[Paper Review] A Receding Horizon Multi-Objective Planner for Autonomous Surface Vehicles in Urban Waterways

Tixiao Shan, Wei Wang+3 more|arXiv (Cornell University)|Jul 16, 2020

Robotic Path Planning Algorithms29 references4 citations

TL;DR

This paper proposes a receding horizon multi-objective planner for autonomous surface vehicles (ASVs) in urban waterways, using lexicographic optimization to hierarchically prioritize navigation objectives—such as collision risk, heading consistency, and path length—within a dynamically updated graph of sensed obstacles. The method enables real-time, safe path planning with minimal tuning, validated in both simulation and real-world experiments on a quarter-scale Roboat with moving obstacles.

ABSTRACT

We propose a novel receding horizon planner for an autonomous surface vehicle (ASV) performing path planning in urban waterways. Feasible paths are found by repeatedly generating and searching a graph reflecting the obstacles observed in the sensor field-of-view. We also propose a novel method for multi-objective motion planning over the graph by leveraging the paradigm of lexicographic optimization and applying it to graph search within our receding horizon planner. The competing resources of interest are penalized hierarchically during the search. Higher-ranked resources cause a robot to incur non-negative costs over the paths traveled, which are occasionally zero-valued. The framework is intended to capture problems in which a robot must manage resources such as risk of collision. This leaves freedom for tie-breaking with respect to lower-priority resources; at the bottom of the hierarchy is a strictly positive quantity consumed by the robot, such as distance traveled, energy expended or time elapsed. We conduct experiments in both simulated and real-world environments to validate the proposed planner and demonstrate its capability for enabling ASV navigation in complex environments.

Motivation & Objective

• To develop a real-time path planning framework for autonomous surface vehicles (ASVs) navigating complex urban waterways with dynamic obstacles.
• To address the challenge of multi-objective motion planning where competing objectives like collision risk, heading deviation, and path length must be prioritized hierarchically.
• To enable efficient, parameter-free planning by leveraging lexicographic optimization in a receding horizon framework without iterative constraint tuning.
• To validate the planner in both simulated environments and real-world experiments using a quarter-scale Roboat platform with moving obstacles.

Proposed method

• The planner uses a receding horizon approach, dynamically constructing a graph from the global reference path and sensor-observed obstacles within the field of view.
• Obstacles are detected via a Velodyne VLP-16 lidar, with range returns within 5 meters treated as obstacle data.
• A modified lidar odometry system enables real-time localization of the ASV in the environment.
• The path search employs a variant of Dijkstra’s algorithm that supports lexicographic optimization over multiple cost criteria ranked by priority.
• Costs are hierarchically penalized: higher-priority objectives (e.g., risk) are minimized first and may be zero in safe regions; lower-priority objectives (e.g., distance) are used only for tie-breaking.
• The framework scales linearly with the number of cost criteria, ensuring computational efficiency for real-time deployment.

Experimental results

Research questions

• RQ1 How can a receding horizon planner effectively balance multiple conflicting objectives—such as safety, heading consistency, and path length—in dynamic urban waterway environments?
• RQ2 Can lexicographic optimization be efficiently applied to graph-based path planning for ASVs without requiring iterative parameter tuning?
• RQ3 How does the hierarchical prioritization of objectives improve path quality compared to single-objective or weighted-sum approaches in obstacle-rich scenarios?
• RQ4 To what extent can the planner maintain safe, feasible paths in the presence of moving obstacles and dynamic environmental changes?

Key findings

• The planner successfully generated safe paths that avoided obstacles while maintaining heading consistency, even when obstacles moved or blocked the reference path.
• In real-world experiments, the robot remained stationary when blocked by two obstacles and resumed navigation once a path cleared, demonstrating robustness to dynamic environments.
• With three criteria (risk, heading, distance), the path achieved a risk cost of 32.7, a heading deviation of 56.4°, and a distance of 8.1 m, showing effective trade-offs via lexicographic prioritization.
• The lexicographic search method achieved high-quality solutions with minimal tuning, as only the objective hierarchy needed to be defined, not weights or constraints.
• The computational complexity scales linearly with the number of cost criteria, enabling real-time performance on embedded systems.
• The method outperformed single- and two-criteria approaches by significantly reducing risk while maintaining acceptable increases in heading and distance costs.