Remote Control of Land-Air Amphibious Vehicle with Multimodal Human-Machine Collaborative Obstacle Avoidance

To address the challenge of simultaneously maintaining operator intent, ensuring safe obstacle avoidance, and balancing multiple motion modes in remote control of amphibious unmanned platforms in complex environments, a multimodal human-machine collaborative path planning and trajectory tracking method was proposed. Based on ground-based tracked and airborne ducted flight dynamics models, motion primitives were used to structurally express operator input and unified optimization and tracking of the reference path was achieved within a Model Predictive Contour Control (MPCC) framework. The method was validated using the Gazebo-PX4 co-simulation platform in a complex environment containing narrow passages and windows. Results show that the platform achieved a minimum safe clearance of 0.32 m in ground mode and stably passed through a 0.8 m × 0.8 m narrow window with a minimum safe clearance of 0.24 m in flight mode. Furthermore, both motion modes achieved short task completion times and smooth, controllable speed output, validating the effectiveness of the proposed method in complex and constrained environments.