Real-Time Planning with Multi-Fidelity Models for Agile Flights in Unknown Environments

Abstract

Autonomous navigation through unknown environments is a challenging task that entails real-time localization, perception, planning, and control. UAVs with this capability have begun to emerge in the literature with advances in lightweight sensing and computing. Although the planning methodologies vary from platform to platform, many algorithms adopt a hierarchical planning architecture where a slow, low-fidelity global planner guides a fast, high-fidelity local planner. However, in unknown environments, this approach can lead to erratic or unstable behavior due to the interaction between the global planner, whose solution is changing constantly, and the local planner; a consequence of not capturing higher-order dynamics in the global plan. This work proposes a planning framework in which multi-fidelity models are used to reduce the discrepancy between the local and global planner. Our approach uses high-, medium-, and low-fidelity models to compose a path that captures higher-order dynamics while remaining computationally tractable. In addition, we address the interaction between a fast planner and a slower mapper by considering the sensor data not yet fused into the map during the collision check. This novel mapping and planning framework for agile flights is validated in simulation and hardware experiments, showing replanning times of 5-40 ms in cluttered environments.

FASTER: Fast and Safe Trajectory Planner for Flights in Unknown Environments

“…The first six methods are described deeper in (Oleynikova et al 2018), while (Tordesillas et al 2019a) is our previous proposed algorithm. The results are shown in Table 2, which highlights that FASTER achieves a 8 − 51% improvement in the total distance flown.…”

“…An alternative strategy is to create and plan trajectories in a map of the environment built using a history of perception data. Within this second category, in some works (Schouwenaars et al 2002;Tordesillas et al 2019a;Oleynikova et al 2018), the local planner only optimizes inside F, which guarantees safety if the local planner has a final stop condition. However, limiting the planner to operating in F and enforcing a terminal stopping condition can lead to conservative, slow trajectories (especially when much of the world is unknown).…”

“…The typical way to solve the speed versus safety tradeoff is to ensure safety by planning only in F, and then impose a final stop condition near the global plan. This can be achieved by either generating motion primitives that do not intersect O ∪ U (Mueller et al 2015;Lopez and How 2017a,b;Tordesillas et al 2019a), or by constructing a convex representation of F to be used in an optimization (Deits and Tedrake 2015;Liu et al 2017b;Preiss et al 2017). The main limitation of these works is that safety is guaranteed at the expense of higher speeds, especially in scenarios where F is small compared to O ∪ U.…”

“…Velocity profile in the office simulation. On the left the results of(Tordesillas et al 2019a) and on the right FASTER.…”

“…Comparison between(Tordesillas et al 2019a) and FASTER of flight times in the forest simulation. Results are for 10 random forests.…”