Free Access
TABLE 1:
Brief comparison of centralized and decentralized systems.
| ASPECT | CENTRALIZED CONTROL | DECENTRALIZED CONTROL |
|---|---|---|
| Decision-making | All decisions made by a ground station or central planner | Decisions made locally by each drone, onboard |
| Control philosophy | Top-down, explicitly commanded behavior | Bottom-up, emergent behavior from local interactions |
| Predictability | High; behavior is deterministic in space and time | Lower; behavior can be non-deterministic and emergent |
| Global optimization | Strong capability for globally optimal planning | Limited; optimization is local and distributed |
| Scalability | Limited by computation and communication load on central unit | High in principle, but complex in practice due to interactions |
| Robustness to agent failure | Moderate; individual failures manageable, but central node is critical | High; no single point of failure at control level |
| Communication structure | Hierarchical, drone-to-ground dominant, strong dependence on reliable link to central station | Local, often peer-to-peer or broadcast-based, less dependent on long-range communication, more on local links |
| System complexity location | Centralized (ground station / controller) | Distributed across all agents |
| Onboard intelligence required | Low; drones execute predefined commands | High; drones must make autonomous decisions |
| Hardware cost per drone | Lower (simpler agents possible) | Higher (more computation and sensing onboard) |
| Verification & certification | Easier to test, simulate, and certify | Difficult due to emergent and nonlinear behavior |
| Failure mode | Potential single point of failure | Gradual degradation rather than collapse |
| Operator role | Direct control of “what and how” | Defines only “what”; system determines “how” |
| Adaptability to local changes | Slower; requires central re-planning | Fast; local real-time adaptation |
| Typical applications | Surveying, inspection, precision agriculture, drone shows | Research systems, exploration, dynamic environments |
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.