RSG is an action-strategy game, currently in development, about robots and aliens.
The Plot
You play as a roboticist named Soan, who arrived for his first day of work on a space station to find that it was being taken over by strange and hostile aliens.
Soan has some powerful weapons at his disposal, but his true weapon is his knowledge of robotics. He knows that there are a number of robots in operation at the station, but soon realizes that the aliens have secured most of the station. His challenge is to fight his way through it, room by room, finding robots and recruiting them for the cause.
Gameplay
Each robot needs a power source, and there are a limited amount of rechargeable batteries on the station. This means that sometimes the robots can’t all fight at once and Soan has to be smart about which ones to bring into each battle.
Each robot has a handful of major modes that can be adjusted during battle to control their overall strategy in the fight, and they also have a slew of fine-tuning settings that can be tinkered with to emphasize certain behaviors. Robots are built to excel at different jobs. Some robots are better at combat, while others are able to provide support such as repairing or recharging other robots on the team.
Each room represents a descrete encounter, so the player can take time to form a battle plan in terms of which robots to use and how to configure them. Curiously, although the aliens are hostile, they are not particularly malicious–if they win a fight, they’ll let Soan and his team flee, fully recover, and try again.
Technical Information
- Development is using Java using Libgdx.
- The game engine and graphics engine are programmed from scratch.
AI
Both robots and aliens make decisions using an AI model built on A* Pathfinding. AI agents run a routine that which:
- For each accessible destination square:
- For each ability that can be used from the square:
- For each target that the ability can be used on from that square:
- Calculate a score based on factors such as travel distance, travel risk, range, target value, current energy, current health, etc.
- Save the target with the highest score.
- Save the ability/target combo with the highest score.
- Choose the destination/ability/target combo with the highest score and use this as an action plan.
I plan to write another article going into more depth with the AI action planning system in the future.