Abstract
Obstacle avoidance is one of the essential and indispensable functions for autonomous mobile robots. Most of the existing solutions are typically based on single condition constraint and cannot incorporate sensor data in a real-time manner, which often fail to respond to unexpected moving obstacles in dynamic unknown environments. In this paper, a novel real-time multi-constraints obstacle avoidance method using Light Detection and Ranging (LiDAR) is proposed, which is able to, based on the latest estimation of the robot pose and environment, find the sub-goal defined by a multi-constraints function within the explored region and plan a corresponding optimal trajectory at each time step iteratively, so that the robot approaches the goal over time. Meanwhile, at each time step, the improved Ant Colony Optimization (ACO) algorithm is also used to re-plan optimal paths from the latest robot pose to the latest defined sub-goal position. While ensuring convergence, planning in this method is done by repeated local optimizations, so that the latest sensor data from LiDAR and derived environment information can be fully utilized at each step until the robot reaches the desired position. This method facilitates real-time performance and has little requirement on memory space or computational power, making it ideal for small low-cost autonomous platforms.
Introduction
Navigating a robot through a dynamic, unknown environment is a complex balancing act. Traditional global path planners, which rely on pre-existing maps, often fail when unexpected obstacles appear. Conversely, local planners like the Artificial Potential Field (APF) method are computationally efficient but prone to "local minima" traps—situations where a robot gets stuck behind U-shaped obstacles because it cannot perceive a way around them.
This research addresses the need for a navigation system that is real-time, robust against moving obstacles, and efficient enough to run on low-cost autonomous platforms using simple 2D LiDAR sensors. The goal is to achieve global convergence through iterative local optimizations.
Methodology
The proposed system operates through a continuous loop of local optimizations. The process is divided into four primary steps, as illustrated in the flowchart below.
Sub-goal Selection
Instead of calculating a single global path, the robot identifies a "sub-goal" within its immediate field of view based on the latest LiDAR scan. This selection is governed by a multi-constraint cost function that balances three factors: the distance to the final goal, the robot's current heading, and the overall direction toward the destination. This ensures that each local step contributes effectively to the global objective.
Improved Ant Colony Optimization (ACO)
Once a sub-goal is established, an improved ACO algorithm plans the local trajectory. The authors introduced two key enhancements to the standard meta-heuristic:
- Corner Heuristic: A mathematical penalty for sharp turns is integrated into the transition probability. This forces the algorithm to find paths that are not just short, but smooth enough for a physical robot to follow.
- Weighted Pheromone Update: A ranking system allows only the most efficient "ants" to update the path pheromones, which significantly accelerates the algorithm's convergence speed.
Results and Discussion
The method was evaluated using MATLAB simulations and real-world experiments with an EAI differential drive robot. The performance was compared against standard ACO, Artificial Potential Field (APF), and Bacterial Foraging Optimization (BFO) methods.
As shown in the comparative analysis, the proposed method consistently outperformed existing techniques. In real-world trials, the robot reached its goal in 81 seconds, notably faster than APF (92.5s) and BFO (88s). Furthermore, the path length was reduced to 16.2m compared to 18.5m for APF, which tends to take inefficiently wide turns to avoid obstacles. The improved ACO also demonstrated superior convergence, reaching optimal paths in approximately half the iterations required by traditional versions. When faced with sudden external disturbances, the system proved capable of re-planning nearly instantaneously.
Conclusion
The proposed multi-constraints method provides a robust solution for real-time navigation on resource-constrained hardware. By prioritizing iterative local optimizations over global mapping, the system remains highly responsive to dynamic environmental changes while ensuring path optimality. Future work will explore extending this framework to 3D environments for aerial robotics and integrating predictive modeling to anticipate the trajectories of moving obstacles.
Summary updated