A cell-decomposition based path planner for 3D navigation in constrained workspaces
For robotic path planning in constrained 3D environments, this work offers a memory-efficient optimization method that balances solution quality and computational cost.
The paper proposes a cell decomposition algorithm for binary occupancy grids that ensures mutual complete visibility, enabling a simplified framework for path feasibility verification. The KSP-SOCP method achieves time performance comparable to MISOCP while using less memory, tested in 9 city-like workspaces.
This paper proposes a cell decomposition algorithm for binary occupancy grids that ensures mutual complete visibility from each cell to at least one adjacent cell. This decomposition establishes a simplified framework for verifying path feasibility that can be easily embedded in optimization problems. To illustrate its utility, we formulate both second-order cone programs (SOCP) and their mixed-integer variant (MISOCP) within the proposed framework. Furthermore, we propose the KSP-SOCP method, which combines Yen's k-shortest path algorithm with the SOCP, achieving improved solutions compared to a standard SOCP approach while avoiding the computational burden of MISOCP. The cell decomposition algorithm, KSP-SOCP, and MISOCP approaches were evaluated in 9 city-like workspaces. The decomposition efficiently partitioned each map, enabling both optimization methods to compute feasible paths. The proposed KSP-SOCP achieved time performance comparable to the MISOCP while requiring less memory, making it highly suitable for large-scale problems.