FASTER: Fast and Safe Trajectory Planner for Flights in Unknown Environments (1903.03558v3)

Abstract: High-speed trajectory planning through unknown environments requires algorithmic techniques that enable fast reaction times while maintaining safety as new information about the operating environment is obtained. The requirement of computational tractability typically leads to optimization problems that do not include the obstacle constraints (collision checks are done on the solutions) or use a convex decomposition of the free space and then impose an ad-hoc time allocation scheme for each interval of the trajectory. Moreover, safety guarantees are usually obtained by having a local planner that plans a trajectory with a final "stop" condition in the free-known space. However, these two decisions typically lead to slow and conservative trajectories. We propose FASTER (Fast and Safe Trajectory Planner) to overcome these issues. FASTER obtains high-speed trajectories by enabling the local planner to optimize in both the free-known and unknown spaces. Safety guarantees are ensured by always having a feasible, safe back-up trajectory in the free-known space at the start of each replanning step. Furthermore, we present a Mixed Integer Quadratic Program formulation in which the solver can choose the trajectory interval allocation, and where a time allocation heuristic is computed efficiently using the result of the previous replanning iteration. This proposed algorithm is tested extensively both in simulation and in real hardware, showing agile flights in unknown cluttered environments with velocities up to 3.6 m/s.

• The paper introduces a dual trajectory strategy that optimizes paths in both known and unknown spaces while ensuring a safe fallback route.
• It employs Mixed Integer Quadratic Programming to dynamically allocate time intervals, achieving up to 52% improvement in time efficiency.
• Extensive simulations and hardware trials demonstrate enhanced UAV navigation with up to 51% improvement in path efficiency and effective obstacle avoidance.

Overview of FASTER: A Trajectory Planner for UAVs in Unknown Environments

The paper introduces FASTER (Fast and Safe Trajectory Planner), an advanced trajectory planning algorithm tailored for unmanned aerial vehicles (UAVs) navigating high-speed flights through unknown environments. The objective of FASTER is to address the trade-off between computational efficiency and the trajectory's speed and safety. Traditional approaches often lean towards either optimizing speed without comprehensive safety guarantees or ensuring safety at the cost of speed. FASTER innovatively combines local and global planning methods to ensure fast and safe navigation.

Key Contributions

Optimization in Combined Spaces: Unlike conventional methods, FASTER allows the local planner to optimize trajectories in both free-known and unknown spaces (denoted as $\mathcal{F}$ and $\mathcal{U}$, respectively), while guaranteeing safety by maintaining a feasible backup trajectory within the free-known space.

Mixed Integer Quadratic Programming (MIQP): FASTER uses MIQP for trajectory planning, where the solver can dynamically choose the allocation of trajectory intervals. The method optimizes time allocation using a heuristic based on prior solutions, allowing it to adapt more flexibly compared to fixed interval strategies.

Two Trajectory Approach: The framework distinguishes between "Whole Trajectory," which may traverse unknown spaces, and "Safe Trajectory," confined to known safe spaces, always ensuring there exists a viable path back to safe conditions. This dual trajectory approach enhances both exploratory capability and safety assurance.

Simulation and Real-World Validation: The algorithm's efficacy is validated through extensive simulations and hardware trials, demonstrating agile and efficient flight capabilities in cluttered environments at high velocities (up to 3.6 m/s).

Experimental Insights

The authors benchmarked FASTER against several established strategies, including random goal selection, RRT* variations, and other local exploration planners. The results underscore FASTER's significant improvements in path efficiency and completion times—demonstrating up to 51% improvement in navigation distance and 52% in time efficiency compared to predecessor methods.

The practical viability of FASTER was further illustrated through hardware experiments involving real-time navigation in complex indoor environments, showcasing effective obstacle avoidance and trajectory adaptation.

Implications and Future Prospects

FASTER represents a substantial stride in autonomous UAV navigation technologies, primarily for applications requiring rapid decision-making and adaptive trajectory adjustments in unexplored settings. The proposed MIQP-based optimization contributes to broader prospects for enhancing real-time planning algorithms, potentially applicable in other robotic motion planning scenarios with dynamic constraints.

Future developments could explore integrating learning-based models to further improve the dynamic adaptation of flight paths based on incoming environmental data. Additionally, expanding this methodology to multi-agent systems could enhance cooperative navigation strategies, enriching UAV applications in complex collaborative tasks.

In conclusion, FASTER advances the capabilities of UAVs to operate autonomously in uncertain environments, highlighting the importance of balancing speed and safety in autonomous systems—a critical advancement for practical deployments in numerous exploration and surveillance tasks.