Abstract
This paper addresses mutual collision avoidance between multiple mobile robots based on a layered strategy. In this strategy, static motion generation and dynamic motion generation of several levels are provided, and a proper level of dynamic motion generation is selected for mutual collision avoidance according to the complexity of the situation. We have implemented two typical methods in the layered strategy, which are rule-based local collision avoidance and negotiation-based global one using communication. In each method, a robot detects collision and applies as a local method as possible. Experimental results show two actual mobile robots can achieve mutual collision avoidance based on the layered strategy.
Introduction
In the mid-1990s, multi-robot systems faced a significant hurdle: how to avoid collisions efficiently without succumbing to deadlocks or overwhelming communication networks. Sensor-only methods, while fast, often lack awareness of other agents' intentions. Conversely, purely communication-based methods provide global awareness but can be computationally expensive. This paper proposes a layered strategy that balances these extremes by being "as local as possible but as global as necessary."
The Layered Strategy for Collision Avoidance
The core contribution of this research is a hierarchical framework that organizes different avoidance behaviors into levels of complexity and communication requirements.
The system operates on two main levels of motion generation:
- Static Motion Generation: Initial path planning based on a known map of stationary obstacles.
- Dynamic Motion Generation: Real-time adjustments to avoid other moving robots.
Within the dynamic layer, the system transitions from simple "traffic rules" to complex wireless negotiation as situations become more difficult.
Operational Flow
The robot's decision-making process is designed to prioritize local resolution before escalating to global communication.
When a robot detects a potential collision during motion execution, it triggers the dynamic motion generation module. It first attempts Rule-based Avoidance (Local), using sensors to apply simple rules like "move to the left" for head-on encounters. If these rules fail—for instance, if an obstacle prevents the required maneuver—the robot initiates Negotiation-based Avoidance (Global) via wireless communication.
Priority-Based Negotiation
During negotiation, robots exchange IDs and calculate a Priority Value ($C$) to determine right-of-way: $$C = Ce + Ct + Cp$$ Where $Ce$ represents environmental constraints, $Ct$ is task urgency, and $Cp$ is the robot's physical maneuverability. This logical framework ensures that the robot with the most critical need or the least flexibility is given precedence, effectively resolving potential deadlocks.
Experimental Evaluation
The proposed system was validated using two custom mobile robots (MR-1 and MR-2) in a hallway environment. The robots were equipped with ultrasonic sensors and wireless modems, coordinated by a "Global Environment Manager" (GEM).
Key findings include:
- Efficient Local Resolution: In simple scenarios, robots successfully avoided each other using only local rules, preserving communication bandwidth.
- Deadlock Resolution: In complex scenarios where obstacles blocked standard avoidance paths, the robots successfully transitioned to negotiation, exchanged priorities, and resolved the conflict.
- Robustness: The layered approach allowed the robots to reach their goals even when the environment was more complex than initially mapped.
Conclusion
By integrating reactive behaviors with deliberative negotiation, this layered strategy provides a robust foundation for multi-robot coordination. While the 1996 study focused on two robots, the principles of hierarchical complexity and quantified priority remain highly relevant to modern autonomous vehicle and swarm robotics research.
Summary updated