Abstract
This paper introduces a distributed "Rolling Dispersion Algorithm" that allows a small team of robots to fully explore unknown, hazardous environments like disaster zones. By prioritizing "sweep coverage" over traditional "blanket coverage" and using wireless signal intensity to guide movement, the system ensures robots maintain connectivity even under extreme communication restrictions.
Introduction
In the wake of disasters like earthquakes or fires, sending human teams into unstable structures is dangerous. While robot teams can be deployed to map these areas, most existing "coverage" algorithms are designed for surveillance (blanket coverage), which requires a massive number of robots to monitor every inch of a space simultaneously. Furthermore, disaster zones often suffer from "communication-denied" or restricted environments where standard Wi-Fi fails due to structural or chemical interference. This research addresses the challenge of achieving full exploration with a minimal number of robots operating without a pre-existing map or reliable, high-bandwidth communication.
Methodology: The Rolling Dispersion Algorithm
The core of the research is the Rolling Dispersion Algorithm. Unlike static coverage models, this is a distributed approach where robots move through the environment in a "rolling" fashion.
Key Technical Features
- Sweep Coverage: Instead of attempting to monitor the entire area at once, the team performs a methodical single pass. This significantly reduces the required team size compared to blanket coverage.
- Signal-Based Movement: Robots utilize wireless signal intensity (RSSI) as a primary sensor to direct movement, ensuring they remain within range of neighbors.
- Beacon Deployment: Physical beacons (e.g., ZigBee motes) are dropped to act as breadcrumbs, marking paths and identifying unexplored areas or dead ends.
- Behavioral States: The algorithm defines specific behaviors based on local context, such as Avoid Collisions, Seek Connection, and Guard.
Experimental Evaluation
The system was evaluated through both simulated and physical experiments:
- Simulated Environments: A complex cave-like environment was used to test exploration efficiency. This setup was designed to be challenging, typically requiring ten robots for traditional coverage.
- Physical Lab Experiments: Tests were conducted using physical robots tracked by overhead cameras. To simulate restricted communication (analogous to chemical signaling), researchers used a screen below the robots to display colors detected via filtered IR sensors.
Key Results and Insights
The "Rolling Dispersion" approach treats communication as a limited resource rather than a given. By leaving beacons behind, the robots create a temporary infrastructure that allows the team to "clear" an area and move forward.
As shown in the performance data, while a team of five robots can eventually explore the entire area, increasing the team size to eight robots significantly improves efficiency, reaching full exploration in approximately 150 seconds compared to 250 seconds for the smaller team. Furthermore, the sweep approach successfully explored environments that would typically require at least ten robots using traditional methods.
Conclusion and Future Work
The proposed algorithm demonstrates resilience in communication-restricted environments, proving that high-bandwidth Wi-Fi is not a prerequisite for complex coordination. Future work will focus on formalizing the minimal message count required for full exploration and integrating these autonomous networks with human rescue teams to facilitate search and rescue operations in dynamic environments.
Summary updated