsistently generates motions with equal or fewer reconfigurations compared to the alternative approaches. The $x$-axis of the plot corresponds to a total of 70 end-effector trajectories. The y-axis shows the number of reconfigurations in logarithmic scale.