Reactive Whole-Body Control of Mobile Manipulators for Dynamic Target Tracking Via Adaptive-Predictive Visual Servoing
Andrea Monguzzi, Giuseppe Alfonso, Navvab Kashiri
AI summary
Problem
Mobile manipulators struggle to track dynamic targets in unbounded workspaces because coordinating the base and arm while keeping the target visible requires fast, robust adaptation to unknown trajectories and estimation errors.
Approach
The authors develop a quadratic program-based whole-body controller that dynamically shifts priorities between field-of-view centering and pose alignment, integrated with an adaptive-predictive visual servoing scheme that uses adaptive gains and Kalman-filtered velocity estimates to generate smooth Cartesian references.
Key results
- Adaptive QP weighting strategy that prioritizes camera field-of-view centering at long range and precise pose alignment at close range
- Adaptive-predictive visual servoing with variable gains to prevent base vibrations and dynamic feedforward terms using Kalman-filtered target velocity
- Real-world experimental validation on a holonomic mobile manipulator demonstrating superior tracking performance over state-of-the-art whole-body controllers
Why it matters
Provides a robust, real-time control solution for mobile manipulators operating in unstructured environments, advancing applications in logistics, inspection, and autonomous manipulation.
Abstract
This paper addresses the challenging problem of enabling a mobile manipulator with an eye-in-hand camera to track dynamic targets with time-varying positions and orientations in an unbounded workspace. Specifically, we pro- pose an optimization-based whole-body control framework for dynamic target tracking. The framework enables the mobile manipulator to maintain the target within the camera’s field of view while reaching the desired pose, by dynamically regulating the priorities of the optimization constraints and objectives according to the task execution state. Moreover, we present an adaptive-predictive position-based visual servoing strategy to generate the Cartesian references sent to the controller. To enhance the tracking performance, we introduce (1) adaptive gains to avoid abrupt motions and the resulting vibrations while preserving final precision; (2) dynamic addition of a feedforward term incorporating a velocity estimate of the target using a Kalman Filter. The proposed approach is validated on a real robotic setup, as compared to a state-of-the-art approach, demonstrating superior performance in dynamic target tracking.