Kinodynamic Motion Planning for a Team of Multirotors Transporting a Cable-Suspended Payload in Cluttered Environments

TL;DR

It is demonstrated in a software-in-the-loop simulation and real flight experiments that there is a significant benefit in kinodynamic motion planning for such payload transport systems with respect to payload tracking error and energy consumption compared to the standard methods of planning for the payload alone.

Abstract

We propose a motion planner for cable-driven payload transportation using multiple unmanned aerial vehicles (UAVs) in an environment cluttered with obstacles. Our planner is kinodynamic, i.e., it considers the full dynamics model of the transporting system including actuation constraints. Due to the high dimensionality of the planning problem, we use a hierarchical approach where we first solve the geometric motion planning using a sampling-based method with a novel sampler, followed by constrained trajectory optimization that considers the full dynamics of the system. Both planning stages consider inter-robot and robot/obstacle collisions. We demonstrate in a software-in-the-loop simulation and real flight experiments that there is a significant benefit in kinodynamic motion planning for such payload transport systems with respect to payload tracking error and energy consumption compared to the standard methods of planning for the payload alone. Notably, we observe a significantly higher success rate in scenarios where the team formation changes are needed to move through tight spaces.

Figures (5)

• Figure 1: Highlighted in the red box is our full kinodynamic motion planning algorithm. The geometric output of a sampling-based motion planner is used to initialize an optimizer, which generates the full feasible reference trajectory of the payload $\mathbf{p}_{0_r}$ and the cable states $\mathbf{q}_{i_r}$. One can also use our sampling-based motion planner and compute the first order derivatives of the geometric states $(\dot{\mathbf{p}}_{0_r},\boldsymbol{\omega}_{i_r})$ to provide a reference trajectory. Each reference trajectory is then tracked by our controller wahba2023efficient.
• Figure 2: Validation scenarios. From left to right: empty (3 robots), forest (4 robots), window (5 robots). Green UAVs (left on each picture) show the initial state, red UAVs the desired state. The red line represents the reference trajectory, and the white line is the tracked trajectory by our controller. For window, the obstacles necessitate a formation change to pass through a narrow passage.
• Figure 3: Left: Computational effort in seconds for the optimization to compute a solution in the forest environment over different numbers of robots. Right: Solution quality (in terms of energy) when sequentially solving the kinodynamic optimization over multiple iterations.
• Figure 4: Examples for the sampling-based geometric planner using different environment and sampling strategies. The plot shows the mean and standard deviation (shaded) for cost convergence over runtime (log-scale), if the success rate is over $50 \%$ (50 trials).
• Figure 5: Real flights validation scenarios. left: forest (3 robots), right: window (2 robots). The payload in both scenarios is modeled as a 10g point mass.