UAV and ground robot cooperate without global position using active camera control
Robust Active-Perception Control for Global-State-Free Aerial-Ground Cooperation
Robotics
Summary
Getting drones and ground robots to work together needs them to know where each other are all the time. Instead of using expensive GPS or maps, the authors designed a system where the drone's camera is on a movable mount, so it can keep the robot in sight even when the drone is tilting or moving. They also teach the system to guess how the ground robot will move next, helping the drone and its camera move smoothly to keep watching it. Their tests show this method works well both in computer simulations and real-world experiments.
What this means in practice
- •For drone operators: Maintain reliable drone tracking of ground robots without GPS or external localization by using active camera control and predicted robot motion.
- •For robotics integrators: Build collaborative aerial-ground robot teams that operate effectively in GPS-denied environments using vision-based relative navigation and predictive control.
Authors
Mingxuan Zhang, Jiajun Yu, Baozhe Zhang, Pengxiang Zhou, Wentao Liu, Fei Gao, Chao Xu, Yanjun Cao
Abstract
Aerial-ground cooperation requires real-time UAV--UGV relative-state information. Instead of maintaining global estimates for both robots, direct control in a UGV-attached non-inertial frame avoids reliance on global localization. Vision-based relative pose estimation with a passive marker offers a low-cost and effective solution. However, a fixed camera may lose sight of the moving UGV when the required UAV attitude conflicts with the field-of-view (FOV) constraint. To address this, we propose COPA, a robust active-perception framework for global-state-free aerial-ground cooperation. We use a single-axis gimbal to decouple the camera optical axis from the UAV pitch attitude. We derive an active-perception model that relates UAV motion, gimbal angle, and UGV motion to the target image-plane state.A Temporal Convolutional Network (TCN) predicts short-horizon UGV acceleration and angular velocity from recent motion history without global-state measurements. The model predictive control (MPC) uses these predictions to jointly optimize UAV and gimbal control. Simulations show that COPA maintains continuous target visibility, while ablation studies confirm that the TCN reduces peak errors during UGV motion transitions. Real-world experiments with UGV accelerations up to 3m/s^2 and yaw rates up to 1.0rad/s demonstrate robust tracking.