Hierarchical Collision Avoidance for Adaptive-Speed Multirotor Teleoperation

TL;DR

This paper tackles safe, high-rate teleoperation of multirotors in caves with varying clutter, aiming to reduce operator cognitive load. It presents a hierarchical collision avoidance framework that jointly modulates maximum speed and local map resolution at 10 Hz. The key contributions are a variable-resolution local occupancy map, adaptive velocity bounds tied to map alpha, and a three-level collision-checking loop that re-plans across up to three map levels. The results from simulation and real-world cave experiments show faster task completion and reliable traversal through open, cluttered, and narrow regions without requiring the operator to set a maximum speed. This work advances CSAR teleoperation by enabling automatic adaptation to unknown, structured environments while maintaining safety.

Abstract

This paper improves safe motion primitives-based teleoperation of a multirotor by developing a hierarchical collision avoidance method that modulates maximum speed based on environment complexity and perceptual constraints. Safe speed modulation is challenging in environments that exhibit varying clutter. Existing methods fix maximum speed and map resolution, which prevents vehicles from accessing tight spaces and places the cognitive load for changing speed on the operator. We address these gaps by proposing a high-rate (10 Hz) teleoperation approach that modulates the maximum vehicle speed through hierarchical collision checking. The hierarchical collision checker simultaneously adapts the local map's voxel size and maximum vehicle speed to ensure motion planning safety. The proposed methodology is evaluated in simulation and real-world experiments and compared to a non-adaptive motion primitives-based teleoperation approach. The results demonstrate the advantages of the proposed teleoperation approach both in time taken and the ability to complete the task without requiring the user to specify a maximum vehicle speed.

Figures (8)

• Figure 1: Tele-operated multirotor adapts the motion planning speed and local map resolution to \ref{['sfig:cave_entrance']} enter a cave and \ref{['sfig:cave_passage']} traverse a tight passage inside. \ref{['sfig:cave_entrance2']} illustrates the surroundings near the cave entrance, which is embedded in a sloping hillside. A video of this experiment can be found at https://youtu.be/VjyoPVXT8WY.
• Figure 2: Information flow diagram for the technical approach.
• Figure 3: Bounding box extents for a scenario where the robot traverses a window. The teleoperator gives maximum joystick input in the forward direction for these three figures. \ref{['sfig:bbox1']} When the robot is far from the window, the bounding box extents and local occupancy map are large because the voxel size is also large. \ref{['sfig:bbox2']} As the multirotor gets closer to the window, the voxel size decreases and so does the bounding box extent because the number of voxels in the map stays the same. \ref{['sfig:bbox3']} After exiting the window, the bounding box expands to the original size. Note that the change in bounding box extents is achieved by varying the voxel size and keeping the number of voxels constant.
• Figure 4: The robot used in the field experiments is equipped with a forward-facing Intel Realsense D455, downward-facing mvBluefox global shutter color camera, and Pixracer flight controller.
• Figure 5: Performance comparison for the simulated window teleoperation task. \ref{['sfig:window_scenario']} depicts the initial conditions for the task. A multirotor hovers at a distance of 10m from a window of dimensions $0.9m \times 0.9m$. The operator controls the multirotor via the joystick shown in \ref{['sfig:joystick']}. The operator intends to go forward at the highest speed possible. \ref{['sfig:speed']} and \ref{['sfig:map_res']} show the variation of the forward speed and the local map voxel size as a function of the distance from the window for the three teleoperation approaches.