tabathaviso / uav-swarm-path-planning

uav-swarm-path-planning's Issues

User inputs

RunSim.py user input for scenario selection is convenient.
Let's expand this feature to ask the user for the desired number of swarm agents.

Separate Files Please

Please organize classes into separate files
Please

Missing evaluation methods

Add the following evaluation methods:

• check ability to generate successful flight path in all scenarios (3 obstacle configs x 3 swarm sizes)
• completeness
• execution time, time complexity
• space complexity
• optimality (cost)
• convergence effect
• statistical results -> optimal, mean, worst values + std deviation (20 runs per scenario)
• scalability (compare results from 1-, 3-, 5-agent swarms)

Multi Agent Environment Needs Fixing

Single UAV works, issue is that multiple agent sim throws errors:

• A* might be getting confused
• Or path might not be clearly returned
• Need to add some kind of test to see what is actually happening

If Multi-agent sim doesn't work by EOD 7/2/23, transition to single UAV sim for phase 2 progress report

Simplified GWO needs continued development

Tabatha to:

• initialize search space
• update fitness equation
• develop hierarchy function
• implement GWO search loop with wolf population initialization, fitness + search iterative loop, path reconstruction

make a pretty readme

Rough Draft of MADDPG

Code isn't ready to commit to my branch but wanted to start sharing it:

https://wpi0-my.sharepoint.com/:f:/g/personal/sltseng_wpi_edu/EjQ7N-97b95Iu9Y81REkWFwBsWHLIjFxbCaDpsT_kUZTDg?e=m2B01V

@wyatt522, basically I got an outline of what I believe are all the needed functions.
Working on:

• creating functions to modify parameters for each agent after every iteration
• passing correct variables and values between functions
• having MADDPG output compatible data for the main MultiAgentEnvironment.py file to plot