BPMP-Tracker: A Versatile Aerial Target Tracker Using Bernstein Polynomial Motion Primitives

TL;DR

BPMP-Tracker presents an online aerial target-tracking framework that couples target motion prediction with chasing trajectory planning using Bernstein polynomial motion primitives. The method leverages a sample-check-select strategy and Bernstein properties to perform fast feasibility checks (collision, visibility, and dynamic limits) and advance to a best-priority trajectory. Key contributions include a versatile predictor for multiple targets, a chasing planner with safe-corridor generation, and extensive validations in large-scale simulations and real-world flight tests, including dense dynamic obstacles. The work offers a scalable, real-time solution for multi-target tracking in unstructured and dynamic environments, with potential impact on autonomous surveillance, filming, and search-and-rescue missions.

Abstract

This letter presents a versatile trajectory planning pipeline for aerial tracking. The proposed tracker is capable of handling various chasing settings such as complex unstructured environments, crowded dynamic obstacles and multiple-target following. Among the entire pipeline, we focus on developing a predictor for future target motion and a chasing trajectory planner. For rapid computation, we employ the sample-check-select strategy: modules sample a set of candidate movements, check multiple constraints, and then select the best trajectory. Also, we leverage the properties of Bernstein polynomials for quick calculations. The prediction module predicts the trajectories of the targets, which do not overlap with static and dynamic obstacles. Then the trajectory planner outputs a trajectory, ensuring various conditions such as occlusion and collision avoidance, the visibility of all targets within a camera image and dynamical limits. We fully test the proposed tracker in simulations and hardware experiments under challenging scenarios, including dual-target following, environments with dozens of dynamic obstacles and complex indoor and outdoor spaces.

Figures (9)

• Figure 1: Various target chasing scenarios. A chaser (blue) follows the targets (red, magenta) in unstructured spaces (top) and among dynamic obstacles (green, bottom).
• Figure 2: Overview of the chasing pipeline.
• Figure 3: {Black, red, magenta} and {black, blue, cyan} splines represent {primitives, feasible and the best trajectories} in the prediction and planning, respectively. The pink-shaded region on the left is a generated safe corridor ($\mathcal{S}^{q}$) in point cloud, and grey cylinders on the right are obstacles.
• Figure 4: Simulation results in a city park (left, single target) and a futuristic building (right, dual target). The position histories of the drone (blue) and targets (red, magenta) are plotted.
• Figure 5: Chasing results amidst 69 dynamic-obstacles (black). Blue, red and grey splines represent the position histories of the drone, the target and the obstacles, respectively.

Theorems & Definitions (3)

• proof
• proof
• proof