The goal of this project is to achieve a collaboration between UAVs to extinguish fire outbreaks in forests and finding a balance between the number of UAV to be deployed and the on-board hardware communication range per UAV using CNP Protocol.
A multi-agent system (MAS) is a computerized system composed of multiple interacting intelligent agents. They can solve problems that are difficult or impossible for an individual agent. Intelligence may include methodic, functional, procedural approaches, algorithmic search or reinforcement learning. Multi-agent systems consist of agents and their environment. Typically, multi-agent systems refer to software agents. However, the agents in a multi-agent system could equally well be robots, humans or human teams. A multi-agent system may contain combined human-agent teams.
CNP is a task-sharing protocol in multi-agent systems, consisting of a collection of nodes or software agents that form the `contract net'. Each node on the network can, at different times or for different tasks, be a manager or a contractor.
When a node gets a `composite task' (or for any reason cannot solve the present task) it breaks the problem down into sub-tasks (if possible) and announces the sub-task to the contract net acting as a manager. Bids are then received from potential contractors and the winning contractor(s) are awarded the job(s).
CNP implements a task assignment which elects a manager in order to allocate tasks to nearest UAV. The manager allocates task to the neighboring UAV’s in the communication range of contractor by making sure that these UAV’s aren’t allocated with any task as of then. The contractor sends out data about fire outbreaks to UAV’s in communication range. The UAV’s calculate distance from their position to the fire Centre as the proposal to contractor based on Proposal with a relation between number of cells on fire and the distance to reach the center of these fire cells. Based on the best Proposal, the contractor allocates task to the UAV’s until the fire is extinguished. The cells that are completely burned cannot be recovered and the contractor doesn’t assign any task for those cells to any of the UAV’s as these cells cannot be recovered. The UAV’s that are not in communication range will do a random walk and checks if there is any fire.
The UAV’s share their position with the Contractor and all other UAV’s in the communication range. If anyone of them first see the fire, it will share the data with the others in the range. If one of them is lost from the range, it will use the data shared to it until then for executing the task or for coming back to the range. If all the fire in the area is extinguished then they share this data and move to another area in fire or the task will be terminated if no fire is existed in the forest.
The project was built using NetBeans 8.2 - Java setup on a MAC system.
The results were obtained after simulating with different combinations of communication range, Number of fires in the forest, Number of cells saved and Number of UAV’s used for extinguishing fire. The Inferences can be drawn using one of these parameters as constant and relation was drawn between the other two. All the results were taken after a single run with the changed parameters. Starting with the communication range of 20 and changing the No. of UAV’s to see the Cell Recovery trend. When the UAV’s have increased the no. of cells recovered have increased and when the drones are more enough the extinguished cells are less as they started extinguishing cells so fast so the fire didn’t spread across the neighboring cells.
The Number of UAV’s are increased up to 60 UAV’s with a step of 5 starting with 20 UAV’s with a communication range of 20. Again, the simulations were made with increased communication range by 10 till the end communication range is 60. Every simulation config ran for 5 times and their average is taken into result for that particular config. Total Number of Simulation = 9(UAV’s)* 5(CRange) *5(Runs) = 225 simulations
The simulation time taken to extinguish fire is decreasing with increase in number of UAV’s and Communication Range. The number of cells saved including both recovered and kept normal are increasing with increase in Number of UAV’s and Communication Range. A little change of trend is observed as the number of UAV’s increased beyond 55 as there is collision in UAV’s for the assigned task as they are more than needed.
