top of page

Let's assume..

Let's assume that in Australia, in the vicinity of Alice Spring, 15 unmanned aerial vehicles (UAVs) were deployed on special launch platforms, which are designed to perform rescue missions.

Each UAV is equipped with appropriate medical cargo (of various kinds).

Each UAV (or group of UAVs in the swarm) waits for a remote signal from the Emergency Center, which triggers the appropriate UAV and directs it to the indicated place of occurrence of the event (e.g. accident, distress calls or other medical events).

CobraX 1

Let's assume that the Rescue Center received a call from one of the locations (marked with the symbol TAX003) requiring sending 3 (out of 15) UAVs to the scene of the incident. Each of the UAVs has the appropriate cargo (e.g. defibrillator, insulin ampoules, first aid kit), which are necessary at the scene.

 

The operator in the Rescue Center must decide which 3 UAVs (out of 15) are most suitable to complete the mission.

What decision will he make? There are many factors that affect the selection of the best solution:

  • The distance that each UAV has to cover seems to be the most important factor; in addition, the maximum range of the UAV should be taken into account

  • Technical condition of the UAV (e.g. counted by the number of completed missions)

  • The status of IoT devices responsible for communication with the Rescue Center with the UAV (measured, for example, by the strength of the data transmission signal)

  • Weather conditions occurring at the UAV parking place, at the destination and on the flight route

  • The type of medical cargo that is necessary at the scene of the incident

  • ...

  • ​

CobraX 2

Compare the 3 solutions that are presented below:

CobraX
CobraX
CobraX

Each of these solutions seems good enough, but to be sure it is optimal, you need to take into account all the parameters of the mission (UAV properties, target properties, distance properties, rules and limitations).

If the operator of the Rescue Center were to perform such an analysis manually (e.g. by assigning specific properties and then drawing all possible combinations of solutions), the implementation of the mission would be ineffective (try to draw on a piece of paper all possible combinations of the allocation of 3 UAVs from 15 to 1 goal; you will find that this is not an easy task).

 

Below all possible connections of 15 UAVs with 3 target locations are graphically presented:

CobraX

Let's assume that your task (as the operator of the Rescue Center) is to allocate all 15 UAVs of 5 for each of the 3 target locations.

 

Do you know how many possible combinations of allocation there are? 756 756.

 

If, in addition, it is necessary to take into account the properties of objects (UAV, target location, distance), take into account the rules (e.g. that the cruising speed of the UAV cannot be greater than 10 m/s) and limitations (e.g. that the safe range of the UAV is 80% of its nominal range), perform the task in a limited time (and this is required by rescue missions) is impossible ...

cobra_X_blue_transparent.png

That's why we designed The COBRA X

​

an expert decision support system,  

that is using artificial intelligence mechanisms (genetic algorithms)

to optimise UAV allocation for the purpose of

and the execution of

autonomous swarm missions

Below are the results of the COBRA X genetic algorithm for the example of allocating 15 UAVs of 5 for each of the 3 target locations presented above.

CobraX

This solution is suboptimal (best possible), taking into account the values of unfitness (i.e. the degree of adaptation to the feasibility of the task) of all objects (UAV, target location, distance, IoT devices). Please note that there is no risk of in-flight collisions in this solution (no UAV routes intersect with each other).

The total distance of all UAVs to individual locations (TAX001, TAX002, TAX003) is the shortest and the adoption of the solution proposed by COBRA X guarantees the operator of the Rescue Center that all existing factors have been taken into account in the optimisation.

The middle solution from the entire solution space (756,756) is different. There is a potential risk of collision and the distance to TAX is longer.

CobraX
CobraX

The worst (in terms of non-adaptation to the mission) solutions is presented on the left. The number of potential collisions is the highest and the lead time is many times longer than for the best solution.

 

You can analyze all the solutions in COBRA X. You can also trust COBRA X and automatically choose the best (optimal or suboptimal) solution.

 

How did we design our proprietary genetic algorithm?

We map the properties of objects into genes, from a set of genes we create chromosomes, from a set of chromosomes we create genotypes, then from genotypes we create parental pairs, we perform crossing and optionally mutation and normalization.

cobra_X_blue_transparent x.png

In the "Functions" tab, we present how we have translated very complex algorithms into a clear and understandable user interface (GUI) of COBRA X.

​

COBRA X is part of the DABI (Drones Advanced Business Intelligence) platform.

 

Therefore, in the "Links" tab you will find links to our other pages that describe the individual components of this platform.

​

Download CobraX information sheet.

bottom of page