STITCHER: Constrained Trajectory Planning in Known Environments with Real-Time Motion Primitive Search
🤖 AI Summary
Addressing the challenge of ensuring real-time performance, dynamic feasibility, and safety simultaneously for high-speed autonomous navigation in known environments, this paper proposes STITCHER—a real-time trajectory planning framework that avoids numerical optimization. STITCHER integrates graph-search-driven motion primitive matching with short-trajectory-segment stitching, enabling millisecond-scale handling of non-convex state and actuator constraints (e.g., tilt angle, motor thrust limits). By leveraging a precomputed trajectory library, efficient online querying, and lightweight dynamic feasibility verification coupled with collision checking, STITCHER generates collision-free trajectories over long horizons (<5 ms) in complex 50 m × 50 m environments. Hardware experiments on a quadrotor demonstrate robust real-time trajectory tracking. Quantitative evaluation shows STITCHER significantly outperforms two state-of-the-art optimization-based planners in both computational efficiency and tracking accuracy under dynamic constraints.
Application Category
📝 Abstract
Autonomous high-speed navigation through large, complex environments requires real-time generation of agile trajectories that are dynamically feasible, collision-free, and satisfy state or actuator constraints. Modern trajectory planning techniques primarily use numerical optimization, as they enable the systematic computation of high-quality, expressive trajectories that satisfy various constraints. However, stringent requirements on computation time and the risk of numerical instability can limit the use of optimization-based planners in safety-critical scenarios. This work presents an optimization-free planning framework called STITCHER that stitches short trajectory segments together with graph search to compute long-range, expressive, and near-optimal trajectories in real-time. STITCHER outperforms modern optimization-based planners through our innovative planning architecture and several algorithmic developments that make real-time planning possible. Extensive simulation testing is performed to analyze the algorithmic components that make up STITCHER, along with a thorough comparison with two state-of-the-art optimization planners. Simulation tests show that safe trajectories can be created within a few milliseconds for paths that span the entirety of two 50 m x 50 m environments. Hardware tests with a custom quadrotor verify that STITCHER can produce trackable paths in real-time while respecting nonconvex constraints, such as limits on tilt angle and motor forces, which are otherwise hard to include in optimization-based planners.
Problem
Research questions and friction points this paper is trying to address.
Real-time agile trajectory planning for autonomous navigation
Ensuring dynamic feasibility and collision avoidance in complex environments
Overcoming computational limitations of optimization-based planners
Innovation
Methods, ideas, or system contributions that make the work stand out.
Stitches trajectory segments with graph search
Replaces optimization with real-time motion primitive search
Handles nonconvex constraints like tilt angle limits
🔎 Similar Papers
No similar papers found.
