US Army Combat Capabilities Development Command conducted research on robotics over the past decade. But, only recently has the military begun to integrate these technologies into its battlefield. They have been testing robots, using them to detect IEDs or recover Ravens. The US Marine Corps also experiments with recon robots for urban warfare. These robots work together with Soldiers. These robots can execute tactical actions according to plan, and can suggest changes to plans.
The US Army has developed an algorithm-based teaching system that will allow Soldier robotics to communicate with each others. This system will enable the robots to understand the Soldier's intent through spoken language, and execute the appropriate actions. The robots can also ask questions to overcome uncertainties. This will allow the robots to perform joint tasks at operational speed, and it will help to reduce the learning curve for the Soldiers.
The most difficult part of creating an algorithm-based learning system was figuring out what questions to ask. The team created a natural language understanding system that can process Soldier's spoken language. It was based on statistical classification techniques. The system was trained on a small number of humanrobot dialog. This allowed researchers to gradually reduce the variance in natural language. The team is currently trying to connect the output semantic representation of Soldier's speech language with the grounding system.
The goal of the algorithm-based learning system is to improve Soldier-robot teaming in tactical environments. This will allow the robots and Soldiers to communicate using natural language. The research team developed a natural language processing pipeline that interprets the intent of a Soldier's spoken language. The system will eventually be able control the variation in spoken languages. This will allow robots that understand spoken language to be more collaborative and will enable them act on it.
Robots will be able to ask Soldiers questions using the algorithm-based learning system. This will allow them to gain a better understanding about their environment. These questions will be based upon information from maps. This information will enable robots to make more autonomous decisions based on this information. These questions are tactically relevant. These questions examine the process by which robots make decisions, and the information the Soldiers need. These questions were created to get feedback from Soldiers. The Soldiers' information was used to ensure that the robots could identify information on maps that was relevant to tactical decision making.
Soldier-bot critical variant is faster in attack and fires critical rockets more quickly than other Soldierbot variations. These rockets inflict six times the damage as regular rockets. Critical variants have higher health and can kill players without crit resistance.